The all-seeing eye

State-of-the-art vision sensors for unmanned craft

Driverless 4WD
MIRA’s autonomous Land Rover
Danielson diesel
Inside the new generation Scion
SA-400
Autonomous helicopter technology

UST 02 : SPRING 2015

UK £15, USA $30, EUROPE e22 [Link] [Link]

 Spring 2015 | Contents

40 68

18
62 48

04 Intro 40 Focus: Vision sensors

Interest in swarming unmanned systems is growing, as the Developments in imaging technology are paving the way for
benefits of having them work in harmony become clearer more integrated onboard vision, as Nick Flaherty explains

06 Platform one: Mission-critical info 48 Dossier: Danielson Trident 100 TD2

High-altitude internet base station UAV unveiled, details of Ian Bamsey investigates this state-of-the-art French turbodiesel
a flexible and high-efficiency solar sheet, a real-time tool for that’s been created specifically for UAV use
handling complex imagery that runs on a standard web browser, 
new swarming algorithms, and much more… 56 Insight: Security and safety systems
How to ensure an unmanned system is secure and safe
14 In conversation: Graham Smethurst
BMW’s head of its autonomous vehicle programme explains the 62 Dossier: MIRA MACE
challenges that lie ahead and how to overcome them The UK’s MIRA automotive research institute is trialling unmanned
ground vehicles for the military. What lessons are being learnt?
18 Dossier: Scion SA-400
Ian Bamsey investigates an optionally piloted helicopter which, 68 Focus: Additive manufacturing
unmanned, can fly sideways at 46 knots and land on a ship at sea Making components using an additive process is ideal for the
unmanned systems industry. David Cooper explains why
32 UST online
Full details on the easy way to find new suppliers and stay 76 Insight: Marine UUVs
up to date with the latest technological developments Torpedo-shaped autonomous underwater vehicles are taking over
the oceans. Nick Flaherty examines some new developments
34 Report: Commercial UAV Show
Ian Bamsey and Nick Flaherty highlight some of the new 82 PS: Wankel engines
technology on show at this trade fair launch The rotary engine has found a niche in unmanned systems

Unmanned Systems Technology | Spring 2015 3

Intro | Spring 2015

Going with
the flow
The all-seeing eye
State-of-the-art vision sensors for unmanned craft
Driverless 4WD
MIRA’s autonomous Land Rover
Danielson diesel
Scion Volume One | Issue Two

W
Inside the new generation

SA-400 Spring 2015

hen an airstream enters a Autonomous helicopter technology

convergence, the individual High Power Media Limited

molecules speed up so that
Whitfield House, Cheddar Road,
they can all get through the
Wedmore, Somerset,
restriction without delay. As you
will appreciate, that acceleration is a basis on BS28 4EJ, England
which the clever aerodynamicist can generate
UST 02 : SPRING 2015

Tel: +44 (0)1934 713957

reduced pressure where that might be useful.
UK £15, USA $30, EUROPE e22 [Link] [Link]

Fax: +44 (0)20 8497 2102

I mention it here simply to emphasise what [Link]
inevitably happens when a flow of traffic meets a
restriction. ISSN 2056-9823
What, though, if that traffic is under robotic Printed in Great Britain
rather than direct human control? ©High Power Media
It isn’t surprising that there is increasing All rights reserved. Reproduction
development of the swarming of unmanned Editorial Director (in whole or in part) of any article
vehicles – let’s not just develop one that operates Ian Bamsey
or illustration without the written
in isolation but a group working in harmony to
mutual benefit. Is it any wonder then that the ripe Technology Editor permission of the publisher is strictly
field of autonomous transport is so attractive to Nick Flaherty prohibited. While care is taken to
today’s best inventive brains? ensure the accuracy of information
Contributor herein, the publisher can accept no
Welcome to the second issue of Unmanned David Cooper
Systems Technology, the communications liability for errors or omissions.
hub for a new breed of engineering pioneers. Technical Consultants Nor can responsibility be accepted
Whatever your interest in this field, I feel sure Charles McCarten, for the content of any advertisement.
you will find inspiration through our pages, Paul Weighell
whether from helicopters operating in unison or US COPIES ONLY
swarming submarines. Production Editor
Guy Richards Unmanned Systems Technology
Go with our flow. You won’t suffer the pressure
ISSN No: 2056-9823, is published
of data congestion – you will accelerate vital
Design by High Power Media in England
information gathering!
Andrew Metcalfe and is distributed in the United States
Ian Bamsey | Editorial Director andrew@[Link]
by SPP, 17B S Middlesex Ave,
UST Ad Sales Monroe, NJ 08831. Periodicals
Please direct all enquiries to postage paid at New Brunswick, NJ.
Editorial opportunity Simon Moss Postmaster please send address
simon@[Link] changes to: Unmanned Systems
Are you enthused by the world of
Unmanned Systems Technology to the Technology, 17B S Middlesex Ave,
Subscriptions
extent that you would like to help document Jim Cavanagh Monroe, NJ 08831.
its emergence? jim@[Link]
This publication has had such an SUBSCRIPTIONS
enthusiastic reception that we are expanding Publishing Director Subscriptions are available from
fast. That has opened the door for more Simon Moss High Power Media at the address
experts to contribute to our pages. If you above or directly from our website.
General Manager
have a strong technical knowledge of any Overseas copies are sent via air mail.
Chris Perry
aspect of the field and would like to explore
the possibility of writing for us, do let me Office Administrator UST SPECIAL 1 YEAR
know. You don’t need prior experience of Frankie Robins
the media, simply a desire to explore SUBSCRIPTION OFFER
further and share your expertise: 4 issues per year; 15% off:
ian@[Link] UK – £50; Europe – e72
USA – $100; ROW – £65
Make cheques payable to
High Power Media. Visa, Mastercard,
Subscribe today @ Online Publisher
Mike Rees
Amex and UK Maestro accepted.
Quote card number and expiry date
[Link] mike@[Link] (also issue/start date for Maestro)

4 Spring 2015 | Unmanned Systems Technology

 Platform one
Mission-critical info for UST professionals

A four-fuselage configuration of
UAVOS’ atmospheric satellite

UAVOS flexes its wings

Atmospheric satellite The individual operation of each buffer battery that can be protected in a
Autopilot specialist UAVOS used the fuselage and the inherent stability of thermal box in its specific fuselage.
Commercial UAV show at Olympia (see the combined unit help protect against The prototype on show at Olympia
page 34) to showcase its new-concept turbulence – indeed, the control system had six fuselages and a 50 m wingspan.
atmospheric satellite, an autonomous manipulates the entire craft to obtain the Head of r&d at UAVOS, Aliaksei
UAV that hovers at high altitude, most efficient flight. The wing carries solar Stratsilatau, remarked, “The proposed
normally to act as a base station for panels, and these top up the batteries aerodynamic arrangement has a linear
internet connection. A long wingspan that power the propeller motor in each weight increase relative to the span,
is a fundamental requirement for this fuselage. The batteries also power the unlike a classical arrangement. This
application, and rather than have a single autopilot-based control system as well as allows us to create an aircraft significantly
central fuselage carrying the propulsion the payload. At night, in the absence of different in its performance from the
system and the payload, the UAVOS craft solar energy, the craft is allowed to glide. previously designed airframes.”
consists of four or more fuselages linked Over ten hours it will drop from around Stratsilatau added that the design makes
by a single wing. Each fuselage takes an 30,000 m to around 15,000-20,000 m – use of the company’s distributed control
equal share of the payload, and each has still above civil aviation airspace – and system, which he said the company has
its own propulsion system. then the next day it will be powered back been working on for more than ten years
The wing has a high degree of up to its original altitude. across a number of unmanned systems.
flexibility, and the individual fuselages The craft has been designed to “Our concept is different because it
each have an autopilot that is central to allow a total flight duration of about has dual CAN potential and distributed
controlling its dedicated propeller motor, five years, which is the planned routine processing,” Stratsilatau said. “The system
its horizontal tail elevator and winglets maintenance interval given its (advanced consists of several nodes with different
with optional air brakes on the adjacent composite) construction. Having landed functionality and all connected in a CAN
wing sections. The various controllers for refurbishment, it needs additional network. We connect as many nodes as
have (wired) data links, and the control power for its initial climb to 30 km, and necessary to achieve redundancy and
software operates the entire craft as a this is provided by a battery system that distribute the computing. This means that if
single unit. If one of the fuselages starts is dropped back by parachute once it you lose a node nothing changes and the
to move out of a desired position relative has done its job. There is a temperature system remains stable.”
to the others then the control system will issue for batteries at high altitude, so this The base modules measure just 25 x
react to stabilise the craft through the approach allows the ongoing use of a 36 mm and are stackable to form a more
common wing, which can twist and bend. very small, lightweight lithium-polymer complicated node if required.

6 Spring 2015 | Unmanned Systems Technology

 Platform one

Cool approach to power

Solar energy

MicroLink Devices has developed a

lightweight, flexible, high-efficiency
solar sheet to enable UAV applications
calling for extended flight endurance,
the company’s director of solar sheet
products, Ray Chan, has reported.
MicroLink manufactures the individual
solar cells that are incorporated into
the solar sheet using a proprietary
epitaxial lift-off process resulting in
cells that operate at greater than 30%
AM1.5 efficiency at room temperature.
New-generation solar cells from MicroLink
The efficiency rises further at low Devices are extending UAV flight times
temperatures, unlike some solar cell
technologies which struggle to operate
efficiently under similar conditions, Chan MicroLink has also been working with
noted. The cells operate Imperial College London, Rochester
Completed solar sheets can be placed Institute of Technology, and the US
over curved surfaces such as aircraft at more than 30% Naval Research Laboratory on a new
wings and can generate more than triple-junction solar cell that could
250 W/m2 and over 500 W/kg under efficiency, which provide an unprecedented 50%
AM1.5 illumination. conversion efficiency. At present, the
The cells are produced using metal rises further at world record for triple-junction solar
organic chemical vapour deposition. A cell efficiency is 44%, and Chan added
release layer is first deposited directly low temperatures, that it is generally accepted that a
on top of a gallium arsenide (GaAs)
substrate, and multi-junction solar cells unlike some cell major technology breakthrough will be
required for the efficiency of these cells
are then deposited on the release layer.
A thin, flexible metal backing layer is
technologies that to increase much further.
The three-year programme is to
then applied to the solar cell structure to
provide mechanical support for the solar
struggle under develop multi-junction solar cells
where each junction is ‘tuned’ to
cell after epitaxial lift-off, which allows the
solar cell structure to be removed from
such conditions different wavelength bands in the solar
spectrum to increase efficiency. High-
the GaAs substrate without damage. bandgap semiconductor material is
Conventional solar cell fabrication is then used to absorb the short-wavelength
performed to create the multi-junction radiation, with longer wavelength
solar cells which generate the power. of up to 60 mph with a flight time of parts being transmitted to subsequent
This approach brings significant cost an hour, while the larger Puma has a semiconductors.
reductions, Chan reported, since the range of 9.0 miles and an approximate The group is using a combination of
GaAs substrate can be re-used to grow endurance of two hours. Chan remarked indium, aluminium, arsenic and antimony
other solar cells. that in collaboration with the US Air as a high-bandgap material layer that
MicroLink’s solar sheet technology has Force Research Lab’s Materials and can be grown and matched to a lattice
been applied to the Raven and Puma Manufacturing Directorate, MicroLink’s of indium phosphide. As a result, the
small, electric-powered UAVs made solar sheets were integrated with the team has created a solar cell design that
by AeroVironment. Using a lithium-ion existing power systems of both UAV is a potential route to 50%-plus power
battery the Raven has a range of up platforms and demonstrated a doubling conversion efficiency under concentrated
to 6.2 miles and can reach speeds of flight endurance. solar illumination.

Unmanned Systems Technology | Spring 2015 7

Platform one

New orbital endurance record

Space flight Force. The vehicle is a quarter of the size It is controlled by an autopilot to keep
The X-37B space vehicle has smashed of the Space Shuttle, being 9 m long it in low Earth orbit and to oversee the
the endurance record for an autonomous and having a wing span of 4.5 m, and is landing, and is being used to test out new
vehicle, staying aloft for 674 days, or just powered by an Aerojet AR2-3 hydrazine ceramic materials for heat protection. A
over two years. The craft, also called the engine and gallium arsenide solar cells key feature is that there are no hydraulics
Orbital Test Vehicle (OTV), was on its and lithium batteries while in flight, on the craft – all the actuation is electro-
third mission and was built by Boeing although it is launched using a traditional mechanical to enhance the endurance
before being transferred to the US Air rocket system. and controllability.

UST diary

AUVSI Europe UAS

Tuesday 3 March – Wednesday 4 March Tuesday 7 July – Wednesday 8 July
Crowne Plaza, Brussels, Belgium Twickenham, UK
[Link]/UnmannedSystemsEurope/Home [Link]

DoD Unmanned Systems Summit World Congress on Unmanned Systems Engineering

Wednesday 18 March – Thursday 19 March Thursday 30 July – Friday 31 July
Alexandria, VA, USA Palacio de Santa Paula, Granada, Spain
[Link] [Link]

SkyTech InterDrone
Friday 24 April Wednesday 9 September – Friday 11 September
Business Design Centre, Islington, London, UK Rio, Las Vegas, USA
[Link] [Link]

Small Unmanned Systems Business Exposition DSEI

Wednesday 29 April – Friday 1 May Tuesday 15 September – Friday 18 September
San Francisco, CA, USA ExCel, London, UK
[Link] [Link]

AUVSI USA Unmanned Systems Institute (USI) Conference

Tuesday 5 May – Thursday 7 May Wednesday 23 September – Friday 25 September
Georgia World Congress Center, Atlanta, GA, USA Las Vegas, USA
[Link]/auvsi2015/public/[Link] [Link]

ICUAS’15 – The 2015 International Conference Unmanned Global Systems (UGS)

on Unmanned Aircraft Systems Wednesday 14 October – Thursday 15 October
Tuesday 9 June – Friday 12 June Bordeaux Exhibition Centre, Bordeaux, France
Denver Marriott Tech Center, Denver, CO, USA [Link]/en
[Link]
Commercial UAV Show
Paris Air Show Tuesday 20 October – Wednesday 21 October
Monday 15 June – Sunday 21 June ExCel, London
Le Bourget, Paris, France [Link]/exhibition/the-commercial-uav-show
[Link]/EN
Unmanned Systems Canada
Commercial UAV Show Asia Tuesday 3 November – Thursday 5 November
Tuesday 30 June – Wednesday 1 July World Trade & Convention Centre, Halifax, NS, Canada
Suntec Singapore Convention & Exhibition Centre, Singapore [Link]/
[Link]/exhibition/commercial-uav-asia

8 Spring 2015 | Unmanned Systems Technology

Worldwide web of access
The browser-based system can handle complex imagery in real time
UTS software
UTC Aerospace Systems (UTCAS) has
launched the first real-time tool that
can run in a standard web browser for
handling the complex imagery acquired
from unmanned systems anywhere in the
world. Using a browser means the SCI
(Simplifying Complex Information) Toolset
can be used on any terminal to access and
search a federated network of libraries of
images or data, including real-time feeds,
which could be anywhere in the world. The
tool can then combine any location and
time information automatically or manually,
and integrate these into reports quickly
and easily, simplifying document flows in
enterprise systems.
“The reason you fly is to answer a and access the information purely from
question, and we give that from the We are looking the map location.
software,” said Richard Glyn-Jones, UTC Aerospace Systems has also
business development manager for the at a simplified developed an exploitation tool called
UK at UTCAS, “You have a thumbnail for SCI-X that is hosted on a local machine
the overview, then you stream the still version in a or local server rather than being browser
image so you only pull down the pixels based. This is able to annotate and
you want.” This allows the imaging to be browser, for measure details on the images and save
used in a standard web browser using snapshots to library, as well as create
the JPEG 2000 image standard. example, with reports from the images.
For example, a remote assessor The next stage is to support multi-
working offshore in the Shetland Islands the most popular spectral imaging alongside SAR

features but
can upload still and video image data (synthetic aperture radar) and electro-
from a UAV, annotate and measure optic imaging. “We are looking at a
the key features and have the data
available in real time at head office in it is customer simplified version in a browser, for
example, with the most popular features
London, or anywhere else in the world,
without needing to rely on a complex IT
dependent but that depends very much on the
customer requirements,” Glyn-Jones
infrastructure. Head office can then find said.

Sea change in deck landings

UAV software its small unmanned system to land on ground with a precision of 10-20 m,
Danish start-up Little Smart Things has the deck of a ship. so the company is working with the
been awarded 8.5m kroner ($1.3m) by The project uses the 2 kg Kevlar- Technical University of Denmark and
the Danish Innovation Fund to develop based Cumulus One hand-launched Aalborg University to develop the new
new GPS navigation sensors and more fixed-wing aircraft. This can currently sensor technology by 2017. The project
accurate mathematical models to enable handle automated landings on the starts in March 2015.

Unmanned Systems Technology | Spring 2015 9

Platform one

Clever swarming indoors

DARPA plans to test new algorithms to enable small
UAV software
UAVs to find their way through a labyrinth of rooms
US defence technology research lab
DARPA is looking to test out completely
new algorithms for fast-moving, small UAVs
with a view to extending the technology
into ground and marine systems. The Fast
Lightweight Autonomy (FLA) programme
aims to create a new class of algorithms
to enable small, unmanned aerial vehicles
to quickly navigate a labyrinth of rooms,
stairways and corridors or other obstacle-
filled environments, both individually and
as a swarm.
The programme aims to develop and
demonstrate autonomous UAVs small
enough to fit through an open window and could be applied to autonomous ground, The performance of the algorithms
able to fly at speeds up to 20 m/s (45 mph) marine and underwater systems, which will be evaluated on a DARPA testbed,
while navigating within complex indoor could be especially useful in areas where a six-rotor small UAV called VeloHex.
spaces independent of communication GPS doesn’t work well. This will carry a Gigabyte mini-PC using
with outside operators or sensors and “Urban and disaster relief operations an Intel Core i7 processor with two
without relying on GPS waypoints. would be obvious key beneficiaries, but Lidar laser systems from Hokuyo – its
The algorithms could enhance applications for this technology could UTM-30LX on top, with a 30 m range,
unmanned system capabilities by extend to a wide variety of missions and its UST-2 on the bottom with a 5 m
reducing the amount of processing power using small and large unmanned range – as well as four USB cameras
and communications needed for low- systems linked together with manned from mvBlueFOX. The VeloHex will have
level tasks such as navigation around platforms as a system of systems,” said a total weight of 2100 g and a flight time
obstacles in a cluttered environment. The Stefanie Tompkins, director of DARPA’s of 8-10 minutes from the 4400 mAh
initial focus is on UAVs, but the algorithms Defense Sciences Office. lithium-polymer battery.

Super-chip hits the road

Driverless cars The Tegra X1 has eight ARM CPU “Mobile supercomputing will be
A multicore graphics chip previously cores and 256 graphics cores using central to tomorrow’s car,” said Jen-
used in supercomputer designs is now NVIDIA’s latest Maxwell architecture, Hsun Huang, CEO and co-founder of
being used for image processing for and will drive camera-based ADAS NVIDIA. “With vast arrays of cameras
driverless car systems. The Tegra X1 (advanced driver assistance systems) and displays, cars of the future will
processor from NVIDIA aims to help self- applications such as pedestrian see and increasingly understand their
driving cars advance from the realm of detection, blind-spot monitoring, lane- surroundings. Whether finding their
research into the mass market with its departure warning and street sign way back to you from a parking spot or
automotive-grade version of the same recognition. In the DRIVE PX autopilot using situational awareness to keep out
GPU that powers the world’s ten most platform, two of the X1 chips can of harm’s way, future cars will do many
energy-efficient supercomputers. The process video from up to 12 onboard amazing, seemingly intelligent things.
key has been lower power consumption, cameras to run capabilities for a Advances in computer vision, deep
allowing more powerful applications to seamless 360º view around the car, and learning and graphics have finally put
run within the same thermal limits. true self-parking. this dream within reach.”

10 Spring 2015 | Unmanned Systems Technology

Platform one

The vehicle-to-infrastructure technology is a key element

for driverless cars and is powered partly by an i.MX6
processor from Freescale Semiconductor

Autonomous vehicle driver

Software the ARMv7 architecture, while the i.MX8 OpenCL,” said Lees. “The [Link] 7 has
The increased demand for processing will focus on the automotive sector with morphed into low-power, low-cost
power in automotive ADAS (advanced larger quad and octal devices with 64-bit consumer, non-automotive applications
driver assistance systems) applications ARMv8 architecture such as A53 and A57. with a power consumption of 800 mW
for driverless cars is having a profound These will be coupled with more power- and a target power of 500 mW and
effect on Freescale Semiconductor’s efficient Coretex-M controller cores. lower. The i.MX8 is the focus for the
multicore roadmap. All of this will need a new quad and octal devices for automotive
The i.MX6 family has seen great programming approach, Lees said, infotainment and cockpit imaging with a
success in a wide range of applications, and it is committing to the OpenCL pure v8 A core with ARM-Cortex M sub-
ranging from single-core to quad-core language to do so. OpenCL is a higher processing.”
implementations, but this is changing, level programming language that is This will also see a massive step up
according to Geoff Lees, senior vice- open rather than tied to a particular in GPU processing with the Vivante
president and general manager of the architecture, and is aimed at multicore GC700 cores across the range. “We
microcontroller division at Freescale. systems with elements for detecting will implement massive arrays in the
Instead, the i.MX7 range will focus on which resources are available for i.MX8 family – automotive is a key driver
general consumer applications with running code. and there’s a lot of graphics and image
single- and dual-core devices using “Our [Link] roadmap is all about capture applications to follow,” said Lees.

12 Spring 2015 | Unmanned Systems Technology

Compact lidars
from
For autonomous operation, navigation, terrain mapping
or inspection nothing beats hard data.
Small light and very accurate, Hokuyo lidars provide a
2D digital representation of the environment without
compromising payload or battery life.
Unlike other laser systems these sensors are a
customisable component not a fully integrated solution.
So there are no system constraints and the associated
cost / weight burden is reduced.
Ranges of up to 80m with resolution to 0.125⁰

Contact Sentek Solutions via info@[Link]

or [Link] to find out more.

Sentek Solutions Ltd

[Link] +44 (0)1223 923 930
Safety in numbers
BMW’s manager of its autonomous vehicle programme
tells Nick Flaherty about the challenges in this realm
that lie ahead, and the plans to overcome them

T
he key technical challenge and steering with data from Lidar laser BMW already has its own
facing driverless cars is not cameras, 3600 radar, ultrasonic sensors communications network for its
the technology itself – that and optical cameras that monitor the cars, which tracks the (anonymised)
is pretty much already there, environment. movement of every BMW across Europe,
says Graham Smethurst, Instead, says Smethurst, the technology and Smethurst has detailed video footage
manager of the networked and challenge is very much about the to show this. He points out though that
automated driving programme at the infrastructure that driverless cars will use BMW is not unique here: other major
BMW Group. This was demonstrated in everyday operation. Vehicle-to-vehicle car makers have similar systems, and
clearly at the Consumer Electronics (V2V) and vehicle-to-infrastructure (V2I) one of them, for example, plans to offer
Show in the US in January, where BMW technologies are being developed to a service where parcel couriers can find
showed two autonomous petrol-driven provide data to the vehicles, either from your car, open it and leave your delivery
cars, a 2 Series and a 6 Series Gran cars a few hundred metres in front or in the boot (trunk).
Coupe, which drove themselves around from roadside transmitters, and this is So the challenge is for these car
a test track at speeds of up to 150 mph a vital capability for autonomous cars, makers to collaborate, he says, as no one
and were able to handle drifting in he says, as it provides more time for an company can install and run networks
corners, a high-speed slalom course and autonomous system to respond rather right across the world. The standards
precise lane changes. The technology than relying on local sensors with a here for frequencies and data packets
in the cars combines active braking range of only a few metres. have been defined, but they are slightly

14 Spring 2015 | Unmanned Systems Technology

 Graham Smethurst | In conversation

Graham Smethurst
eventually becoming head of vehicle
information systems at Rover Group. In
2000 he moved across to Rover’s owner
BMW, and since 2009 he has been at
the forefront of the Genivi Alliance, a
collaboration between car component
makers to develop open source
standards for in-car entertainment and
comms systems.
As the infotainment console is also
a candidate to host the central car
controller through devices such as
Nvidia’s X1 processor (see Platform
one), and can also host the V2V and V2I
comms systems, it makes sense to hand
over the autonomous car programme to
the head of the Genivi project, Smethurst
says. The data coming in then has to be
linked to the navigation system within this
console, for example to offer a different
route to avoid problems. All of this
data packets coming into the car, and requires a huge amount of collaboration
then act on them; others want to decrypt between the various hardware and
only high-priority packets. software suppliers, he says.
All this comms data has to be The growing demand for functionality
integrated with the local sensors as such as all kinds of wireless
well. If the V2V or V2I data calls for connectivity, entertainment systems and
an immediate stop, for example if it secure comms networks is pushing
has been informed of a crash ahead development costs ever upwards, so
BMW has shown that an autonomous
car can handle drifting in corners and other cars are stopping sharply, Genivi’s approach is to allow OEM
(Photos courtesy of BMW) then that command may have to be solutions to compete successfully with
validated by cameras in the car to consumer solutions. That is spurring
avoid the risk of spoof data disrupting the use of standards to bring these
the vehicle. The next stage is for a together as non-competing functions
different between the US and Europe. driverless vehicle to be able to take in the car. In future, standards relevant
This has been addressed though with data from any of the V2V networks as to automotive infotainment will emerge
software-defined radio (SDR) strategies well as the V2I infrastructure of the road, from traditional routes such as the
and the latest optimised processors from whether it’s public or private. All this AUTOSAR software standard and as
companies such as Cadence Design requires significant, detailed technology code from open community initiatives.
System’s Tensilica microprocessor agreements, says Smethurst. The challenge is how to enable direct
division (see Platform one). interoperability between automotive
However, at this stage the car makers The Genivi project standards and open source ‘standards’
have different ways of handling the data Taking on the role of promoting while still having the performance,
packets, which are encrypted for security. technology collaboration is perhaps functionality and security that is vital for
This is essential to prevent hacking of a natural step for Smethurst, as a UK driverless systems.
driverless cars, sending spoof data to automotive engineer working in the But the technology collaboration goes
perhaps re-route vehicles to a different German engineering heartland of further than that, Smethurst says. Maps,
destination or even create a crash. Munich. During 11 years with Rover for example, are a key technology for
Security and safety are therefore key Group in the UK he was responsible a driverless system, so the accuracy of
issues for the developers, and have to be for delivering many diverse projects, the map data becomes correspondingly
managed throughout the system. Some ranging from embedded electronics more important. There are plenty
manufacturers want to decrypt all the to information technology solutions, of stories about GPS systems

Unmanned Systems Technology | Spring 2015 15

In conversation | Graham Smethurst

guiding drivers down lanes too small

for the vehicle, or into rivers, and while Graham Smethurst
luckily many of these problems can be
overcome by linking into the local sensor
networks, GPS systems are not infallible. Smethurst began his career in telecoms, in the UK, transferring to the automotive
The map data also has to be transmitted industry in 1989. During 11 years with Rover Group in the UK he was responsible
to the vehicle, and the protocols and for the successful delivery of projects ranging from embedded electronics to IT
usage of the network agreed. In fact the solutions.
problem extends even further, into a From March 1998 to 2000 he was head of vehicle information systems at Rover
building say, where GPS navigation will Group. In 2000 he transferred to BMW in Munich to lead the development of the
not even work. Car Communication Computer, which was launched in the company’s 5 Series in
2003. Between 2003 and 2006 he led the team that defined the infotainment system
Issues with self-parking launched in the new 7 Series in 2008.
One predicted use of driverless He is currently chairman of the Genivi Alliance, formed in March 2009, and over
technology is the self-parking mode. This the past two years he has been the BMW Group representative in an industry
allows a user to drive (or be delivered) to collaboration investigating the feasibility of alternative approaches to delivering
a location, get out and have the car find automotive infotainment solutions. The target of the collaboration is a method of
a pre-arranged, prepaid parking space ensuring the sustainable delivery of consumer-based infotainment functionality
somewhere else. This avoids some of within the technical and commercial constraints of the automotive industry.
the problems with finding a parking
space and makes more efficient use of
city-wide resources.
If a car goes off to an underground underground, and the car may also have BMW’s collaboration here is with
garage, however, then the GPS stops to pay for the parking. perhaps a surprising partner: Baidu, the
working. The car has to find a parking All of this is a major challenge, but it is Chinese web search engine equivalent
space but still know where it is (to let being overcome by a form of self-parking of Google. Baidu has bought Chinese
the owner know where it has parked, technology, BMW’s Remote Valet Parking mapping technology company RITU, as
for example). It also has to be able to Assistant, also demonstrated at CES in well as investing $10m in US company
receive an authenticated command to an i3 research electric vehicle, which Indoor Atlas, whose technology uses
pick up the owner at a particular location, uses a laser scanner and a digital site a magnetometer in a smartphone and
which may also be in a building or map to replace the GPS signal. natural variations in the geomagnetic
field to navigate with an accuracy of a
metre, even indoors.
BMW is solving the challenge of
underground parking by collaborating
with other technology companies Conclusion
We are seeing a clear window for
collaborative technology in autonomous
systems, says Smethurst. With the
technology for autonomous operation
being demonstrated on the road, the
challenge now becomes the wider
infrastructure to support the operation of
these vehicles.
Both the V2V and V2I technologies
will play a key role in this, and these
require technology alliances across
a wide range of companies and
organisations around the world. He is
certain that it is no longer possible for
one car maker to do everything – to
make the driverless car work, he says,
we need to work together.

16 Spring 2015 | Unmanned Systems Technology

Innovative Lightweight Power
AIE Wankel-rotary engines take the heat like no others do.
Specifically engineered for small or tactical unmanned
vehicles, the 40S, 125CS, and 225CS deliver unrivalled
performance for demanding and long endurance operations.
With an industry-redefining power-to-weight ratio, AIE brings
you the future in UAS propulsion.
• Patented Compact SPARCS liquid cooling
• Industry leading power-to-weight ratio
• Heavy-fuel operation
• ISO9001 compliance
• Increased TBO
• No wet oil loss
• Low vibration

Visit us at AUVSI Atlanta, DSEI London,

and the Commercial UAV Show Singapore!
 An SA-400 in flight – note optional side tank

Day of the Jackal

Ian Bamsey travels to Scion UAS in Colorado to
get the low-down on a helicopter for which the pilot
is optional, the SA-400 Jackal

S
cion UAS, based near The SA-400’s origins be a completely new project, and by
Denver in Colorado and co- In 2012, Scion UAS responded to an offer comparison would have a length of
located with sister company for tender by the United States Naval 232 in (589 cm) and a rotor diameter
Scion Aviation, is developing Research Laboratory (NRL), calling for of 250 in (635 cm). It would, however,
a very special helicopter, the an unmanned rotorcraft to be used in exploit the same advanced autonomous
SA-400 Jackal, which can accommodate its evaluation of specific payloads under operation technology initially developed
one person although they can opt not to certain arduous conditions. For this for the compact SA-200.
fly it themselves. In fact, that person can purpose the helicopter the company then In the end the SA-400 proposal won
be merely a passenger, with the craft had under development, the SA-200 the NRL contract. That was at the end
operating in fully autonomous mode. Weasel, with a length of 89.2 in (227 cm) of September 2012, and the deal was to
Alternatively, it can fly without anyone on and a rotor diameter of 81.5 in (207 cm), supply a pair of SA-400s with the option
board. Moreover, even flying unmanned, wouldn’t be large enough. More carrying for a third after satisfactory trials of the
it can operate from a ship, regardless of capability was called for. prototype. In the light of this, the SA-200
the sea causing the deck to pitch, yaw The far more powerful SA-400 Jackal project was put on the back burner,
and heave. that Scion offered for tender would just after its first ground tests had been

18 Spring 2015 | Unmanned Systems Technology

 Scion UAS SA-400 optionally piloted helicopter | Dossier

Detail of the SA-400 during assembly, showing the tail

rotor drive shaft, the main transmission, pulley with one-
way clutch, dual alternators and main (Kevlar) drive belts

carried out, although Scion UAS CEO Currently the two initial SA-400s are platform. The first vehicle was delivered
Steen Mogensen emphasises that it will flying under an ‘Experimental’ FAA to the NRL immediately after this
be reactivated this summer, once the classification. That is an airworthiness successful demonstration.
NRL contract has been completed. certificate granted for a specific r&d craft, Flight trials of the second SA-400
The fact that the SA-400 optionally accepting that there hasn’t been the began in early December 2014, the time
carries a person and that, if suitably likes of crash testing but acknowledging of UST’s visit, with delivery scheduled for
qualified, they can become its pilot that there has been an FAA inspection early 2015. Scion UAS is offering both
at the flick of a switch has been very of it. Under an Experimental licence the the SA-400 and the SA-200 on the open
useful for development. Not only can craft cannot be used for commercial market, and in due course it also plans
they concentrate on monitoring an operations. It can be flown for r&d smaller as well as larger rotorcraft in its
autonomous flight, their presence in the purposes, initially within defined product line-up.
cockpit means the FAA considers it as a geographic boundaries, but after a
manned craft from the perspective of its certain number of hours have been The design brief
operational regulations, which are much logged then anywhere nationwide. The SA-400 was designed to meet a
tougher for unmanned craft. Given that a The capabilities of the SA-400 without certain payload capability requirement of
pilot is on board, there isn’t the restriction a pilot on board were demonstrated the NRL but was not otherwise tailored
on the use of commercial airspace in September 2014. This exercise, to a specific application; it is suitable for
that there would be in the case of an conducted in restricted airspace, multiple uses. The NRL requirement
unmanned craft. The ability to ‘See and included autonomous take-off and was that it must be capable of carrying
Avoid’ is satisfied by the onboard pilot. landing using a (truck-towed) mobile a 100 lb payload for four hours’ flying

Unmanned Systems Technology | Spring 2015 19

Dossier | Scion UAS SA-400 optionally piloted helicopter

The SA-400 undergoing flight trials

an autopilot and servomotor-based
controls in parallel with the standard
controls so it could evaluate its potential
as an unmanned craft, and it heavily
instrumented the craft for the purpose of
data gathering. In this form it made its
maiden flight on December 23, 2012.
While Mogensen notes that most of the
SA-400’s components are upgrades from
the Helicycle base, he adds that the re-
engineering of it has involved “changing
90% of the parts. Looking back, it might
seem that it would have been better for us
to have started with a clean sheet of paper,
but that wouldn’t have been possible within
the timescale, especially as we needed a
proven base to win the bid.”
Mogensen remarks that Scion UAS has
developed a good working relationship
with Eagle R&D, from which it directly
time. Moreover, it had to be capable buys Helicycle kits. As an example,
of taking off from and landing on a Eagle R&D didn’t have a complete
seaborne platform in conditions up to sea We had to make CAD model of the craft and Scion UAS
state three, which implies wave heights needed one for its development of the
up to 1.25 m (49 in). “That means landing it fly sideways SA-400, so with agreement from Eagle
on a pitching, rolling, heaving platform,” R&D, Scion UAS reverse-engineered the
notes Mogensen. “In addition, we had to alongside the ship missing components of the Helicycle in
make it fly sideways alongside the ship SolidWorks and has now shared that data
in a headwind and be controlled by the in a headwind with Eagle R&D in an ongoing spirit of
payload. Being able to do all of that is cooperation between the two companies.
part of the trick!” and be controlled Scion UAS chairman Jim Sampson
The brief from the NRL imposed a
very tight timescale for delivery of the first by the payload. adds, “We were taking an aircraft
that typically flies at 850-900 lb gross
vehicle. Mogensen says that ruled out a
clean-sheet-of-paper design. Instead, to
Being able to do weight and making it fly at 1200 lb. To
do that, everything had to be analysed
win the bid it was necessary to re-engineer
an existing and proven rotorcraft.
all of that was for strength, robustness and durability.
We strengthened the frame, and in the
Having studied the market, Scion UAS
identified Eagle R&D’s Helicycle model
part of the trick! moving parts we upgraded the bearings:
we got rid of as many grease bearings
as being ideal as a base for the SA-400. as we could and went to self-lubricated
This is a one-man helicopter with a single versions such as Teflon-lined stainless
main rotor, powered by a turbine engine. steel – bearings that work in a wet or
Sold as a kit for assembly by the customer, been logged by Helicycle owners, and gritty environment.
it was designed to be as affordable as when it came to analysing the Scion UAS “That process of upgrading went on
an SUV. With its 90 bhp power unit, as bid, the NRL was impressed with the throughout the aircraft, so there are some
standard it cruises at 95 mph and has an integrity of the Helicycle as the platform fairly significant changes to the basic
11,000 ft operational ceiling. for the SA-400 Jackal. mechanisms. Then of course we had to
Designed by Buford John (‘BJ’) Having won the contract in September allow for the servo motor integration into
Schramm, the Helicycle has been for sale 2012, Scion UAS quickly took delivery of the hardware components, so that we can
since 1997, and more than 70 examples a Helicycle kit and built it up essentially actuate them from our control system. Also,
have been completed and flown. Since as a standard product to establish a there is an entirely new carbon chassis,
1997, more than 5000 flying hours have baseline for development. It also added with the cabin and the fuel tanks.”

20 Spring 2015 | Unmanned Systems Technology

The engine UAV,” remarks Sampson. “All the sensors,
One of the attractions of the Helicycle for including the inertial sensors for the
Scion was its use of a turbine rather than In a turbine there autopilot, are very sensitive to vibration.
a reciprocating or rotary engine. Given The turbine doesn’t need engine cooling,
the operational requirements of the NRL is a lot less stress aside from some cooling for the oil.
and the fact that the craft would need That’s the beauty of it: all the waste heat
to use Navy-specification, kerosene- and a very low goes out of the exhaust pipe.”
based JP5/JP8/Jet-A heavy fuel (which It is worth noting that the turbine
is very difficult to ignite compared with
other liquid fuels) the engine options,
part count, which in question is of the turboshaft type,
meaning it is optimised to provide power
given what was available on the
market, narrowed down to four-stroke
helps make it to an integrated output shaft rather than
jet thrust. Sampson emphasises that
compression-ignition (CI) or turbine. Of
those the turbine was the lighter solution
extremely reliable in any turboshaft unit, as much of the
energy of the hot gas as possible is
– indeed, Sampson reckons that a CI
engine of comparable power output
and it requires extracted as useful shaft power.
“A jet engine extracts only sufficient
would weigh at least twice as much.
“In choosing to use a turbine we gave
little servicing energy out of the turbine wheel to power
the [charge] air compressor; the rest is
up some fuel efficiency,” Sampson nozzled and blown out of the back to
admits, “but what we gained is that in create thrust for propulsion. In a turboshaft
the turbine all the parts are moving engine we extract as much of the energy
in the same direction – they just spin, of that exhaust stream as we can as torque
they aren’t trying to beat themselves to make a turbine extremely reliable, and to drive the rotor. It comes out as rotation
death, like they do in a reciprocating it requires little servicing relative to a of a mechanically linked shaft. As much
engine. In a turbine there is a lot less reciprocating engine.” energy as we can get is extracted using
stress, there is a very low part count Clearly a turbine also runs inherently that output shaft, and there is not much
and most of the parts don’t touch, aside smoother than any reciprocating engine. thrust left in the exhaust.”
from a couple of bearings. That helps “Smooth running is really important for a The lack of cooling requirement

During proving trials the SA-400 took off and landed

from a towed platform while in autonomous mode

Unmanned Systems Technology | Spring 2015 21

The SA-400’s Solar T62-32
engine awaiting installation with a lighter production by Eagle
R&D, made possible since the power
output is de-rated from 160 to 90 bhp.
“A turbine operates like a naturally
aspirated engine: if you want to have
power at altitude on a hot day, you
can’t get much more than 100 bhp out
of it due to the altitude and the heat,”
explains Sampson. “In view of that, we
de-rate it to where we are going to be
operating, which is an 8000 ft maximum
[beyond which the air is too thin]. That
is standard practice for turbines. You
can’t supercharge a turbine; you have to
oversize it and then de-rate it.”
In the SA-400, the T62 is fitted with a
digital governor and a number of sensors
for temperature, pressure and fuel flow.
There is no throttling of a turbine’s charge
eases installation and improves overall air. “Turbines run ultra-lean,” notes
craft performance. It is no coincidence Sampson. “There is always more air
that these days almost all manned If I fly manually going through a turbine than is needed
military and commercial helicopters for a stoichiometric combustion. There is
are turbine powered. Turbines are not and aggressively, constant combustion and you control a
well suited for applications that involve turbine by how much fuel you give it.”
a lot of stopping and starting, such as it still only The fuel supply to the T-62
cars, but they are very well suited to incorporates an engine-driven gear
high-power constant-load applications. changes speed pump that creates constant pressure,
“Helicopters operate at high power all beyond which is a metering valve: a
of the time,” observes Sampson. “You momentarily by needle valve that in the case of the SA-
are hanging on your horsepower –
helicopters don’t glide well at all! about 100 rpm 400 is operated by a servomotor, which
in turn is controlled by the governor.
“An equivalent gas engine would be
far heavier,” he emphasises. “Put a diesel
out of 61,000. Sampson notes that while the engine
is running, the needle valve is always
in there and you wouldn’t lift it off the
ground.”
That is moving, feeding half-a-dozen constant
flow (fixed orifice) injectors via a common
Eagle R&D designed the Helicycle
around the Solar Turbines T-62 Titan gas
remarkable! fuel rail, so it is the only control over the
amount of fuel supplied to the engine.
turbine, as used as an auxiliary power “We run at optimum engine speed
plant in the Sikorsky Super Stallion all of the time: 61,000 rpm,” remarks
and Chinook military helicopters; these Sampson. “We have three speed settings
days Solar Turbines is a wholly owned addition of a fuelling stage). This is a – zero, idle and then, once we have
subsidiary of Caterpillar. Scion UAS so-called single-spool turbine, meaning engaged the rotors, flight speed. Once
uses reconditioned T-62s since they are the compressor and turbine wheel are we have engaged the rotors it stays at
perfectly acceptable for its needs, making directly connected to the output shaft. the 61,000 rpm flight speed. If a rotor is
them cost-effective. The assembly rotates at 61,000 rpm and pitched, that loads the engine, and with
The T-62 uses a centrifugal charge the bearings are lubricated by a very fine that the servomotor moves a lever so that
air compressor, while its power section oil mist. This is a dry-sump oil system; the needle supplies more fuel. We call
– having introduced the fuel supply there is a gear-type oil pump driven off that throttle-by-wire.”
and then ignited the mixture – uses the gearbox. Flight speed is adjusted by manipulating
a radial inflow impeller (echoing a The T-62 comes complete with a the rotors, so it follows that the turbine
typical automotive turbocharger with the gearbox, but Scion UAS replaces that can be kept at constant revs at all times.

22 Spring 2015 | Unmanned Systems Technology

 Scion UAS SA-400 optionally piloted helicopter | Dossier

“You just vary how much fuel you flow in, shafts are connected via drive belts. As
but you are always flowing some fuel into well as sending power back to the tail
the engine,” remarks Sampson. “When When we put rotor, the belt-powered driveshaft sends
we put pitch into the rotor blades the the drive forward and up to the main
turbine is in danger of losing rpm; the pitch into the rotor via a 90º bevel drive.
governor sees that and gives it more fuel In the T-62 the drive to the output shaft
to maintain 61,000 rpm. If I fly manually rotor blades the is reduced from 61,000 compressor/
and aggressively, it still only changes turbine wheel speed to 3000 rpm via
speed momentarily by about 100 rpm out
of 61,000 rpm. That is remarkable!”
turbine risks gearing. The belts then drive the main
transmission and the tail rotor at 3000
Moreover, the vibration level is so low
that the turbine can be hard-mounted
losing rpm, but rpm. The bevel drive has a fixed ratio
of 4.8:1 so that the main rotor (since it
to the airframe. In fact, more vibration
comes from the rotors than the engine.
the governor is much bigger in diameter than the tail
rotor) runs at 620 rpm. Sampson adds,
“We do have anti-vibration mounts on
the electronics though, specifically for the
sees that and “The bevel drive runs in an oil bath; it
is positively lubricated, not splash, and
autopilot, and normally for the payload,”
says Sampson.
gives it more fuel we have an oil cooler to maintain oil
temperature in that pressurised system.”
Some gears are transferred from the
The transmission original bevel drive gearbox, others are
The T-62 is centrally mounted in the new in the Eagle R&D production. The
SA-400 with its output shaft horizontal, various gearshafts run in a combination
underneath and in parallel with a shaft a shaft coaxial with the tail rotor shaft. of ball and taper-roller bearings.
taking the drive back to the tail rotor. Both The driveshaft runs above the forward- Despite the fact that for start-up a
the tail and main rotors are driven from projecting T-62 output shaft, and both single-spool turbine needs all load

The SA-400 has simple landing gear but those skids house
load sensors that are crucial to its autonomous operation

Unmanned Systems Technology | Spring 2015 23

Dossier | Scion UAS SA-400 optionally piloted helicopter

For the SA-400, Scion UAS designed

its own tail rotor mechanism
SA-400 general
specification

Craft height: 82.1 in

Craft width: 51.5 in
Craft length: 232 in
Rotor diameter: 250 in
Empty weight: 700 lb
Gross weight including
full tank and payload: 1200 lb
Turbine length: 33.4 in
Turbine diameter: 21.4 in
Turbine weight: 142 lb dry
Maximum speed: 95 knots
Cruising speed: 82 knots
Mission duration: in excess
of four hours.

taken off it, there is no conventional clutch An overrun clutch in the transmission Some key component
in the transmission. Instead the engine allows the rotor to continue spinning suppliers to the SA-400
is literally tilted down from a neutral if the power stops. This is purely an
start-up position so that its output shaft emergency device, to offer the chance of Base helicopter, turbine & gears:
tensions the belts and thus creates drive a safe touchdown in case of unplanned Eagle R&D
to the transmission. This tilt mechanism critical power loss. Autopilot: Adaptive Flight
operates at the front of the engine, and an At the moment Scion is using the main Servomotors: Moog
over-centre approach allows it to lock into rotor blades supplied with the Helicycle Servomotors: Volz
a position whereby once engaged there kit, which is formed as a bonded GPS: Novatel
is constant tension on the belts. aluminium structure – in effect a folded Alternators: B&C Specialty Products
“The use of a conventional clutch sheet with a spar running inside the Battery: Odyssey Batteries
would be heavier,” notes Scion UAS’ leading edge and other reinforcements. Fuel pump: Facet
director of mechanical engineering Scott This is given special treatment to guard Instrumentation: MGL Avionics
Lowry. “Since we have the [vertical] offset against corrosion and abrasion. Communications systems:
between the turbine output shaft and the Clearly, as a rotor spins, the airspeed Persistent Systems
transmission we can get two jobs done as measured along the length of each Navigation lights: Aeroleds
with the belts.” blade increases with distance from Design software: CATIA
A small electric motor, which in the hub. As a consequence it can be Design software: SolidWorks
the Helicycle is activated by a button advantageous to alter a blade’s profile CNC machines: Haas
pressed by the pilot, causes the engine along its length. That is not the case CNC machines: Zimmermann
to snap into its drive position. In the SA- though with the SA-400’s off-the-shelf Pre-preg: Cytec
400 that function is integrated with the production, so Scion UAS is looking to Resin systems: MTM46
autonomous control system. “We have develop a more sophisticated design
the computer check that the belt isn’t that exploits such advanced tailoring as
tensioned,” remarks Dr Phillip Jones, it moves in due course to a composite
Scion UAS’ director of flight control and rotor blade that will be manufactured wider and more efficient – so both
software development, “then it goes in-house. the rotor blades and the tail rotor
through all the process of starting the The stock tail rotor has already hub mechanism were redesigned.
engine, then engages the belt drive and been replaced by Scion UAS’ own For the time being, the production is
then revs up the engine to flight rpm: it production. In effect, for a sideways still aluminium skins over steel spars,
is all automated. Then comes the next flight requirement by the NRL it had whereas in due course there will be a
command, to take off.” to be beefed up – making the blade carbon fibre composite tail rotor.

24 Spring 2015 | Unmanned Systems Technology

JACKAL’S Increase the Distance of
ADAPTIVE FLIGHT your Data Link Over 10X

Linear Bi-Directional RF Amplifiers

When Scion UAS looked for a flight control system
• In Bands from 400 MHz to 6 GHz
for the SA-400, they selected the FCS-20 from
• TX Powers from 1W to 50W
Adaptive Flight. Why? The answer is simple: • 10 dB RX Gain
Capabilities, Flexibility, Price and Support! • Manual or Auto Switching between TX and RX
• Less than 2 μsec switching speed
Highlights: • CPLD Control
• Helicopter and Fixed-Wing Flight Control • Designed to work with all radios including
Silvus, Microhard and Freewave
• DSP + FPGA for flexibility
• Ethernet Port
• High-fidelity simulator
• Easy to Use, Feature-rich Interface

180 Tices Lane Building A Suite 107 East Brunswick NJ 08816

Ph. 678 770 3155 [Link] sales@ [Link] • [Link]
 Scion Aviation
Titanium doesn’t
Now 20 years old, Scion Aviation moved to its current location near Fort Collins, corrode in salt
about an hour north of Denver, Colorado, and almost in the shadow of the Rocky
Mountains, in 2001. It carries out general aviation engineering, including r&d on water, which
behalf of a number of clients. Based close to the small Loveland-Fort Collins airport,
Scion has its own helipad outside its main assembly hanger.
Scion Aviation and Scion Unmanned Aerial Systems (UAS) are separate
means we will
companies that share the same 22,000 sq ft building, which Aviation owns. UAS
rents space and the use of plant and labour as necessary from Aviation. In total,
not have to carry
14 people work across the two companies. The facilities of Aviation include an
aerospace-grade advanced composites construction capability, including complete
out secondary
airframes, which is exploited by some major players in the aerospace world.
Aviation has design as well as manufacturing capabilities. It has CATIA and
treatments such
SolidWorks software and uses CAM as well as CAD. It also has an ISO 9 clean
room for hand lay-up and the facility for vacuum bagging. It produces composites
as anti-corrosion
using pre-preg carbon fibre and small and large autoclaves, the latter being 8.5 ft in
diameter and 50 ft long – large enough to accommodate a Learjet fuselage.
Tooling is mostly MDF wood- or epoxy-based and is five-axis CNC machined
using a 16 ft stroke water-cooled mill with 1 thou precision. It also offers CNC
pattern-making using an infinite-length cutting table. Other capabilities include resistance in the marine environment as
welding, finishing and assembly. It works with customers manufacturing both fixed- well as saving weight. And titanium works
wing and rotor-based craft. well with carbon fibre.
Wind tunnel testing is not seen as being as critical for a rotorcraft as it would be “Carbon fibre is conductive, and when
for a fixed-wing design. For small rotor blade development Scion UAS has a whirl in direct contact with metal of any kind
tower that runs a rotor in the open air and uses load sensors to measure its lifting it creates a galvanic reaction, so you
capability. However, it isn’t feasible to simulate in any test cell the centrifugal forces get galvanic corrosion – the metal, not
to which the entire helicopter is subject during flight, so actual operation is the key the carbon, will corrode. Aluminium and
to development. In addition though, Aviation has a CFD capability that can assist steel will corrode quite rapidly, whereas
rotorcraft development. titanium is electrically active enough that
it and the carbon can be directly bonded
together without any problem.
“Also, titanium doesn’t corrode in salt
The airframe is a lot heavier than the original because water. A titanium frame means we will
The SA-400 is based on a straightforward there is a lot more to it.” not have to add secondary operations
tubular spaceframe, which does not have The Scion UAS version of the frame such as anti-corrosion treatment. We can
structural sheet bulkheads integrated is sandblasted, etched, primed and TIG weld titanium as well as chromoly
into it. It is made from 4130 chromoly, finished with an epoxy paint. Indeed, in-house, and we figure titanium can
the standard tubular steel for aircraft given that the initial SA-400 commission save about 30 lb. We will make further
structures. A modification of the stock is specifically for use in a marine refinements to the frame design when we
Helicycle frame, it has wall thickness environment it was considered crucial bring manufacture in-house for the next-
increased where necessary and that all bare metal in all parts of the craft gen version of the SA-400.”
additional bracing in places. should be covered. In places nickel
“There were a number of spots where plating is used – that is another in-house Fuel system and electronics
we found we needed to reinforce the capability of Scion UAS. The SA-400’s fuel system, including the
frame,” remarks Lowry. “We also had “As we go forward we will be building main tank, was re-engineered from the
to make some significant changes to a custom spaceframe rather than Helicycle base. Fuel is carried in the
it to allow for payload integration and modifying the Helicycle spaceframe,” primary tank as well as optional side
mounting, and to allow for the servos to confirms Sampson. “We are looking at tanks. With or without the side tanks the
be integrated into the aircraft. Our frame making it out of titanium, for corrosion fuel mass is centred on the craft’s centre

26 Spring 2015 | Unmanned Systems Technology

 Scion UAS SA-400 optionally piloted helicopter | Dossier

The SA-400 has simple bodywork

and leaves its tail unshrouded

of gravity so that its dynamic behaviour is it ensures positive head pressure to the
unaffected by diminishing fuel. engine-driven pump, Sampson explains.
“On the original Helicycle the primary We had to go to Ahead of the main tank the carbon
tank fits within the main frame tower,” fibre composite cabin floor and integral
observes Lowry. “We had other things a completely new seat backrest tie into the spaceframe. The
we also needed to fit in that space so tank attaches to the back of the cabin
we had to go to a completely new fuel fuel tank set-up, floor and transfers weight through it to the
tank set-up, with a new central tank spaceframe, but most of the fuel weight
that really hugs around the engine and with a new central is transferred through the side payload
the control system and landing gear, to mounts to the spaceframe, Lowry reports.
maximise the use of space. That was a tank that really The requirement from the NRL
real design challenge.”
Since Scion UAS couldn’t move the hugs around the included that the electrical system would
be 24 V (28 V), consistent with standard
fuel load away from the centre of gravity,
it had to expand the tank into every
engine and the naval vessel power systems. Since
the base Helicycle vehicle has a 12 V
available adjacent nook and cranny.
The consequently very complex design,
control system electrical system, a number of items had
to be changed.
which expands around the engine
and control system and landing gear
and landing gear The SA-400 power system has two
alternators with separate control units.
zones, was produced from carbon fibre Each power system, one of which is
composite (the simpler Helicycle tank dedicated to the payload, has its own set
is moulded plastic). It accepts 17 US of batteries.
gallons (64.4 litres), versus 16 for the are fitted with internal baffles. There is Mogensen says, “With a separate
standard item. an electric booster pump between the payload power system, if anything
The complex form of the main tank tanks and the mechanical fuel pump electrical goes wrong on the payload
(evident from the exterior) helps dampen (with filtering of the fuel either side of it). side it will not affect the aircraft power
slosh, whereas the optional saddle tanks The engine will run without that pump but system. Flying without a pilot on

Unmanned Systems Technology | Spring 2015 27

Dossier | Scion UAS SA-400 optionally piloted helicopter

The SA-400 is capable of a very wide range of

missions, from agriculture to naval defence

board, we are highly dependent on the Consider a situation in which a primary that frequency is not authorised for such
power system.” transmitter can reach a certain device but use in the EU, so there is an alternative
Standard aircraft navigation lights not a more distant one, over the horizon. available for vehicles being exported
are used. The primary communications It is conceivable that if the transmitter there). Thus there are two separate
between ground station and the craft is were positioned at the location of the first radios on board the craft; two separate
an off-the-shelf 2.4 GHz Wave Relay wi-fi device, it could then reach the further links between it and the ground station.
system supplied by Persistent Systems, device. That is where MANET comes Moreover, the SA-400 is designed
and is a Mobile Ad Hoc Networking in, with the first device automatically re- to allow for easy replacement of the
(MANET) solution originally developed transmitting the signal to the second. standard communications radios, either
for military use. The basic concept is that This infrastructure-less Wave Relay mesh with other civilian radios or military radio
each device in the network, in addition to network is self-forming and self-healing. systems.
its regular functioning, acts as a router, As a back-up there is a more
forwarding traffic unrelated to its own use. conventional 900 MHz wi-fi link (although SA-400 control
To control a helicopter, the main
rotor has blade pitch control which
Foreground left, Steen
Mogensen, centre, Jim is activated through an articulated
Sampson, right, Phillip swash plate that pivots on its hub to
Jones with an SA-400 manipulate the individual rotor blades.
To control the direction of force,
front/rear/left/right, the blades are
manipulated individually depending
on their position through the circle of
rotation. The input for this action comes
from the Cyclic control, which tilts the
swash plate in the desired direction.
To control the lifting force of the rotor,
the blades are manipulated together by
moving the swash plate up or down,
which will increase or decrease lift
throughout the entire rotation. This is
accomplished with the Collective control.

28 Spring 2015 | Unmanned Systems Technology

 fuel valve used to provide additional fuel
Scion UAS during engine start-up.
Jones notes that during flight the
onboard autopilot is a key element of
Having worked together in 2009-10 as subcontractors on an IED Detection UAV control. “It senses how the helicopter
project known as the Yellow Jacket, the five founding members decided to form moves in space and what its attitude is
Scion UAS. “We realised that together we had critical-mass capabilities and at any given moment using its internal
experience to design and build highly sophisticated autonomous helicopters,” says sensors – accelerometers and gyros.
Scion UAS CEO Steen Mogensen. “Taking advantage of the already established It can detect any movement in any
company, Scion Aviation, and its world-class facility in Fort Collins, Colorado, it was direction at any given time, which allows
decided to locate the new company there.” it to stabilise the helicopter.
Scion Aviation is owned by Jim and Betsy Sampson. They are also joint owners of “If a gust of wind upsets the helicopter
Scion UAS together with long-term Aviation employee, director of operations Scott then the autopilot will issue commands to
Lowry, plus Phillip Jones and Steen Mogensen (all five holding an equal share). bring it back to where it needs to be. But
Jones and Mogensen both joined Scion UAS from autopilot supplier Adaptive Flight, of course, it is one thing if it is hovering at
bringing very valuable knowledge of flight control technology – the heart of any the time; another if it is moving forward.
unmanned system the company might devise. Then there is flying sideways – which is
Jim Sampson is chairman of the board of Scion UAS. He is an engineer like balancing on a ball!”
specialising in all mechanical aspects of aviation, including the design and The SA-400 has an autopilot
production of metallic and advanced composite components. He is also an expert made by Adaptive Flight, which was
in aircraft engines and is a licensed pilot for fixed-wing and rotorcraft. Betsy established to commercialise technology
Sampson runs the financial side of both Scion Aviation and Scion UAS. developed at the Georgia Institute of
Steen Mogensen is the chief executive officer and technical director of Scion Technology (‘Georgia Tech’) in Atlanta.
UAS, and is an engineer specialising in electronics. Dr Phillip Jones is Scion UAS’ The autopilot itself was developed
director of flight control and software development, and as such is responsible for under the leadership of Professor Eric
all software-related items. Scott Lowry is its director of mechanical engineering, and Johnson, the Lockheed Martin Associate
oversees design using CAD and CAM, the in-house facilities of Scion Aviation. Professor of Avionics Integration in
Mogensen remarks that, aside from its internal expertise, Scion UAS is assisted the School of Aerospace Engineering.
by several subject matter experts in the fields of helicopter design, aircraft structure Scion UAS continues to work with
analysis and aerodynamics. He adds that, until his passing in October 2014, Scion Adaptive Flight and with Prof Johnson
UAS consulted with Ray Prouty, widely acknowledged as one of the pioneers on development of the autopilot for the
of the modern rotorcraft industry and author of seminal works on helicopter SA-400, in particular for the advanced
aerodynamics. control necessary to land in the unstable
environment of a sea-going vessel.
“The autopilot is detecting the actual
attitude and comparing it with what it
In the cockpit of a manned helicopter controls can be operated by servomotors should be and keeping that balance,”
there is a lever for each of these two rather than the physical actions of a pilot. Jones says. “We have an additional
functions, Cyclic and Collective. In the Four servomotors grouped on a tray are computer on board, which controls
SA-400 these levers are optional, along connected to the mechanical linkages, the SA-400 itself. This ‘flight controller’
with a centre console that holds the allowing for easy conversion from interfaces through the servos and is
required instruments and switches for a manned to unmanned operation. getting input from the autopilot. So the
pilot to operate the helicopter manually. In addition, there is the aforementioned autopilot tells the flight computer to, say,
The tail rotor is necessary to control fuel control servo which acts on a move right and the flight controller knows
the yaw of the craft; it serves as an anti- closed-loop basis. It is controlled in how to get that done. It sends appropriate
torque device, counteracting the main response to the signal from an rpm signals to the appropriate servos.
rotor’s rotational force, and is also used sensor on the engine output shaft, so as “The flight controller isn’t navigating but
to control the heading of the helicopter. to keep that shaft running at constant is, for example, handling functions such
In a conventional manned helicopter its speed. That is the only control over the as babysitting the engine for temperatures
blade pitch angle is controlled, again engine, aside from a solenoid valve to and pressures. In fact, a good example
mechanically, by the pilot using foot cut the fuel supply in an emergency and is starting the engine, in which case
pedals. In the SA-400 all those rotor an additional electronically operated all those functions are handled by

Unmanned Systems Technology | Spring 2015 29

the flight controller. We tell it to start the
engine and it starts the engine – that has The SA-200 Weasel
nothing to do with the autopilot.”
It follows that core control of the SA-
400 is concentrated in two modules Prior to the establishment of Scion UAS in 2011, the founders started the design of
– an autopilot and a flight controller. a small rotorcraft which became known as the SA-200 Weasel. Further developed
In addition, says Jones, “The control by Scion UAS following the formation of the company, this design was based on
electronics include sensor interface feedback received from the US Air Force after the completion of the Yellow Jacket
boards and things like that. We run a project in 2010.
lot of Ethernet between different sensor The SA-200 is a general-purpose unmanned rotorcraft which is sized such that it
types and different systems. Then there is can be transported in the back of a pickup truck, and is designed to be launched
an advanced GPS and there are the two from its carrier within 10 minutes. It is envisaged for roles such as pipeline and
radios to talk back down to the ground, a wind turbine inspection, mineral exploration, security and other tasks calling for a
primary link and a back-up. camera in the sky. Another possibility is the use of chemical sensors, for example
“For the flight controller we are using one that allows the identification of radiation on sea-going vessels or one that
a standard off-the-shelf single-board allows leaks to be found in a natural gas pipeline, through odour recognition.
computer, albeit one that has a lot of The SA-200 project went on the back burner when Scion UAS won the contract
I/O. Aside from its primary function of for the larger SA-400, but will be restarted this summer.
interfacing between all onboard
systems, the flight controller acts as
a data logger, although all the data
from the various onboard sensors is “So long as the radio link is open then back and replay everything that went on,
constantly sent to the ground station all the data is sent to the ground station. should we have an issue with anything.
in real time via the radio links. The At any time the operator at the ground But if at any time we lose the data link
onboard logger is simply a back-up station can monitor those inputs, and all between ground control and the aircraft,
for data logging. the data is logged there, so we can go that data is still being stored on board the
aircraft.”
The SA-400 is designed primarily to
operate in autonomous mode rather
than as a remotely piloted aircraft. In
the normal manner of autonomous
aerial craft, it will automatically follow
a predetermined course defined by
waypoints. While it can also be remotely
controlled within line of sight or via
onboard cameras in First Person View
style, Mogensen says it isn’t normally
operated by Scion UAS in that way.
Autonomous operation is no mean feat.
As Mogensen remarks, “The pilot can
see, sense, smell and so on, but once
you take him out of the equation you
actually lose about a million sensors and
a data processor.”
The SA-400’s landing gear is
essentially a bent tube on each side
of the craft. “But we do have load cells
there, to use as a weighing scale,” notes
Mogensen. “The autopilot needs to know
when the skids are off the ground. This
triggers a state change in the autopilot,
Start of another mission for the SA-400 – without pilot which is required since trying to fly

30 Spring 2015 | Unmanned Systems Technology

 Scion UAS SA-400 optionally piloted helicopter | Dossier

and the interactions between them;

when you then add another force onto
that, it will be different according to
which side it affects.
“And that all gets into some very
complex dynamics. If we have Jim in
the seat, he can feel that the helicopter
doesn’t really want to do something any
more, and he knows what to do to get
out of it. But how do you explain that to
a computer? You have to tell the flight
controller when to say ‘no’, and then
when it says ‘no’, what should happen.
That all amounts to an enormous amount
of programming.”

Potential uses of the SA-400

Interestingly, the SA-400’s autonomous
control system is sophisticated enough
to allow a payload to command the
autopilot. If, for example, the payload
were a military one in the form of an
The SA-400 can leave its pilot behind early detection system, identification of
a threat could lead to a given response.
The strategy might be that the SA-400
sideways when still on the ground would acts as a sacrificial decoy to protect a
be a disaster! convoy against an incoming missile, in
“Then the autopilot has to react to any The pilot can see, which case the payload would call for the
gust of wind as soon as it takes off. Then appropriate response of the autopilot.
think about the added complication of sense, smell and “The SA-400 control system can
the ground being the deck of a ship that even be integrated with those of sister
is moving at sea. As a pilot you just do so on, and once craft,” Mogensen says. “In the case of
what you need to do, but we have to get our missile scenario, there might be a
a computer to do it.” you take him out number of SA-400s acting in convoy. The
The NRL needed the SA-400 to be first to detect the threat might not be in as
able to fly sideways alongside a ship of the equation good a position as another to respond,
travelling at 30 knots into a 16 knot
headwind. In other words, effectively you actually lose so the integrated control system would
ensure that the one in the best position
it had to be able to fly sideways at 46
knots (53 mph/85 kph). “That is not
about a million takes the decoy action.”
However, Mogensen emphasises that
normal and not easy – a tall order,”
reflects Mogensen. “Jim [Sampson] can
sensors and a such military scenarios are no more than
an example of the sort of capability that
do it using manual control, having a lot
of experience, but there are some very
data processor the SA-400 has, and that the long-term
strategy is to also develop its potential
strange dynamics involved. It is hard to in the civilian world. For example, it
teach an autopilot to do that safely. could be used in agriculture, not only
“A helicopter doesn’t really want to fly to inspect crops but, by using the side
sideways; it is like a weathervane trying wind resistance and your tail rotor’s tanks, to spray them as required. Without
to go against the wind. The main rotor ability to overcome that is a factor. Then the fuel in the side tanks the flight time is
doesn’t care – it is all about having the there are some very strange helicopter still about 45 minutes to one hour.
rest of the structure pushing against dynamics, which has to do with which Many and varied are the potential uses
the wind. The balance between the way the tail and main rotors are turning of the very clever SA-400 Jackal.

Unmanned Systems Technology | Spring 2015 31

Digitally enhanced
Stay up to date with industry developments and
research suppliers on the all-new UST website

Get updates as they happen – find

us on Twitter and Facebook, and
join our LinkedIn discussion group!

@unmannedsystem

[Link]/unmannedsystem

Search “Unmanned Systems

Technology” in groups to find us.

32 Spring 2015 | Unmanned Systems Technology

 UST online

Video channel

Whether you want to examine the features of a new

product or technology, view a milestone event or find
out more about a supplier’s capabilities, the new video
channel on UST puts relevant unmanned systems video
content in one easy-to-browse location.

Aeryon SkyRanger VTOL

sUAS by Aeryon Labs

More technology
UST now features more suppliers and technology
categories than ever before, so it is now even easier
to find what you’re looking for. Each company profile
provides in-depth insight into their capabilities, with
images, videos, documents and white papers, as
well as clear contact information.

GNC1000 integrated guidance, navigation and

control system by Moog Crossbow

+ Add your company

If you supply equipment or services to the

unmanned systems sector, visit
[Link]/add
to create your supplier profile.

Phoenix 15 multi-rotor unmanned

aircraft system by UAV Solutions

• Cameras, Imaging & Vision Systems • Materials & Manufacturing • Navigation & Guidance Systems • Propulsion & Power
• Unmanned Vehicles & Platforms • Development Services & Facilities • Simulation & Training • Software • Robotics
• Electronics, Subsystems and Components • Integration Services • Hydraulic & Pneumatic Systems • Data & Communications
• Ground Control Systems and Equipment • System & Mechanical Components

Unmanned Systems Technology | Spring 2015 33

 Taking off
and far, representing many potential
With the market for commercial civilian users ranging from filmmakers to
UAVs set to soar, Ian Bamsey and local authorities and security specialists.
The show also saw the launch of
Nick Flaherty highlight some of the Unmanned Systems Technology

new technology on show at this trade fair magazine and put the spotlight on
innovative new technology in this

T
burgeoning field. UST writers were there,
he world is on the verge of exhibition centre in London on October walking the aisles to identify the most
explosive growth in the use of 21-22, was billed as “the first marketplace important developments at this busy and
UAVs for civilian applications, for commercial UAVs”. As such it intriguing event.
and the Commercial UAV attracted an international line-up of
Show, held at the Olympia exhibitors and drew in visitors from near Commercial UAV 2014
Less than three years old, Advanced
Innovative Engineering (AIE) offers
single-rotor Wankel-type engine
technology in a range of displacements
from 40 to 650 cc. Nathan Bailey told
UST that the company is scaling up
and down from its initial engine, the
225CSR, which as its name suggests
has a displacement of 225 cc; it first ran
in February 2014, and produces 40 bhp.
The 40 cc engine variant offers 5 bhp,
the 650 cc some 120 bhp, and the target
market includes automotive applications
as well as UAVs.
Bailey said AIE has taken rotary
engine performance a step forward in
terms of performance and durability,
thanks partly to a revolutionary cooling
technology called the Self Pressurised
Wankel-type engine from AIE Air Rotor Cooling System (SPARCS). He

34 Spring 2015 | Unmanned Systems Technology

 Commercial UAV Show | Report

Also seen at Olympia

Studies show that
adding hydrogen VulcanUAV has been working The LidarPod in action
with RouteScene, part of
to diesel fuel Mapix Technologies in
Edinburgh, since December
can significantly 2013 to develop a system

reduce the
optimised for Lidar 3D mapping
using RouteScene’s 2.5 kg

particulates LidarPods, we were told. These

use a Velodyne HDL32 Lidar

emitted by a scanner which has 32 lasers

and collects more than 700,000

compression distance measurements per second. The measurements are then integrated with a
highly accurate IMU and comprehensive GPS navigation system from RTK that uses

ignition engine additional GPS antennas on the UAV for ultra-accurate heading reference. There
is onboard data storage and a two-way data link for RTK GPS updates and data
verification.

Singular Aircraft’s
The unusual aspect of the Flyox 1 seapane
explained that in general the key to rotary UAV from Singular Aircraft,
engine durability is adequate cooling and Xavier Martinez said, is that it
lubrication of the rotor. “If the engine’s is a seaplane. He told us that
rotor is not cooled sufficiently then the company founder Louis Carrillo
engine’s lubrication will break down due was particularly concerned to
to the high rotor temperature, and the exploit the potential for UAVs
rotor bush and bearing will wear rapidly, to assist in fire fighting, where
reducing overall engine life.” the ability to land on water is an
AIE uses conventional liquid cooling obvious advantage.
for the rotor housing, and combines this The news at Olympia from the Barcelona-based company, which was established
with its patented SPARCS technology in 2011, was a new model of amphibious UAV. Martinez said that with a wingspan of
for the rotor. “SPARCS uses blow-by 14 m the twin-engined, fixed-wing Flyox 1 is the largest UAV outside the military. It
gases from the combustion process can carry a 2050 kg payload and has an endurance of more than 60 hours.
to pressurise the engine core,” Bailey
explained.
“This pressurised gas is then
recirculated within a sealed system by a outperform a battery in the context of a particulates emitted by a compression
fan, passing first through the rotor and UAV. At Olympia, Stephen Bennington ignition engine. In effect, it will burn
then, to reject heat, through an air-to- told us that it had recently signed a up nanoparticles that would otherwise
water heat exchanger integrated into the contract with French aerospace company escape through the particulate filter.
rotor housing.” Safran, to provide a back-up power While UAVs are not currently bound by
Wankel-type engines are popular in supply for manned aircraft. This is Cella emission regulations, they will doubtless
the UAV sphere, and it is worth noting Energy’s first revenue from a commercial come under the same scrutiny as
that such engines lend themselves to deal, and Bennington was pleased to modern road vehicles as the commercial
being fuelled by hydrogen rather than report that it gives further credibility to his market for them booms in due course.
petrol or diesel. company’s pioneering technology. Jon Skelton told us that Topcon
We highlighted Cella Energy’s Interestingly, Bennington also told Positioning provides a very lightweight
hydrogen storage pellet technology in the us that studies have shown that using GNSS OEM board for UAVs that offers
previous issue of UST, explaining how hydrogen as an additive in diesel an accuracy down to less than 5 cm
it can supply a fuel cell to collectively fuel can significantly reduce the in all directions. This, combined

Unmanned Systems Technology | Spring 2015 35

Report | Commercial UAV Show

Also seen at Olympia (continued)

A key feature of
the receiver is Aerosonde’s UAV

that it dispenses
with the need for
ground control
points, opening
up terrain where
ground access
is a problem

with the ability to take 100 readings per

second in combination with an onboard Textron Systems Unmanned Systems is now offering the Aerosonde Mk 4.7G
camera, enables precise mapping of the UAS platform. The company owns Aerosonde, which co-manufactures the 3.6 m
environment. A key feature of the Topcon wingspan small unmanned aircraft system, and the Mk 4.7G, is offered for sale or
receiver is that it dispenses with the need operated by Textron Systems on a fee-for-service basis around the world. Fellow
for ground control points, saving time Textron Systems business Lycoming Engines manufactures the purpose-built EL-
compared to conventional aerial systems 005 direct-injected, two-stroke heavy fuel engine for the craft.
and opening up terrain where ground Sean Baity, principal systems engineer, told us that the Mk 4.7G has attained
access is problematic. nearly 100,000 flying hours over the past two years, that it is runway-independent
The Topcon GNSS receiver exploits and is suitable for operations in harsh environments.
Real Time Kinematics (RTK), a differential
GNSS (Global Navigation Satellite
System) technique that provides high
positioning performance in the vicinity of via the hand-launched, battery-electric included in the package.
a base station. The technique is based propelled, fixed-wing Topcon Sirius The best solar cells are typically in
on the use of carrier measurements and UAV, in a package designed for a the high teens in terms of the efficiency
the transmission of corrections from wide range of applications. He said it with which they convert solar energy
the fixed-position base station to the operates successfully even in the rain, sunlight into electrical output, and
UAV, so that any errors from standalone at airspeeds of up to 60 kph. “The the best flexible solar cells are barely
positioning are cancelled out. challenge was to integrate the GNSS above 10% efficiency, Rich Kapusta of
In essence, the mapping process is receiver with the platform’s existing Alta Devices told us at Olympia. There
one of photogrammetry. Photographs hardware and software for repeatable his company was showing a new
of the terrain are overlapped, and accuracy,” he said. “We succeeded in cell technology which he said has an
shape-matching imagery analysis using achieving that repeatability regardless of efficiency of 28.8% and is lightweight,
advanced algorithms is combined with weather conditions, altitude and so on.” thin, and flexible.
the x-y-z coordinate data from the GPS The Sirius can fly for up to 55 “This is a world record for a
to create a 3D map of the surface over minutes capturing high-resolution commercially available solar cell,” Kapusta
which the UAV has flown. images and RTK coordinates. This assured us. “The secret is in the use of
Skelton said the Topcon GNSS data is stored onboard for post-flight gallium arsenide [GaAs] rather than silicon.
receiver is now available in this role download and analysis using software GaAs has been used in satellites for

36 Spring 2015 | Unmanned Systems Technology

Report | Commercial UAV Show

Also seen at Olympia (continued)

The secret is in
the use of GaAs, Selex ES, a member of the
Finmeccanica group of companies
which has been part-owned by the Italian
government, plans to improve the
used in satellites payload capability and flexibility of its

for decades but

established Falco UAV, the company
confirmed at Olympia. The original

until now, nobody Falco unmanned fixed-wing aircraft

was introduced in 2007, and by 2009 The Selex ES Falco is in use by the UN

has made it it was in active service. In 2014 it was

deployed by the UN for surveillance work in the DR Congo.

cost-effective for Powered by a rotary engine, the Falco has a 7.2 m wingspan, a 70 kg payload
capability and 8-14 hours’ endurance. More recently, Selex ES introduced the

terrestrial use Falco EVO, which has a 12.5 m wingspan, a payload capability of over 100 kg and
an endurance of 20-plus hours. The company said the next step will be to further
extend the payload capability of this craft while developing its configuration options
to make it suitable for a wider range of missions. To do this, the company takes
a modular approach to development, retaining the proven base concept while
decades but until now, nobody has ever increasing size, motive power and capability.
made it thin, flexible, and cost-effective for
terrestrial applications.” Quest UAV has launched the new Q-200 Agri Pro with Q-Pod technology, a system
Kapusta explained that Alta Devices optimised for precision agriculture and professional mapping applications, Nigel
is using GaAs – as used for a variety of King reported.
high-performance semiconductors such “The Q-200 Agri Pro houses our Twin NDVI System including 10.1 megapixel
as those in modern smartphones – in sensors in a rugged roll-axis gimbal, with one camera configured as a visible
a form well suited to UAV applications. sensor and the other as a near-infrared sensor. At the end of a mission, the high-
“The power-to-weight ratio is 1 W/g, definition still images and their corresponding GPS coordinates are entered into
and in full sunlight the output is about Quest modelling software, where Normalised Difference Vegetation Index (NDVI)
data can be generated for analysis of farmland.”
The Q-200 Agri Pro comes with a fully integrated parachute recovery system,
including a rugged ‘trapdoor’ system that protects the cameras on landing. It also
comes with a static gimbal that acts as a replacement camera housing in the event
of damage, and assists with capturing better verticals during the image post-
processing phase.

250 W/m2,” he said. Gardner showed us the Hokuyo laser

Kapusta said AeroVironment’s scanner, introduced in April 2014. It
battery-electric Puma AE UAV had weighs only 130 g and can operate at
been an early testbed for Alta Devices’ up to 20 m distance while requiring only
flexible solar cells. “Normally, on battery 150 mA of current. It produces a 2D
alone, that UAV would fly for two hours,” profile that can be transformed to 3D
he said. “After we added solar cells to through mapped movement of the UAV
most of its wing area, they increased carrying it. Gardner said that, compared
flight time to over nine hours – despite to the use of 3D Lidar equipment, this
The Hokuyo laser scanner cloud cover.” solution is lighter and consumes less
from Sentek Solutions At Sentek Solutions’ stand Adam energy as well as costing less.

38 Spring 2015 | Unmanned Systems Technology

Image of a car tyre from the Lepton infrared camera,
which can be used for monitoring the condition of
components in an autonomous vehicle (Courtesy of FLIR)

40 Spring 2015 | Unmanned Systems Technology

 Vision sensors | Focus

Developments in imaging technology are

paving the way for better and more integrated
onboard vision, as Nick Flaherty explains

Clear view
ahead
V
ision is a key capability for The payload is often a camera as well, images, or capturing a video stream for
autonomous systems; it can monitoring mountain terrain perhaps transmission to the searchers. In this
input to control and often as part of a search and rescue mission type of application the bandwidth of the
is the payload function. using high-resolution sensors for still wireless link is the critical factor, or the
Weight and the integration image or video data can be stored on
into the rest of the system is key for both board the craft for later use.
airborne and ground vehicles, whether The challenges with vision systems –
for controlling the craft or delivering The challenges from their field of view to their weight and
an image feedback to base from a power consumption – are similar across
standalone camera. with vision autonomous aircraft and road vehicles,
Image sensors are used in many ways but vary with the different spectra of light
in autonomous systems. Visual data is systems are being sensed, from visible light to infrared.
vital for guiding aircraft and vehicles
using a video feed to a processing similar across InGaAs sensors
unit to detect obstructions or objects. One approach for sensing in the
This requires the sensors to be tightly autonomous shortwave infrared (SWIR) range
integrated with the system controller.
Developments in both the sensor aircraft and road (wavelengths from 0.9 to 1.7 microns)
is the development of Indium Gallium
technology, now allowing lower weight
infrared sensing through the use of
vehicles but vary Arsenide (InGaAs) sensors, as the
bandgap of the material matches the
new micro-machining techniques, and
a new generation of controller chips
with the spectra energy of the photons. SWIR interacts
with objects in a similar manner to
for processing the data, is driving new
ways of designing vision systems for
of light sensed visible wavelengths, bouncing off
objects much like visible light and so
autonomous systems. gives shadows and contrast in

Unmanned Systems Technology | Spring 2015 41

Focus | Vision sensors

This intelligent vision acquisition

board uses a field-programmable
gate array (Courtesy of 3D-One)

the image, unlike the deeper infrared. reduce the field of view by using optics
This means images from an InGaAs with a longer focal length, but then if you
camera can be comparable to visible Infrared vision have a really narrow field of view it’s like
images in resolution and detail, and looking down a tube, limiting the ability to
are increasingly used as the payload has become see what’s happening. That is driving the
camera for a UAV for both day and industry towards developing sensors with
night use. more significant more pixels.
The InGaAs sensors can be made Another way to increase system
extremely sensitive so when built as for controlling angular resolution is to stitch the images
an array with millions of pixels, SWIR together from multiple cameras looking
cameras will work in very dark conditions. autonomous in different directions. There are systems
Using SWIR at night has another major
advantage. An atmospheric phenomenon systems as the such as the ‘thermal grenade’ camera,
which is a sphere with cameras around it
called night sky radiance emits five
to seven times more illumination than
cost and weight pointing in all directions, allowing a larger
field of view to be synthesised using
starlight, nearly all of it in the SWIR
wavelengths. This means a SWIR camera
of the sensors multiple sensors.
Alternatively there are multi-spectral
can use this night radiance – often called
nightglow – to ‘see’ objects with great
have come down algorithms that blend visible-light images
with IR images. This takes the visible
clarity on moonless nights. image and throws out all the information
except for the edges of objects, and this
Falling cost and weight is added to the thermal image.
Infrared vision has become a more varies depending on whether it is used
significant capability for controlling for control or payload. This often depends IR-visible combination
autonomous systems as the cost and on the sensor characteristics, making it a For example, some cameras combine
weight of these specialist sensors have complex development that has to balance the smaller IR sensor with a 640 x 480
come down. This has been a result of the the sensor array and the optics. visible camera using these algorithms.
combination of new sensor technologies The future of small infrared cameras The visible camera produces an image
and their integration into a camera, which lies in increasing the number of pixels. with decent edges if there is enough
is not simple. Previous IR cameras have The current generation has a small visible light. The advantage is that the
had to be cooled, making them too heavy number of pixels, typically an 80 x 60 visible camera doesn’t have to be high
for most autonomous systems, but a new (4800) array with a horizontal field of view resolution to work, and when those
generation of uncooled cameras opens of 50º, giving a field of view of each pixel edges are blended with the 80 x 60 IR
up fresh applications, and these – either of 0.625º per pixel. That’s a wide angle sensor it can be very effective for both
as modules or finished units – are being per pixel and really limits the ability to see payload and control applications.
used increasingly for control systems and small targets at any distance – cameras The pixels are made using micro-
as the payload. with a small number of pixels and wide machined mechanical systems (MEMS)
However, the post-processing of fields of view are essentially near-sighted. – tiny silicon structures a few microns
images and the integration of the camera There’s nothing you can do to change in size that are grown on a substrate.
into the autonomous control system that for a given sensor size unless you These silicon micro-bridges are then

42 Spring 2015 | Unmanned Systems Technology

 The problem with IR cameras is
that they have all this thermal energy
hitting the sensor from all directions.
You want the camera to ignore all IR
radiation except for that coming from the
scene being imaged through the lens.
However, it’s never really dark for an IR
camera – even with the lens cap on, the
sensor sees it – so the camera has to
calibrated to compensate for the out-of-
field radiation, and if that changes then it
has to be dealt with. One way of doing so
is to fit the camera with a shutter so that
This shortwave infrared module is used in An unboxed camera developed for the sensor momentarily sees a uniform
airborne cameras (Courtesy of 3D-One) UAV systems to minimise weight field of view and compensates each
(Courtesy of 3D-One) pixel based on that. This process, called
tuning, can be done with an external and
released from the substrate and coated manually controlled shutter.
with a vanadium oxide temperature-
sensitive material. The design of these The technical Sensor combinations
elements is complex. The bridges have One application where a combination of IR
thin legs to get the signal out of the micro- challenges lie sensors is being used is for studying the
bridge quickly, but at the same time they Arctic Ocean, where researchers are using
have to have a high thermal resistance to in dealing with UAVs with two hyper-spectral sensors to
avoid introducing errors into the sensor investigate climate change. These sensors
array. At the same time, the pixels in the more data while look at more than just the red, green and
array have to have a low thermal mass, to blue components of visible light: digital
give the sensor a short time constant of meeting the signal processing captures the light
20 ms or less and allow a fast response to
changes in the IR image. power needs intensity (the radiance) for a large number
(typically a few tens to several hundred) of
The optics are made out of etched
silicon. These are special lenses that
that influence spectral bands. Every pixel in the image
thus contains a continuous spectrum
are a couple of millimetres across, and
they can be designed with different
battery life for (in radiance or reflectance) and can be
used to look at the objects in a scene with
fields of view.
The industry is pushing towards
portable systems greater precision and detail.
One sensor covers the visible near-
more pixels in these micro-cameras, infrared spectrum (400-1000 nm) via a
for example with IR sensors with 160 x standard silicon CCD (charge-coupled
120 pixels, but this then raises the issue device) with a focal plane array of 7.4
of processing the data. Taking the data raw sensor data, as it is not easy to x 7.4 µm pixels. This gives a resolution
and processing it into an image has a get a good image out of these arrays. down to 1.9 nm with a dynamic range
computing cost to it – the more pixels You have to calibrate the image sensor of 60 dB, and a frame rate of greater
you have, the more computing steps per in the camera housing to produce a than 90 fps to capture more detail. The
frame, and that drives up the camera’s good image, regardless of the ambient second sensor covers the near-infrared
power consumption. That’s where the temperature, whether it be -10 C to 65 C, spectrum, from 900-1700 nm. This is a
technical challenges lie, in dealing with which is a serious technical challenge. cooled sensor with an InGaAs IR array
more data while meeting the power There are also thermal gradients across and a 25 x 25 µm pixel size.
requirements that influence battery life for the camera, where one side of it can be Similarly, other developers are
portable systems. hotter than the other because of where combining SWIR and long-wave IR
One way around this is to use a it sits in the overall system, so you have (LWIR) sensors for the 0.9-1.7 µm and
custom controller with proprietary image to make it immune to these effects or 8-14 µm bands into a single camera with
processing algorithms for taking the compensate for them. its own interface for unmanned intelligent

44 Spring 2015 | Unmanned Systems Technology

 Vision sensors | Focus

systems, fusing multiple sensors into Most of the pre-processing, such

compact and lightweight systems. as balancing multiple cameras to run
A key factor for airborne systems is For road vehicles, synchronously in frame lock and specific
the ability to combine daylight imaging features such as false colouring and
sensors, hyper-spectral imaging sensors the integration of synchronising all image channels from
and SWIR and LWIR sensors, along the sensors, is handled in the FPGA.
with the electronics, optics and vision the vision sensor A key processing algorithm on the
system programs. This is the way to get FPGA is ‘decubing’ the hyper-spectral
the minimum power consumption and algorithms has image data. An ordinary camera carries
weight, both for control of the UAV or as out RGB colour processing, but a
a payload system. already moved to hyper-spectral sensor has to handle
One application here has been to
combine two hyper-spectral sensors for more dedicated hundreds more ‘colours’, so the image
becomes a cube, with the x and y axes
precision agricultural monitoring. One
camera sees deep UV to visible, from
programmable as the spatial resolution and z axis as
the colours. This is handled in real time
400 to 600 nm, and the second from
blue to IR (600-1000 nm). Coupled with
devices for ADAS in the FPGA, taking position data from
the GPS and inertial navigation systems,
this is an intelligent data acquisition time-stamping the data and streaming
board controlled by a processor running it to an onboard solid-state disc drive
a customised embedded Linux operating (SSD) or to a modem for wireless
system. Other mission sensors such as transfer to a ground station. The Q7
a GPS, motion sensors and radiance board then handles image processing
sensors to measure the spectrum of the (FPGA), a microchip that can be user- and control, file management for the
incoming light can also be integrated into programmed after manufacture, which SSD and system management.
a complete data capture system, either implements the control algorithms in
for control or payload. hardware rather than software. This is Integration
Low power – so that a smaller, lighter then integrated into the OEM image Compared with driverless vehicles, UAV
battery can be used – and low mass are sensors. The company designs its own integration is relatively new. For road
key design requirements, as every extra daylight cameras and optics, while the vehicles, the integration of the vision
100 g reduces flight time by a minute for IR cameras come from a third-party sensor algorithms has already moved to
many UAVs. To help achieve this, one supplier. The FPGA board then connects more dedicated programmable devices
company has designed an acquisition to a standard PC board in the Q7 format that are being used for advanced driver
board that measures 95 x 95 mm to that can run the latest single-, dual- or assistance systems (ADAS). These
host a field-programmable gate array quad-core processors. processors provide functions such
as lane departure warning, advanced
cruise control, traffic sign recognition,
The image processing used for advanced driver assistance (ADAS) systems (below) is being
extended into the control of driverless cars using the same data and algorithms (Courtesy of Toshiba) pedestrian and object detection, forward
collision warning and preventing
reversing if there is an obstruction
behind the vehicle. These are all key
algorithms for using the data from a
vision sensor to control a self-driving
car, and it is here that the architecture of
the controller is vital.
The camera applications are now
so complex – with up to eight image
sensors running in parallel, and other
sensors such as radar and even Lidar
laser detection systems – that it is
important to understand which functions
are required for controlling the vehicle.
For that you need the hardware

Unmanned Systems Technology | Spring 2015 45

Focus | Vision sensors

The architecture of the latest advanced driver

capability and make the most intelligent assistance chip is a key step towards handling
assignment of the functions to the video in driverless cars (Courtesy of Toshiba)
hardware architecture. This is a key
step towards having a controller for a
self-driving car with sensor fusion that However, the ARM architecture is
combines the data from all the sensors still a key one for automotive systems,
– not just the image sensors – to make and another third generation of image
decisions. Having a processor that can system combines two 500 MHz digital
process other information alongside the signal processing cores with a fully
image data is a key step in the gradual programmable vision acceleration
introduction of autonomous functions. core alongside two 200 MHz ARM
One chip designer expects to have Cortex-M4 cores, along with an image
237 customer designs for ADAS signal processor.
by 2017 using a family of vision Another, relatively new entrant into
processors based on the MIPS this technology is using two ARM
processor architecture. This sensor Cortex-A15 processor cores with a
data processing capability was the graphics core and four of its own image
driver for the definition of a new family processing cores, as well as a separate
of processor cores, the M51xx, that will video decoding core. There is also a
sit at the heart of the next generation of core dedicated to viewpoint conversion
devices for customers such as General of the video, where six channels can
Motors, Volvo and Honda. each support a high-resolution camera
The processing architecture is not such that the combined output can
necessarily fixed though. One supplier, produce high-definition surround view
which launched its first automotive images. The part started sampling at
image processor in 2004 around the the end of 2014 and is set for mass
architecture from chip designer ARM, production in 2016.
has changed its approach with its fourth- The challenge The challenge for vision sensors in
generation design. The hardware-based autonomous vehicles has been to bring
architecture of the latest device moves for vision sensors all the systems together and make them
from ARM to ten dedicated media work together, and that’s the direction in
processor engines (MPEs) supported by in autonomous which the market is heading. Combining
a custom-designed very long instruction electronics and physics allows more
word coprocessor core called MeP vehicles has focus on the detection of light and
running applications such as traffic signal turning it into digital information, which is
detection and lane detection warning. been to bring helping OEMs build better systems and
These are optimised for transferring the
large amounts of data generated by the all the systems allowing the different types of optics to be
integrated with the thermo-mechanical
vision systems.
together and elements into a small housing, along
with the corrections needed in the optical
Architectures
The MPE media processor cores are
make them system to see what the customer wants
to see.
themselves supported by 14 hardware
accelerators for processing the image
work together Acknowledgements
data, and the company now supplies The author would like to thank
the application programming interface Austin Richards at FLIR Commercial
and libraries that developers use to Systems, Marco van Hout at 3D-One
program the devices as camera functions control code from the image recognition and Klaus Neuenhuskes at Toshiba
become more standardised. Moving from algorithms running on the MPE for their input into this article, plus
one processor architecture to another coprocessors, it is easier to move the contributions from Sensors Unlimited,
has traditionally been regarded as a control tasks to the MeP using a standard Headwall Photonics, Xenics, congatec,
difficult task, but by separating out the C++ compiler. Renesas and Mobileye.

46 Spring 2015 | Unmanned Systems Technology

 Cutaway of Danielson’s Trident, three-cylinder turbodiesel

Three-pronged attack
Ian Bamsey investigates a state-of-the-art French
turbodiesel that’s been created specifically for UAV use

D
anielson Aircraft Systems (I3) 100 TD2 customer unit is profiled Everything for the project was designed
(DAS) is part of the Groupe here. The director of DAS, Frederic in-house at DAS, whose extensive
Danielson organisation Hubschwerlen, says the Trident project manufacturing capability meant that very
based beside the Magny was specifically targeted at the UAV little component production had to be
Cours (former Grand Prix) market rather than general aviation. outsourced. Although Groupe Danielson
circuit in France, and well known in the The customer Trident was developed has extensive CI experience, it didn’t
automotive world for its road engines. as a complete package with all ancillaries follow that the Trident engines had to
Danielson has enormous depth of including the cooling system and be turbodiesels; they could have been
experience of compression-ignition (CI) transmission through to the propeller. The spark ignition (SI) like many existing UAV
engines, so popular with French car state-of-the-art 100 TD2 is a 1.1 litre I3 that engines. The design team saw a number
manufacturers. That experience, and its features two-stage turbo-supercharging of advantages from the use of CI.
increasing involvement in aerospace, yet has a dry weight of only 70 kg. It Regardless of power level – and Trident
led to the French military commissioning runs on regular diesel or heavy fuel is catering for 100 to 180 bhp – the
it to develop a clean-sheet-of-paper and exploits mechanical rather than top priority in the UAV market is flight
turbodiesel specifically for UAV use. electronic injection to keep it free from time. In this respect CI is theoretically
Following that commission DAS now electromagnetic interference. Among its advantageous compared with SI, since it
has a range of three Trident engines, innovations is a novel torque-smoothing is inherently more fuel-efficient. A higher
of which the initial inline three-cylinder system to protect the transmission. compression ratio and a lack of throttling

48 Spring 2015 | Unmanned Systems Technology

 Danielson Aircraft Systems Trident 100 TD2 turbodiesel | Dossier

Compact turbodiesel power –

the Danielson Trident 100 TD2

The Trident project

was a response to
issues the French
military had with
operating its UAVs
such as fuel
injection problems
and high total
ownership costs

of the charge air are, on paper, the basis of

efficiency higher than that of a comparable
SI engine. The main inherent drawback of
a CI engine though is lack of valve overlap
to assist charging, but that is counteracted
by the use of turbo-supercharging.
The CI process does reduce the time
available for injection, but that isn’t such
an issue given a combination of forced
induction and a correspondingly reduced
engine speed. That in turn (due to the
relatively low crankshaft speed) implies
reduced frictional losses, while the are better suited to CI rather than SI. injection problems, acoustic signature
relatively high compression ratio and the The exhaust temperature of a CI concerns, rapid wear, complexity of
use of turbocharging imply high torque. engine is inherently lower than that of an maintenance, high fuel consumption,
The only real downside is the structural SI one due to a higher expansion ratio. high total ownership cost and operational
requirement caused by the associated On top of that, the 100 TD2’s dual-stage risks. All of these were associated
elevated cylinder pressures, which turbocharging system further reduces with the existing propulsion system
can result in a weighty engine, but the acoustic and thermal signatures of the military used, which had not been
Danielson’s expertise with in-house the exhaust – important considerations designed for requirements such as high-
aluminium and magnesium castings has for many UAV operators. Most important altitude operation, extreme temperatures
paid dividends here. of all though is flight time, and and high operational availability.
Although mechanically injected, the Hubschwerlen says the Tridents have Trident is designed to address those
Trident family exploits direct injection been measured by DAS to be 40% more issues, exploiting proven Danielson
into the combustion chamber, reflecting fuel-efficient than a UAV four-stroke SI technology in a package designed
state-of-the-art CI practice. Also, CI engine of the same power level. specifically for UAV requirements, in
lends itself to the use of heavy fuels: He also reports that the Trident project particular exploiting a range of fuel
kerosene (paraffin) based fuels are was a response to specific issues chemistries. In general terms, kerosene-
widely used by armed forces. They are encountered by the French military in based fuels ignite quicker than diesel,
less flammable than gasoline, and so operating its UAVs, identified as fuel calling for appropriate adjustment

Unmanned Systems Technology | Spring 2015 49

Dossier | Danielson Aircraft Systems Trident 100 TD2 turbodiesel

Anatomy: Danielson
Aircraft Technology
Trident 100 DT2

Danielson Aircraft Technology

Trident 100 DT2
UAV engine

I3
76.0 x 80.5 mm = 1095.6 cc
Two-stage turbo-supercharging
Diesel and kerosene-based fuel
options (36-70 cetane)
Aluminium structure
Linerless, plasma-sprayed
The Trident 100 TD2 uses
two-stage turbocharging bore coating
Five main bearings, plain
Steel crankshaft, three pins
Steel con rods
of injector timing, and have a lower Light alloy pistons, three rings
cetane rating (the readiness to ignite Chain-driven single overhead
a fuel in a CI engine), which lowers The charge camshaft
power output. Kerosene also attacks Two valves per cylinder
conventional injector pump seals, but air enters the Vertical valves
that is straightforward to counter at the Undisclosed valve sizes
design stage. compressor of Mechanical direct injection
The Trident project started in 2007, 15:1 compression ratio
and the first customer engine was the turbo one then, Maximum rpm, 3600
dual-stage turbo 100 TD2. That in turn
was followed by development of two I4 having been initially Some key suppliers
derivatives, one with the same bore
(76 mm) and stroke (80.5 mm) as the compressed, to this engine

1.1 litre for 1460 cc, the other a larger

(86 mm) bore and longer (97 mm) stroke
travels on to the Main castings: in-house
Bore coating: Oerlikon-Metco
for a 2250 cc displacement.
Early on in the customer engine project
compressor of Bore coating: Capricorn
Crankshaft: in-house
the 100 TD2 met its performance target;
subsequent development focused on
turbo two Camshafts: in-house
Pistons: undisclosed
reliability, and Danielson’s extensive testing Con rods: in-house
facilities include the ability to simulate the Valves: undisclosed
engine running at altitude. The Trident Turbocharger: in-house
series was launched onto the open market of turbo one then, having been Carbon fibre cam cover: Scaero
at the 2014 AUVSI show in Orlando, Florida, initially compressed, travels on to the DLC coatings: Oerlikon-Metco
where this research was conducted. compressor of turbo two, where it is
further compressed before being fed
Charging and fuelling to the engine through an air-to-air
The 100 DT2’s two-stage turbocharging aftercooler. The plenum that then feeds the turbine of turbo two, then to the
sees the use of two individual the three cylinders is integrated into the turbine of turbo one before exiting
turbochargers that are fitted in series. cylinder head. to the atmosphere. The turbos are
The charge air enters the compressor The exhaust gas is fed first into straightforward designs – no

50 Spring 2015 | Unmanned Systems Technology

For over 40 years we’ve been making
high performance oils to solve problems
for race teams worldwide

The challenges faced by UV manufacturers and users are just as

complex. We’d love the opportunity to help you overcome them!
Tel: (562) 595 7208 | NEO OIL USA | 2871 Gundry Ave, Signal Hill, CA 90755
email: neooilusa@[Link] | [Link]

The Flamingo Out of sight –

within reach

Fifty years of remote experience.

If you need more endurance or payload for your
UAV tasks, look for the fixed-wing UAV of choice.
Interchangeable payload nose-cones permit
rapid changing for diverse missions.
Used by universities and governments in
North America and Australasia.
4m wingspan and 25kg operating design.
Top speed 78kts, cruising 52kts. Endurance up to 10 hours.
Single and twin-boom, and
long-endurance variants available.
Payload integration services available.

Silvertone Electronics | 1/8 Fitzhardinge Street,

Wagga Wagga NSW, 2650 Australia | +61 6931 8252
[Link] | contact@[Link]
The Trident TD2 nestling in a typical UAV airframe

complications such as variable geometry of the combustion chamber at TDC. The need for spark plugs and a throttle,
– and are sized to meet the needs of injector is arranged to fire directly into computer-based control of its injectors
the dual boosting process; no wastegate the bowl. The geometry of the porting is clearly theoretically advantageous. On
is required. The three-into-one exhaust is crucial from the point of view of the other hand, in the UAV environment
manifold feeding turbo two is a stainless creating the appropriate turbulence in the there is a huge danger of interference
steel production made in-house. charge air as it enters the cylinder, while between electronic engine control and
With the two-stage compression, the geometry of the bowl is crucial to the other electronics operating the craft
plenum pressure is as high as 4.0 bar combustion effectiveness. and its sensors.
absolute and the compression ratio is Since this is a CI engine, the only “A UAV is full of electronics,”
15:1. That ratio might not be high by CI throttling of the charge air is a butterfly Hubschwerlen points out. “If you have
standards but it has to be seen in the for emergency cut-off only; engine load mechanical engine control you will have
context of the plenum pressure, implying is governed purely through adjustment no compatibility problems, and with fewer
an extremely high effective compression of the fuelling. In this day and age, one components in the engine package you
ratio. As is normal in this type of engine, would expect the Trident engines to will have inherently better reliability. The
the combustion chamber is bowl-in- be electronically controlled, as indeed lack of electronics makes the injection
crown, with in this case the geometry are all advanced automotive engines system immune to electromagnetic
optimised for maximum fuel mileage. The produced by Danielson, whether CI or interference.”
only concern regarding emissions is that SI. An engine management system is Hubschwerlen adds that, in the event
there should be a lack of visible smoke. particularly significant in the context of of any system problem, using a common
The two valves serving each cylinder an SI engine, allowing coordinated rail served by a single pump would put
are upright, so it follows that the piston control of the injection, the ignition and the engine entirely out of action. That
crown can be flat aside from the central often these days a fly-by-wire throttle. danger is avoided by each cylinder
bowl, which in any CI engine forms much While the CI engine avoids the having its own independent injection

52 Spring 2015 | Unmanned Systems Technology

 Danielson Aircraft Systems Trident 100 TD2 turbodiesel | Dossier

pump serving its single injector. “The the aforementioned torque-smoothing

common rail system is a source of device, then behind that the lower of
potential engine failure; here we have a In most respects the two step-up gears that comprise a
fully mechanical system for each cylinder fixed-ratio gearbox taking the drive up to
operating independently,” he says. the performance propeller shaft height. On the very rear of
In the turbodiesel engine, fuel pressure the crankshaft, underneath the propeller
needs to be very high relative to SI to of the Trident’s shaft, is the starter-generator unit, which
suit the combustion environment, and provides power for the UAV’s various
the higher it is the smaller the droplets mechanical electrical systems.
of atomised fuel and the higher the The nitrided steel crankshaft runs in
efficiency of burn: a typical figure for a engine control five main bearings, a pair sandwiching
common rail road engine is 1800 bar.
In the Trident system, fuel pressure system matches each big end with the fifth ahead of the
torque smoothing device, supporting
peaks at just over 1000 bar, which is
perhaps the only compromise with the
that of an the lower of the two gears in the step-up
gearbox. Conventional tri-metal bearings
mechanical system. In most respects,
says Hubschwerlen, system performance
electronically are used for the main and big-end
bearings, of which there are three, each
matches that of an electronically
controlled one.
controlled one cylinder having its own crankpin.
The crankshaft is made in-house, as are
The individual fuel pump for each the steel con rods, which are H-section.
cylinder is driven by the single overhead The con rod small end is bronze bushed
camshaft that also operates its two inline and the piston pin running in it is an
valves. Inside the pump is a piston that, either by Oerlikon-Metco in Switzerland uncoated steel item, again made in-house.
as it reciprocates under the action of the or by Capricorn in the UK, with Capricorn The three-ring piston is aluminium with
cam, opens and closes a port that feeds also honing the bores. a cast-iron reinforcement carrying the
fuel to the injector. The injectors are not The magnesium front cover houses the upper rings, as per conventional high-
switched on and off but dispense slugs of timing drive and forms a vertical oil tank. performance CI practice. Danielson is
fuel as dictated by piston motion. In addition to the carbon fibre top cover, considering the alternative of a steel piston,
The pump piston has a helical delivery a starter-generator at the rear of the as pioneered by Peugeot in its
slot. Turning the piston axially, by crankshaft also has a carbon fibre cover. Le Mans-winning engine.
altering the position of the slot relative The 100 TD2 has its timing drive at The valve seats are sintered steel while
to the port, varies the quantity of fuel the front and its propeller drive at the the valves are solid steel; dual steel
delivered on each stroke. Rotation of rear. On the tail of the crankshaft is valve return valve springs are used.
the piston is achieved through a push-
pull mechanism ultimately controlled by
The Trident 100 TD2 uses a clever mechanical injection system
the autopilot. “This system is simple but
effective, and removes points of potential
failure,” Hubschwerlen says.

Engine structure
A turbodiesel calls for a rugged engine
structure to contain the implicit combustion
pressures. This dry-sump engine has an
all-aluminium structure aside from the
lower crankcase/sump and a front cover,
both of which are magnesium.
Of the relatively few outsourced items,
the head gasket is stainless steel. The
engine is linerless but the bores have
a hard iron-based coating using the
Oerlikon-Metco SumeBore plasma spray
process. The bore coating is applied

Unmanned Systems Technology | Spring 2015 53

Dossier | Danielson Aircraft Systems Trident 100 TD2 turbodiesel

Transmission
The gearbox reduces crankshaft speed
to propeller speed at a fixed ratio of
1.28:1, so that at 3600 rpm engine speed
the propeller runs at about 2800 rpm.
To save weight the engine has no
flywheel, and in any case the propeller
acts as such. However, there is that
feature described by Hubschwerlen
as “an innovative torque-smoothing
system” located between the crankshaft
and the fixed-ratio gearbox. “Due to the
running pattern of a three-cylinder CI
engine there is a danger of breaking
gears,” he explains.
This system overcomes the rotational
irregularity – acyclism – associated with
this type of CI I3: acyclic impulses can
The Danielson Trident 100 TD2 uses a novel ‘torque smoothing system’ cause damage and excessive wear of the
transmission gears. The use of the torque-
smoothing device allows the use of lighter
The vertical valves are each operated by cooled, with conventional water jacketing gears, helping to counteract its weight.
its respective cam via an inverted bucket for the cylinders. Hubschwerlen remarks “This device stores energy during each
tappet. The camshaft and the inverted that specific attention has been paid to combustion stroke and gives it back
bucket tappets are DLC coated, and the minimising the aerodynamic drag caused during the compression stroke, providing
journals of the DLC-coated camshaft by heat rejection. The oil pump is of the a very smooth drive to the propeller,”
run directly in the cylinder head. The gear type and is produced in-house, and Hubschwerlen says. “It isn’t a damper, it
camshaft is chain driven (with cam contains one pressure and two scavenge doesn’t lose any energy, it takes energy
timing fixed). stages in the same housing. The water and gives it back at the right time to
The drive coming from the front of pump is of the centrifugal type and is smooth the drive. It is all-mechanical and
the camshaft sees the injection pump- likewise produced in-house. simple – it has only five moving parts. I
operating cams set between cylinders
one and two, between cylinders two and
Starter-generator for the Trident 100 TD2
three and outrigged beyond cylinder
three. A tenth cam provides the drive for
a feed pump, which lifts fuel from the
tank (through a single line) and delivers
it (via an upstream tri-branch) to the
individual pumps.
Given the pressure and heat of
combustion associated with this CI engine,
like most turbodiesels it uses oil gallery
pistons. That is a step beyond the normal
high-performance SI solution of one or
more oil jets cooling the underside of the
piston. In this case, two jets are arranged
to spray through ports feeding a circular
gallery that runs within the piston crown.
The flow of cooling lubricant escapes
through outlet ports, due to the so-called
‘cocktail shaker’ effect.
The 100 TD2 is both oil- and water-

54 Spring 2015 | Unmanned Systems Technology

don’t want to disclose anything further
until we have patented it.” About Danielson
In addition there is a clutch located
ahead of the gearbox, on the DLC-coated
propeller shaft. This is a centrifugal Danielson was established in the mid-1970s as a motor racing team with engine
clutch, so it operates automatically, and it and chassis engineering capability. In recent years it has concentrated on engine
is intended only to assist starting and to engineering, with race engine involvement accounting for only a small part of
help protect the powertrain in the event its turnover and no other competition activity. Its current operation embraces
of a propeller strike. Danielson offers an automotive, defence and aerospace powertrain work, and it has the capability to
optional variable-pitch propeller that uses design, manufacture, build, test and develop engines. Its manufacturing capability
either electric or hydraulic actuation and includes an in-house aluminium and magnesium foundry.
has carbon fibre composite blades. In recent years Danielson has become a compression-ignition specialist,
reflecting the popularity of diesel over petrol cars in the French market, and the
Performance company does a lot of specialist work for the likes of Renault and Peugeot-Citroen
Weighing 70 kg dry, the 100 TD2 in the roadcar sphere. It produces a number of prototype engines each year,
comes in with an installed weight of primarily for road use although increasingly for aerospace applications.
90 kg. Its quoted maximum crankshaft
speed of 3600 rpm is a nominal figure
– Hubschwerlen says the engine can
tolerate a significant overspeed, to at the 100 TD2 can accept fuels within a
least 4000 rpm. wide cetane rating range. Cetane rating
Hubschwerlen notes that some This device stores is on a scale of 0-100, and typically it
customers want to use a fixed-pitch is around 50 in the case of roadside
propeller, and consequently need to energy during diesel. “It runs perfectly with fuel from
constantly vary its speed and with that 36 and up,” Hubschwerlen says. “99%
the engine speed. Others will use the each combustion of fuel is above 36. If the cetane rating
optional variable-pitch propeller, allowing is below that it will just be more difficult
them to operate the engine within a stroke and gives to start the engine. The cetane rating
more consistent rpm band. In the case can be as high as 70, but when it is
of helicopter deployment, the engine can it back during that high there will be a marginal loss
normally run at a constant speed.
At low altitude the stage one (initial the compression due to its burning characteristics, as
we have had to design the engine for a
pressurisation) turbocharger provides
only about 10% of the total charge
stroke, providing lower rating.”
With rising engine speed, specific fuel
pressure increase. However, as the air
thins with rising altitude the situation
a very smooth consumption in g/kW/h falls to 260 at
12 kW (16 bhp) and continues falling to
changes, to the extent that at high
altitude the situation is closer to 50-50.
drive a low of 230 at 35 kW (47 bhp), a figure
sustained through to 45 kW (60 bhp)
Hubschwerlen remarks that the critical before climbing again, still to less than
altitude, at which the turbos share the 250 at 75 kW. Overall fuel consumption
work evenly and maximum power is is quoted as 0.38 lb (172.37 g) per bhp
produced, is 14,000 ft. Above that the from 50 Nm (about 37 lb-ft) at 1300 rpm per hour. “That is comparable to an
thinning of the air saps power, to the propeller speed through to its peak of automotive engine using a common
extent that it is only 54% of maximum 260 Nm (about 192 lb-ft). Maximum rail system with electronic control,”
at the designed operational ceiling of torque of 260 Nm implies a brake mean Hubschwerlen says. “Compared to a
30,000 ft. effective pressure of 29.7 bar. comparable current Wankel UAV engine,
Maximum cylinder pressure is about Power naturally echoes the torque we cut fuel burn in half.”
160 bar; maximum torque is made at curve, rising from just under 10 kW That emphasises why DAS has opted
3200 rpm, maximum power at 3600 rpm. (10.4 bhp) at 1300 rpm propeller speed to for CI rather than SI. Of course, SI engine
By design, that speed corresponds to a maximum of 75 kW (100 bhp) at 2750, manufacturers would have their own
the normal required operating speed a figure sustained for almost 1000 rpm. opinion but clearly there is a good case
of the propeller. Torque rises steadily Reflecting its multi-fuel capability, for the turbodiesel in this context.

Unmanned Systems Technology | Spring 2015 55

MIRA’s cooperative car project is a key
part of a new cyber security initiative

Stable platform
How do you ensure an unmanned system
is secure and safe? Nick Flaherty reports

A
s autonomous systems situations, and these safety scenarios One recent example of a safety failure
move increasingly into vary from platform to platform. The safety is the UAV that crashed in Iraq and was
the realm of commercial case for UAVs in the air is different from captured by Islamic State fighters. The
UAVs and driverless cars, that for autonomous cars, and different UAV lost its GPS connection and so did
so the focus on their again for maritime and agricultural not have accurate position data. The
security and safety has come to the fore, equipment. UAVs operating in a desert safety case for this eventuality was for it
as it becomes apparent that there are war zone, for example, have a very to return to base, which was fine, except
significant gaps in both, and that a system different set of requirements from an that the last noted base was in Belgium.
that is not secure is of course not safe. autonomous aircraft in civilian airspace Returning to that base, rather than the
The challenges of making unmanned or one used for monitoring farmland. local one, meant the UAV ran out fuel,
systems secure and safe come from Similarly, the requirements for a driverless crashed and was captured.
many directions. There is still work car operating on a motorway are different The safety of an autonomous system
to do to make sure they perform in a from those for an autonomous tractor or can also be compromised by malicious
predictable and acceptable way in all combine harvester. hacking. This can range from hacking

56 Spring 2015 | Unmanned Systems Technology

 Security and safety systems | Insight

into a vehicle-to-vehicle communications

network via a ‘man in the middle’ attack,
for example, where a hacker replicates
another vehicle to divert an autonomous
car, to viruses and Trojans that can
render a system unusable if a ransom
is not paid. There is also the risk of
physical ‘hacking’ that aims to confuse
autonomous systems by changing the
physical environment. With driverless
cars using lane departure warning as
a key control mechanism, one risk that
has been highlighted by an automotive
equipment supplier is a hacker who
directs a vehicle into oncoming traffic,
leading to a crash.
Consideration of issues like these
has led to the formation of a strategic
cyber security collaboration group
for connected vehicles in the UK to
investigate and define security best
practices, including how the vehicles
communicate with each other, with people
and transport infrastructure.
Mapping the functions of a UAV onto intelligent safety agent (top), and what that means for the
functions of the code being developed, above (Courtesy of University of Liverpool)
Countering the threat
The group brings together the Motor
Industry Research Association exposure of autonomous systems to
organisation (MIRA), which has wide New capabilities threats via mobile phone technologies,
experience of vehicle safety and testing onboard diagnostics ports and poorly
as well as autonomous and connected are necessary disciplined interactions between different
vehicles (see Dossier, page 62) and pieces of software.
Plextek Consulting, a specialist in because the Drawing on expertise from other
automotive product and systems design. industries and sectors such as IT and
The third member of the group, Intercede, potential threat the military, along with more industry
is perhaps unexpected, as it is an identity investment, will help deliver better
security specialist with experience in of cyber attacks system protection and ensure that the
the UK and US of managing employee
credentials for large organisations. Its is broad, from connected vehicle – especially the
autonomous connected vehicle – is safe
MyID software enables operators to
create and use trusted digital identities
hackers, malware enough for widespread use in the public
sector in the near future, say leading
for autonomous systems using secure
‘identity agents’ rather than passwords.
developers as developers. This is why the design
process for all connected vehicles has
These new capabilities are necessary
because the potential threat of cyber
well as terrorists to be more rigorous and itself provide
an improved method of reviewing the
attacks is broad, ranging from the process so that both the cyber security
opportunist hacker to malware developers and reputation of the autonomous car
and terrorists. This year will see cyber developers will be enhanced.
attacks and the threat to vehicle comms the same way as functional safety and To tackle these issues, the cyber
systems and driverless cars being physical threats are now. Furthermore, ‘in- security group will also explore how
considered by developers as part of the car vulnerabilities’ need to be addressed a secure connected environment can
essential safety protection of vehicles in urgently as there is potential for massive improve the efficiency of transport

Unmanned Systems Technology | Spring 2015 57

Insight | Security and safety systems

appropriate mathematical analyses can

contribute to establishing the correctness
and robustness of a design.
As the software becomes more
complex, so the cost of fully verifying and
testing autonomous systems becomes an
issue, which is leading to new software
architectures to reduce the cost of the
development of programs such as
agents – small pieces of code that ideally
can be fully verified and developed using
formal methods to be as safe and secure
as possible. The agent then monitors
the rest of the system software to watch
out for aberrant behaviour or attacks.
However, there is still a lot of discussion
as to what to include in the agent,
Mapping the safety requirements of a UAV onto the requirements of a formal development process
(Courtesy of University of Liverpool) the partitioning between the agent’s
functionality and the rest of the system
software.
infrastructure, including interactions with standard from ASTM International in This is largely driven by the experience
traditional as well as electric engines and the US covers maritime systems, with of the PC and business software
the creation of cooperative vehicles. versions for surface and underwater craft, market, where the cost of protecting
The connected vehicle brings with it and deals with autonomy and control, against attacks has become significant.
the potential for severe consequences comms, system interfaces and data Autonomous systems simply can’t afford
should security systems be unable formats. to have the same problems as the PC
to mitigate any cyber threats, so it is DO-178C clarifies the boundaries cyber security sector, where current
essential that vehicle components can between the concepts of High Level approaches are insufficiently robust and
‘trust’ each other and that only the driver Requirements, Low Level Requirements, costly to develop and maintain. This is why
or autonomous driving system can give and Derived Requirements in DO- companies such as Intercede are being
the car instructions, with hackers unable 178B, and gives a better definition included in the initiative at this relatively
to inject their own commands. Identity of the exit/entry criteria between early stage to provide security from the
and credential management is vital in systems requirements and the software beginning of a design.
the fight against these potential security requirements and design. This is the
attacks. The MIRA-Plextek-Intercede key safety and security standard for Other developments
group has been formed to ensure that software in UAVs, but it also impacts on As a result there are other projects
best practice across the whole spectrum other standards such as DO-330 for the elsewhere looking at different ways to
of security is defined and in place early in software development tools, DO-331 implement both safety and security in
the manufacturing process, with the aim for the development and handling of autonomous systems, moving from
of protecting road users. software models, DO-332 for object- purely deterministic assessments to
Adding security into the mix also oriented software development and DO- probabilistic analysis. The industry
increases the issue of fragmentation in 333, which handles formal methods. consensus is that it is increasingly
the development of the safety-critical Formal methods are mathematically impractical to fully test and verify all
software. Rather than trying to develop based techniques for specifying, possible safety cases, so there is a
safety cases that cover all eventualities developing and verifying the software move to developing a consistent set of
across all platforms, the different aspects of digital systems. The behaviours which, for example, would
safety cases vary between the different mathematical basis of these methods define what a system must do, what it
standards. Automotive equipment consists of formal logic, discrete must never do, and what happens when
developers are working to the ISO 2626 mathematics and computer-readable it breaks.
standard, while aircraft suppliers are languages, and their use is motivated At the same time there is a move to
working to the US Department of Defense by the expectation that, as in other using agents to monitor the software
DO-178B and C standards. The F-2541 engineering disciplines, performing in an autonomous system for

58 Spring 2015 | Unmanned Systems Technology

 base to replace a module before it fails is
one way to reduce that security risk and
enhance the safety case.
So a lot of the safety and security
development for autonomous systems
is more around acceptable behaviours
and behaviour boundaries rather than
deterministic rules. This means that
testing and verification is more about
testing the boundary of operation of the
rules, time and time again.
But this also leads to a key new
aspect of autonomous systems, in that
the system can decide for itself about
the best course of action to take. This
‘rational agent’ approach combines
IKV’s medina tool for developing safety code
‘beliefs’ about its environment, the goals
it wishes to achieve and deliberation
changes or activity that are outside the There are a number of different strategies for deciding between options.
accepted behaviours. The argument is aspects to the development of these Clearly, formal verification is needed
that the agents can be smaller, more new tools. Autonomous systems are for this. By verifying the rational agent,
manageable and verifiable code that can increasingly controlled by a rule- we verify not what the system does but
be developed with formal methods. This based decision process, but the what it tries to do and why it decided
also presents a smaller ‘attack surface’ rules can change depending on the to try that option. For this we need
for malicious attacks, making the system environment and the position of sensors appropriate abstractions of the real
more secure, and can be used to monitor and actuators. For some sectors, control and sensing aspects, and tools
the overall health of the system, for such as agricultural equipment, the that can handle the formal verification
example by making a vehicle return to physical security to prevent unattended reliably.
base before a component fails. However, autonomous vehicles being stolen may This has led to projects such as
this latter use increases the complexity be considered to be more important PICASSOS (Proving Integrity of Complex
of the agent and undermines the than the risk of hacking. Automotive Systems of Systems) at the
advantages of using such an approach. University of Warwick to develop tools for
But even the agent approach still Monitoring and diagnostics systems such that they can be verified
needs an overall safety and security This also links to the overall system easily, cheaply, repeatably and quickly.
case to determine an acceptable course reliability, as component failures or It aims to prove the integrity of complex
of action or actions if the agent detects degradation can have a negative impact automotive systems – the systems of
problems, making the whole software on a rule-based decision approach, systems – and so has been looking
system definition a complex matter. so developers need to think about the closely at driverless vehicles.
As a result, determining the optimum effect of health checks on the decision- Led by Ricardo in the UK, the project
partitioning between system software and making process. As an illustration, brings together Axeon Power, D-RisQ,
agent is also an area of intense research. only 20% of the control system of an Jaguar Land Rover and Oxford and
All of this is creating a new focus aircraft is actually control – the other Coventry universities to apply automated
on cognitive tools and reasoning. An 80% is monitoring – so the integrity of formal methods to break away from the
autonomous system has to be seen in the system depends very much on the traditional approach of large amounts of
the context of an environment of some diagnostics, which can influence the processes supported by documentation.
sort, and probabilistic arguments are development of the software. The project aims to develop tools for
not deterministic, so it is not possible to Diagnostics are also often an area developing the initial specifications and
envision and therefore test for every set where the security of a system can be creating a prototype set of tools and
of circumstances. The problem is that it compromised, as this is where data processes to be applied in initial trials,
is expensive to try to get these answers enters and leaves it. Using internal and roll out training in the use of the new
through traditional techniques such as diagnostics to determine when an tools and techniques.
testing. autonomous system should return to Work such as this is vital to the

60 Spring 2015 | Unmanned Systems Technology

 Security and safety systems | Insight

industry. A recent report by McKinsey alternative method is to use a formal models in the SysML language and use
for the UK’s Innovate UK r&d agency proof, which can show the absence of a a formal model checker called FDR3
showed that the overall market for behaviour (which you can’t do with a test) (developed at the University of Oxford)
global robotics, which includes and can assess the responses under that implements parallel processing to
autonomous systems, will be $1.26tn multiple failures, including when under explore autonomous software designs
by 2025 but that a key barrier is attack. There can then be a mathematical for stability, reaching all the different
showing the technology is safe, and demonstration from requirements states automatically through a tool called
the key barrier for that is cost. Adding to object code to show the formal ModelFlow, from D-RisQ.
application security test and verification correctness of the result. At the moment, the translation of the
increases the cost unless it is built in Tools such as medina, from IKV properties is done by hand, which is
from the start of the process. Technologies, can take SysML and UML problematic and could be prone to
Part of the way to do that is to models as well as Simulink models errors, so the next step is to ensure
automate the design tools. If they can developed specifically for ISO 26262 those properties are carried over from
be certified for automatic test and automotive designs to integrate functional the SysML models to the model checker
verification then developers can take safety analysis and development automatically and correctly. This allows
certification credit for the use of that tool activities. It supports safety analysis and the model checker to demonstrate that
in whichever standards process is being design for software-controlled, safety- the design meets the requirements.
used. This then highlights the importance related functions with risk and hazard
of automatic code generation, proof and analysis methods – including hazard Conclusion
formal methods, and has been adopted list, risk graph, fault tree analysis and The move to autonomous systems is
in new, higher-level design approaches failure mode and effects analysis with challenging the testing and verification of
such as the Unified Model Language the modelling and code generation tools software from both the safety and security
(UML) and System-level Model from MathWorks. perspectives. New architectures such
Language (SysML). But moving to new architectures such as rational agents aim to simplify the
The executable code generated as agents means autonomous systems test and verification process, and at the
automatically has to be shown as a need a more complex verification same time reduce the attack surface for
full and correct implementation of the environment, so a project was set malicious software, making systems less
requirements, which can be costly. An up in January 2015 to take high-level prone to hacking.
All of this requires more automation in
the design tools, moving development
The University of Oxford’s FDR3 tool for checking the requirements of safety code
up from the level of the source code
to sophisticated yet flexible models of
autonomous behaviour. Taking these
models and automatically generating
code and verification dramatically
reduces the overhead of certification and
so makes system development more
cost-effective. It also gives the flexibility
to support different standards across
different applications while demonstrating
appropriate levels of safety and security.

Acknowledgements
The author wishes to thank Nick Tudor of
D-RisQ Systems, the UK’s Safety Critical
Systems Club, Mike Bartlett of Test and
Verification Solutions, Jeremie Guichet of
the University of Toulouse, Ireri Ibarra of
MIRA, Michael Fisher of the University of
Liverpool, Richard Parris of Intercede and
Andrew Ashby of Plextek Consulting for
their input into this article.

Unmanned Systems Technology | Spring 2015 61

One of the challenges for the system design of MACE 3 is that is
has to operate in three different modes – self-driving autonomous
mode, remotely driven, and with a driver in the vehicle

Doing some
groundwork
The UK’s MIRA automotive research institute is
trialling unmanned ground vehicles for the military.
Nick Flaherty reports on the lessons being learnt

M
ilitary vehicles face some been working on autonomous vehicle stage this was demonstrating the fusion
of the most challenging technology for more than a decade, and of navigation sensors and algorithms for
conditions possible, and in 2008 it won a Grand Challenge Award obstacle avoidance that allowed accurate
so provide an insight set by the UK’s Ministry of Defence (MoD) operation even in an urban environment.
into the key problems in for Best Use of Autonomy with its UGV At the heart of the MIRA facility is
the development of unmanned ground technology demonstrator MACE (MIRA a dedicated UGV proving ground, a
vehicles (UGVs). MIRA, the UK’s central Autonomous Control Engineering), 4 km series of test tracks for system
automotive research institute, has steered by a remote operator. At an early development. The tracks include test

62 Spring 2015 | Unmanned Systems Technology

 MIRA MACE | Dossier

hills, obstacles such as pot holes and could be a sacrificial vehicle if it triggered it is following predefined control code. It
ditches, complex bends, non-line of a roadside IED. It was first used in 2011 can follow a route on GPS, or predefined
sight testing and urban areas. There is by the British Army, and the contract to routes using a set of map coordinates.
also a fully integrated control room with use it has been extended to 2030. There are several key lessons from
line of sight of the whole UGV proving The difference between MACE 3 and the evolution of the UGVs. The first is
ground to provide a safe and secure Panama comes down to communications that using an optical system gives the
test track location for all of MIRA’s links. Panama is controlled remotely platform an increased level of autonomy
UGV development, as well as giving over a secure high-bandwidth radio rather than relying on map or GPS data
UGV operators training in a safe and network; MACE 3 is controlled over the for positioning.
controlled environment. Internet. The difference between the two The second key lesson is that the
MACE 3 is the third generation of the approaches is how robust that comms design of the platform also depends on
demonstrator, extending the research and link would be: MACE has less complex the level of autonomy required – this
learning of the previous versions into a needs in terms of range and data rate as is the thing that designers are still not
Land Rover Defender. MACE started out the autonomy lessens the need for data. getting their heads around, says MIRA.
as a load carriage vehicle for the MoD, For MACE 3 you can assign it a The key question is: what do you want
carrying kit, food, water and batteries in a route and set it a task to complete in an the autonomous platform to do? Deciding
‘follow me’ mode that would simply follow autonomous manner. You can set the on the task to be carried out determines
the vehicle in front. However, the initial parameters of the route and the task but the algorithms and the accuracy of
specification of this mode highlighted part
of the problem. The project didn’t really
The Vehicle Integration Module sits at the heart of the MACE 3 architecture
develop as the use case was unclear – to pull together all the separate functional elements of the design
what would happen if for example the
vehicle broke down, or the truck in front
went into a ditch?
So the MACE platform has been
extended to add fully autonomous
operation and a remote control mode
alongside the follow-me mode, as well
as the ability for a driver to jump in
and take manual control. This puts the
technology challenge right at the heart of
the development of driverless cars, which
also have to combine autonomous mode
with the ability for the driver to take over.
A key issue here is space. A fully
autonomous military vehicle can have the
control electronics installed in the cockpit, The separate algorithms for controlling MACE 3 are
implemented in an FPGA in the Vehicle Integration Module
but for MACE 3 the electronic control
system – called the Vehicle Integration
Module (VIM) – has to be small enough
to allow a person to drive as well as be
remotely controlled by an operator.
This contrasts with another remotely
operated UGV platform developed by
MIRA, called Panama. This is designed
to be always unmanned as it is aimed
at applications such as finding a route
through hazardous terrain and locating
improvised explosive devices (IEDs), with
an operational range of up to 20 km. It is
based on a modified Land Rover Snatch
which was a surplus item and which

Unmanned Systems Technology | Spring 2015 63

Dossier | MIRA MACE

The three cameras on MACE 3 provide a wide field

of view and are the key obstacle detection system

the sensor systems, and the experience unmanned heavy equipment or heavy on the navigation system, such as the
with all three generations of the MACE engineering vehicles to travel to the perimeter of a military base, while all the
platform highlights the different levels of location autonomously. Once there, the time checking for obstacles on the route.
autonomy for different uses. vehicles would probably then have to be One of MACE 3’s camera systems,
In a military context, the use cases are controlled remotely to achieve the task. called Piranha, provides a field of view of
driven by three Ds – tasks that are dull, All this means that the development of up to 100º and is encased in a rugged
dirty or dangerous. On land, that tends platforms such as MACE has to take into housing that is designed to withstand
to mean mechanical operations such account comms issues such as data rates, vehicle vibrations. It provides images in
as digging ditches or building banks, as latency and power consumption to achieve temperatures from -40 to +70 C, and
well as rescue and retrieval operations, autonomous as well as remote operations. an integrated mounting bracket allows
and for these applications the operation precise elevation and azimuth alignment
of a vehicle is unlikely to be completely Sensor suite to within 2º.
autonomous. For example, if you wanted The main sensor is a camera suite, from The Piranha also uses cooling to
to do unmanned construction tasks or UK company Chess Dynamics. This remove heat from the charge coupled
create a secure base in a potentially handles the optical detection so that the device (CCD) sensor. This reduces
hostile area, you could send out vehicle can follow a set of waypoints random noise in the sensor, resulting
in improved image quality, particularly
in low-light conditions. It also supports
Expanded Hi-Dynamic Range, which
The cooperative car amplifies the signal level in dark areas
and reduces it in very bright areas,
improving the visibility of objects in the
MIRA has developed a vehicle platform capable of adapting throttle and brake image.
control in response to messages received over a wireless network. Using accurate Other cameras are sourced from Sick,
positioning techniques, suitable control commands can be generated remotely to in Germany, which provides cameras for
accurately execute predefined test scenarios. industrial applications such as production
This Network Assisted Vehicle (NAV) has a number of advantages over the two line monitoring. MIRA has used these to
main alternatives for pre-crash testing. These rely on full driver control of the demonstrate that commercial off-the-shelf
vehicle, or a fully automated vehicle. imaging systems can be used as well as
The increasingly complex test requirements for advanced driver assistance ruggedised systems such as Piranha.
systems (ADAS) also means that multiple vehicles or related events need to be The engine transmission is automatic,
synchronised. which is easily adapted for autonomous
The option of using fully automated vehicles is appealing for safe and repeatable operation, although the control software
testing, but the current costs are prohibitively high because of the need for accurate does not use feedback from the
sensor systems for all parts of the vehicle. transmission to the control electronics
MIRA’s NAV has already completed system development and testing, and is being for engine braking, for example. The
used in evaluation projects such as UK Autodrive in Coventry and the connected steering, braking and acceleration are
cyber security collaboration group. all controlled by small actuators inside
the footwell so that a driver can also

64 Spring 2015 | Unmanned Systems Technology

 to the human eye, given some initial object as soon as it comes into the field
practice for the operator to get used to of view, and there are three at the front
The pinch point driving it remotely. A lot of this expertise of the vehicle to provide a wide angle of
in reducing the latency for the remote view. In remote control mode the views
on most of these operation came from the development they give appear on a three-way screen,
of the Panama platform, where operators and the same cameras are used in self-
platforms comes took some time to adapt to the control driving mode.
system because the higher latency The cameras scan the ground ahead,
down to comms caused a lag between moving the at a range of about 20 m so that the
steering wheel and Panama responding. control system doesn’t get confused,
and the data link One way to reduce the latency is allowing the vehicle to stop in response

– this is invariably by using specialist, fast-response

cameras, but on MACE 3 it is all about
to cars and people that may be crossing
its path. The overall control algorithm is

the weak point using standard, commercially available

cameras with standard definition (SD)
relatively simple: as soon as an object
crosses the field of view, the vehicle

and where it can sensors with a resolution of 640 x 480

pixels and basic digital processing, rather
prepares to stop. It tends to react
across a much shorter range, which is

get expensive than a high-definition (HD) sensor that

has a typical resolution of 1080 x 768
adjustable, typically in the 10 m range
when it is driving along, to bring the
pixels. Using SD reduces the bandwidth vehicle to a stop within 5 m of the object.
required for the comms link, and also MACE 3 also uses Lidar laser ranging
serves to reduce the bandwidth required detection as part of the suite of sensors,
use the vehicle in the traditional manner. in the autonomous control unit, allowing using a 32-scan system from Velodyne.
Mechanically they are all separate but it’s a faster response. This uses a high-power laser diode to
the software algorithm that pulls all the The cameras are used to detect an generate a beam that is separated by a
data together for the control system.
The algorithms that control MACE 3
An earlier version of MACE has been used in
are implemented in hardware in a field autonomous operation for monitoring pipelines
programmable gate array (FPGA) in the
VIM rather than in software to achieve the
necessary latency and response times.
The worst-case latency is the time it takes
for an image to be captured at a camera,
relayed to the controller in MACE 3, then
transmitted over the comms link to a
remote operator.
Reducing the latency has been one
of the MIRA development team’s key
aims, and the latency comes primarily
from upgrading the comms system
and tweaking the algorithm to achieve
an end-to-end latency of 150 ms. This
is critical for the remotely controlled
mode, to give the operator as much
time as possible to respond to the
images coming from the vehicle, but it
also determines the time available for
the FPGA to detect obstacles instead of
transmitting and receiving the data.
In remote operation, the aim is for the
latency to be low enough that the control
of the vehicle is more or less seamless

66 Spring 2015 | Unmanned Systems Technology

 MIRA MACE | Dossier

mechanical system into 32 beams that

scan the area in front of the vehicle, About MIRA
capturing any reflections in a sensor.
Lidar sensors don’t lend themselves
to military applications as they emit The Motor Industry Research Institute (MIRA) was founded in 1946 as the UK
so much laser light that they show up government’s automotive research centre. In the past 60 or so years it has built
clearly in the dark, making them an crash labs and system test capabilities, and developed new technologies such as
obvious target. They have been used one of the first production electric cars, the Enfield, in 1972. In 1982 it opened its
though for optical mapping of an area, Engine Test Laboratory, initially with six engine test cells. The facility has since been
with a remote operator driving the expanded with the addition of a further eight test cells.
vehicle, and the map is then used to MIRA has a Technology Park, a 1.75 million sq ft enterprise zone, in the Midlands
define the waypoints for autonomous region of England, and the latest company to take a bespoke facility there is Aston
operation. They are also used for range Martin, whose 15,800 sq ft building will accommodate a key part of its prototype
finding as they provide a longer range and vehicle development.
than optical cameras, detecting objects Aston Martin joins other companies who have already located at the park (which
at 50-85 m. was awarded Enterprise Zone status in 2011), including Haldex, Bosch, Norgren
All the data is fed into a ruggedised and Lockheed Martin. The aim is to house businesses from across the automotive,
laptop running a customised version of aerospace, rail, defence and associated digital technology sectors.
Linux that is optimised for the control
of the system. The critical part of the
obstruction detection algorithm kicks
in at 30 m to interface with the image the environment can and will change vehicle that the number of vectors by
detection to the VIM, triggering the throttle dramatically. This means the key is which a cyber attack could be directed
and braking algorithms simultaneously. having redundancy in the design of the on an autonomous or connected vehicle
The detection algorithm then assesses system so you can throw more variables is myriad, says MIRA. However, because
whether it is an obstacle or terrain, using at it. According to the development the MACE platforms are designed to
the image data rather than data from the team, this in turn means MACE 3 is interpret things that are unpredictable,
engine and transmission. able to make all the decisions that the they will be harder to attack.
The core of the control algorithm is a cooperative car can make but in a much The pinch point on most of these
fusion of data from the camera systems more random environment. platforms comes down to comms and
and GPS so that a series of GPS tracks Interestingly, MIRA is conducting the data link – this is invariably the weak
can be pre-loaded over a digital map, research into a road-edge detection point and is where it can get expensive,
but how it controls its speed is based on system for off-roading, the idea being as the system is only as good as the
the optical systems. As a result, MACE to take a snapshot 15 m ahead that comms, which is where MIRA spends a
3 will know how to drive that route a lot interprets a road edge. That will mean lot of time to make it work.
quicker than in remotely operated mode being able to navigate along tracks
as it has more precise data from the in a much more autonomous fashion Conclusion
cameras and so can stop faster than a rather than a pre-recorded track, and The experience of the MACE and
remote driver could. this is all down to the signal processing Panama autonomous platforms over
In the self-driving mode it is driving algorithms used. the past decade has invaluable lessons
according to the parameters of a route, for the development of the driverless
so it won’t deviate from its route but will Security car. The implementation of the control
vary its behaviour. The same approach Because MACE 3 is a military design, algorithms, fusion of camera data with
is used for the cooperative driverless system security has been built in from other sensors in a simple and efficient
car that MIRA has developed. This takes the start, and accounts for most of manner, and the integration of the
its position and navigation cues from the complexity of the design. This is comms systems are all elements that
messages sent from other cars and from especially the case with the comms have been tested in the most hostile
roadside sensors. system, which has to be agile in environments possible. The technology
For MIRA the driverless car is an frequency, have long range and be developed for these environments will
easier thing. The whole point of the secure from hacking and jamming. translate to more efficient and effective
military application is that a situation But there are so many sensors and control systems for autonomous,
can be guaranteed to go wrong, and systems that are part of an autonomous connected vehicles on the road.

Unmanned Systems Technology | Spring 2015 67

It all
Design freedoms allow new design
approaches to improve part performance
(Courtesy of Concept Laser)

adds
up
Making components using an
additive process is ideal for the
unmanned systems industry.
David Cooper explains what those
processes are and how they work

A
dditive Manufacturing (AM) the manufacturing technologies used that allow ceramic materials to be used,
is one of a number of industrially. Therefore this article will these are not generally mainstream.
widely used terms which focus on those technologies that show
refer to technologies that a readiness for applications that have Plastic AM
manufacture components rigorous engineering performance There are three primary AM technologies
on a layer-by-layer basis, typically direct requirements, such as those in the for making plastic components –
from digital models. While terms such aerospace industry. fused deposition modelling (FDM),
as 3D printing are more often used AM technologies can be readily divided stereolithography (SLA) and selective
in the wider media, it is important to into two groups: plastic and metallic. laser sintering (SLS). FDM uses a heated
draw the distinction between that and Although some solutions are available extruder head to deposit plastic filaments

68 Spring 2015 | Unmanned Systems Technology

 Additive Manufacturing | Focus

into a given shape, in a process often

likened to icing a cake. FDM machines
range from very low cost (sub-£1000)
‘home printers’ to high-end industrial
systems (more than £70,000) with
generally a corresponding difference
in quality, mechanical properties of the
deposited material and productivity.
The second technique, SLA,
consolidates material from a liquid
photopolymer resin contained in a build
vat; a UV wavelength laser then cures the
resin before submerging the cured layer
into the vat by one-layer thickness and
levelling the next layer of liquid resin over
the top, ready for the next cycle. SLS is
very similar in concept to SLA but instead
it uses a thermoplastic powder feedstock
and an infrared laser to sinter the powder
particles together.
Each technology has its own advantages
and disadvantages with respect to the
others, and choosing which is most suitable
depends largely on the manufacturing
requirement. For a one-off part, FDM is likely
to be the most economic, while for large
numbers of components, SLS can be very
cost-effective as it is the only technology
where multiple parts can be ‘nested’ or
stacked together in all three dimensions,
rather than being restricted to a footprint on
the build platform (as with FDM and SLA).
Although the design freedoms are
certainly greater compared to traditional
moulding processes, there is by no
means total design freedom – certain
manufacturing rules must still be
Bespoke instrument housing for satellite observed. For example, both FDM
application (Courtesy of CRP)
and SLA require the use of supporting
structures to build any down-facing
surfaces below a critical angle relative
to the build substrate. These structures
must then be removed either by hand
or, for FDM processes where a soluble
support material has been used, by
dissolving them from around or inside
the final component. While SLS does
not require any supports (and so
perhaps provides the greatest design
freedom), thought must still be given to
the finishing of the rough semi-sintered
surface, removal of unsintered

Unmanned Systems Technology | Spring 2015 69

Focus | Additive Manufacturing

Small UAVs can be built using functional SLS materials

(Courtesy of Graphite Additive Manufacturing) functions with one well-designed part).
The essentially low-volume, batch-based
manufacturing process also reduces the
cost of otherwise expensive plastic parts
where tooling would be required, and
allows increased design flexibility as no
moulds or tools need to be committed to
manufacture, meaning design iterations
are easily possible without a high cost.
Coupled with reasonable functional
material properties, this makes plastic AM
technologies ideal for producing small
volumes or even one-off components for
a range of unmanned vehicles, ranging
from low cost to high value.
Plastic AM technologies are also ideally
suited to small UAVs for the manufacture
of most structural parts or aero surfaces,
although polishing processes such as
powder and to thermal warpage/ such as glass, aluminium or short-strand vibro/tumbling are typically conducted
shrinkage effects, which may be an issue carbon fibres to improve properties; to smooth SLS parts. SLA parts or
in large components for example. special flame-retardant mixtures are soluble FDM materials are also often
The materials available also differ also available which fulfil aerospace FST used to make complex moulds or cores
between each process. FDM offers standards. Polyether ether ketone-based as tooling for carbon fibre parts, which
different thermoplastics, including materials are also available – although would otherwise be very difficult or
ABS, polycarbonate, polyamide, typically they require a dedicated machine impossible to manufacture.
polyphenylsulphone and polyetherimide – and offer exceptional mechanical
(PEI), providing a wide range of material properties as well as temperature and Metal AM
properties (and cost) which allows chemical resistance, but they are more Two primary technologies are available
production of anything from prototypes to difficult and costly to process. for the direct manufacture of metallic
functional end-use parts. In particular, PEI- These plastic-based AM technologies components, and while they have
based materials have a high temperature have great potential for the manufacture many differing trade names they are
and chemical resistance as well as of parts for unmanned vehicles. The fundamentally based on either laser
good Flame, Smoke and Toxicity (FST) design freedom they offer enables beam melting (LBM) or electron beam
ratings, allowing their use in demanding components to be made with lower melting (EBM) of metallic powders.
engineering and aerospace applications. weight and increased functionality Unlike SLS or SLA, where material
SLA materials also include ABS-based (for example by fulfilling two or more is either sintered or cured, LBM
polymers, often mixed as a composite
material with fillers such as ceramic
particulates for improved properties
SLS parts are capable of
in certain applications. However, the high performance, complex internal
greatest limitation with SLA is the need to geometries and integrated moving
work with a specialised liquid feedstock, parts (Courtesy of 3T RPD and the
University of Southampton)
which typically results in a much higher
material cost/kg. SLA materials were
once known for continued UV ageing
after the manufacturing process, but
modern resins have gone a long way to
improving this behaviour.
SLS materials are most usually based
on a Nylon (polyamide 12) powder, which
may also be mixed with different fillers

70 Spring 2015 | Unmanned Systems Technology

THIS IS ONE TOUGH
PIECE OF PLASTIC

METALISE IT…
3DDC specialises in the metal plating of
3D printed parts and prototypes.
If you require additional strength for your 3D printed part to produce a
lightweight tough component our expert team can help.
The metal coating can also be used for RFI shielding or improving the
cosmetic finish of the component.
We offer an unrivalled service and commitment to offer the right metal
plating solution for your project.
Quite simply, the Metalise it... process takes 3D printing to the next level.

3DDC Ltd
Unit 4B Hurst End Farm, North Crawley,
Bucks, MK16 9HS, UK
T: +44 (0)1234 391894 W: [Link]
F: +44 (0)1234 420323 E: 3ddc@[Link]

Layer upon layer What’s on the radar?

Packing a punch Advances in sense-and-avoid systems
Existing and emerging battery technologies compared Focus on additive manufacturing Vision of
the future Signals from space
Signals intelligence Catch some rays Key advances in image GNSS rises to the accuracy challenge
Overlook early integration of antennas at your peril Spotlight on solar power sensor technology

Match points
How to pick the right
electric motor for a UAV

Follow that Cab Multi-tasking monitor Rugged and reliable Who wants data?
What makes the Alpha 800 helicopter
Navya aims to be the first on the stand with Robot Aviation looks to fill a gap in the UAV XOcean’s XO-450 hybrid USV collects and delivers
so versatile and dependable?
a driverless taxi – how was it developed? market with its all-weather SkyRobot FX450
UST 19 : APRIL/MAY 2018 UST 20 : JUNE/JULY 2018 UST 21 : AUG/SEPT 2018 UST 22 : OCT/NOV 2018

Read all back issues online Read all back issues online Read all back issues online
UK £15, USA $30, EUROPE e22 UK £15, USA $30, EUROPE e22 [Link] UK £15, USA $30, EUROPE e22 [Link] UK £15, USA $30, EUROPE e22 [Link]

Direct Laser Metal Sintering

Plug-free power Then there were five

Materials
What 5G will mean for unmanned vehicle control
Wireless charging technologies
explained and compared
Strong attachments
Way points Why you need to latch on
to advances in connectors
Advances in MEMS
and what they mean Blades of class
for navigation Take care How to match a propeller to a UAV
Focus on maintenance
This year’s modelling

• Stainless Steel 316L and 17-4PH

Chain of command The latest in simulation and testing kit
Advances in ground control systems

Burning ambition

• H13 Tool Steel

Plug and play A seamless transition Under the surface Tecdron’s TC800-FF UGV has been
How the Wingcopter uses only one propulsion Development details of the iXblue DriX USV designed to fight fires – at least for now
How Milrem developed its Multiscope commercial UGV system to shift between all flight modes and its underwater payload gondola
UST 23 : DEC/JAN 2019 UST 24 : FEB/MAR 2019 UST 25 : APRIL/MAY 2019 UST 26 : JUNE/JULY 2019

Read all back issues online Read all back issues online Read all back issues online Read all back issues online
[Link] [Link] [Link] [Link]

• Aluminium Al-Si-12
UK £15, USA $30, EUROPE e22 UK £15, USA $30, EUROPE e22 UK £15, USA $30, EUROPE e22 UK £15, USA $30, EUROPE e22

Back issues
• Cobalt-Chrome (ASTM75)
• Titanium
Layer thickness: 0.020-0.080mm

available
Accuracy: +/-0.050 mm

Laser Sintering – Plastic Parts

in print
Materials
• Polyamide (PA)
• Glass filled polyamide (PA-GF)
• Alumide (PA-Alufilled)
Online:
Layer [Link]
thickness: 0.1mm
Telephone:
Accuracy: +44
+/-0.3 % 1934 713957
Email: freya@[Link]
Spires-Tec GmbH
Telephone UK: 0871 237 9007 Fax: 0870 321 1007
Germany: +49 (0) 179 312 3515
Email: info@[Link]
New software tools allow the
generation of highly complex The differences in layer thickness and Electrical Discharge Machining (EDM).
structures (Courtesy of Materialise) pre-heating temperatures mean that EBM While the ability to manufacture
and LBM have their own advantages. complex and unconventional geometries
EBM parts experience lower residual is undoubtedly a significant advantage for
stresses owing to smaller temperature metallic AM technologies, such complex
differentials, and faster build times geometry makes post-processing far
(and so higher productivity) due to more difficult. For example, internal
thicker layers. Conversely, LBM systems cavities that allow much lighter and stiffer
typically provide higher resolution and component designs cannot be easily
so improved surface finish, although with finished to improve surface roughness,
more powerful lasers newer systems and so remain a liability with regard to
now build in thicker layers, improving fatigue crack propagation.
productivity at the expense of resolution, The ability to manufacture complex lattice
while EBM systems have recently seen structures is also a fantastic advantage
and EBM are fundamentally welding an improvement in surface finish. So the but which brings its own set of problems.
processes. Each successive layer two technologies are perhaps beginning New software tools allow such structures
of metallic powder is welded to the to converge, although they still maintain to be generated relatively easily, but
preceding layer in a fully molten welding distinct differences. characterising the mechanical properties
process – long gone are the days of A particular advantage of these of such structures can be difficult, making
sintered green pre-forms which are then metallic AM processes lies in their ability predictive analysis and design unreliable.
infiltrated by bronze materials. Although to produce a complex shape such as Inspecting such a complex internal
they have their own advantages and would more usually be investment cast, geometry is virtually impossible by any
challenges, as with plastic AM techniques but without tooling and with material process other than CT scanning, which
it should be remembered that they are properties closer to those of wrought increases cost. Indeed, non-destructive
simply welding – no black magic occurs material; however, there are important testing of additively manufactured parts
– and as a welding process, a lot of the design limitations. Most parts require is an area of ongoing research and a
same challenges are faced. supporting structures where downfacing substantial topic in its own right.
For example ‘un-weldable’ materials surfaces are too close to horizontal For some time now, the greatest
cannot be readily processed by LBM/ (usually below 400 to the build platform, challenge facing metallic AM has been
EBM, and residual stresses or distortion depending on the material). These to convince potential end-users of the
are perhaps an even greater issue in AM supports must be built from the same mechanical properties that can reliably be
processes than in conventional welding, powder bed material, so they are usually expected from additively manufactured
as they repeat-weld several thousand welded to the part in question and so material. Recent years have seen a
times to form the end component can be quite difficult to remove, requiring general acceptance of the properties
(albeit on a smaller scale compared to intensive manual fettling or further post- that can be achieved, but the industry
conventional arc-welding processes). processing such as CNC machining or must now strive to ensure a reliable and
Typically, metallic AM processes use a
powder feedstock, swept by a re-coating
LBM Ti-6Al-4V lattice samples used to determine bulk
arm or blade onto a substrate in layers properties for a potential unmanned space application
ranging from 20-100 µm depending on (Courtesy of EOS as part of Innovate UK’s LIGHT project)
the process in question. In the case of
EBM, the substrate and build chamber
are pre-heated to about 750 C and
use thicker, 70-100 µm layers, while
LBM systems more typically heat the
substrate to only 35 C – or up to 200
C at most – and process material in
layers of 20-60 µm. EBM systems must
naturally operate under a vacuum, while
LBM systems use an inert atmosphere
of nitrogen or argon depending on the
material being processed.

72 Spring 2015 | Unmanned Systems Technology

 Additive Manufacturing | Focus

Quality control of the raw powder is an important issue

repeatable set of mechanical properties for the AM industry (Courtesy of Innovate 2 Make)
for a given component – quite a difficult
task when one considers the number of
variables involved in the process.
For example, tight controls are required
on the feedstock powder, ensuring not
only a chemical composition that is
within specification but is also free from
detrimental elements (for example, the
fatigue properties of titanium alloys are
adversely affected by a high oxygen
content, either from the feedstock powder
or during processing). The particle size
distribution (PSD) and morphology of the
feedstock powder are also important, as
they affect the packing density of powder
in a layer and the interaction of the energy chamber, how it will be removed from barriers to manufacturing parts for
beam with the new layer of powder, and the substrate (for example, by wire EDM UAVs or satellites, many of the barriers
hence the resulting part density. Humidity or bandsaw) and how each potential to wider adoption of these processes
in the feedstock powder can also affect its orientation might impact the number of are regulatory. As a relatively new
flowability and so re-coating behaviour, as supports required, how easy they are to process, the relevant understanding and
well as the weld behaviour. remove, the resulting surface quality and knowledge base of metal AM is only
All of these factors must be controlled the overall economics of the build. now maturing, and so manufacturing
not only for the first build but for the life Any supports are likely to increase the standards and best practices have yet to
of the powder batch, as it is sieved and time and cost to manufacture the part, the be fully adopted or even written.
re-cycled for multiple builds. Each build material waste and the post-processing
influences the PSD and composition, and costs. Supports are not only required to Applications
so affects the final part quality, and once a allow downfacing surfaces to be built, but Metallic AM components are ideally
powder batch falls out of specification then also to conduct heat out of the part during suited to various applications in
it may have an influence on the properties the build process, and to restrain the unmanned vehicles, for example in
that can be expected of the final product. part against movement due to residual manufacturing stressed structural parts.
While the feedstock powder contributes stresses, particularly in LBM processes. The design freedoms offered allow a
to the part’s quality, there is also an The relative importance of the support’s more efficient part design that can reduce
interaction with the process parameters different functions can vary depending weight and improve functionality by
used – laser power, scanning speed, on the material. For example, maraging incorporating functions once performed
hatching distance, layer thickness and steel tends to need more attention to by several parts into a single component.
so on – which must be tuned to deliver thermal management during a build, More efficient heat exchangers are an
acceptable properties of the part for a while titanium alloys need more stress ideal application, as are any fluid flow
given specification of powder and within management, so support design must components where internal losses may
a given vendor’s machine. A ‘good’ focus on minimising stress raisers to be minimised. The ability to manufacture
powder in one machine may not be prevent cracking during the build and on hard-to-machine alloys such as Inconel
adequate in another or with a different set restraining the part from any movement 718 or Ti-6Al-4V as near-net shape
of laser parameters, for example. that can lead to inaccurate parts. products is also advantageous for parts
Not only are the manufacturing Despite the complex issues that must such as engine exhausts or moderately
parameters important, so too is the be overcome to manufacture a high- sized structural parts, which can be
configuration of parts for the process and quality metal part though, the ultimate complex to fabricate or expensive to
in the build chamber. To obtain a good advantages of metal AM processes machine using conventional methods.
final part, it is important to design for the in reducing part weight, simplifying The manufacture of aluminium parts is
process as well as design an effective assemblies and reducing manufacturing less common, however, as currently only
support strategy in cases where supports time/cost in hard-to-machine materials die casting alloys are readily available
are required. Consideration must be can be highly worthwhile. on AM platforms, and there is only any
given to the orientation of the part in the While there are appreciable technical real benefit if a complex and hard-

Unmanned Systems Technology | Spring 2015 73

Focus | Additive Manufacturing

Some examples of
additive manufacturing suppliers
Belgium 3D Systems (CRDM)
LayerWise +44 (0)8450 514900 [Link]
+32 16 94 64 00 [Link] 3T RPD Ltd
Materialise +44 (0)1635 580284 [Link]
+32 16 39 66 11 [Link] ARRK Europe
+44 (0)1452 727700 [Link]
France CA Models
Poly-Shape +44 (0)1786 464434 [Link]
+33 164 851 310 [Link] Croft Additive Manufacturing
+44 (0)1925 767 070 [Link]
Germany EOS
Alphaform +44 (0)1926 623107 [Link]
+49 89 905 00 289 [Link] I2M
CITIM +44 (0)7850 631 299 [Link]
+49 39 203 51 060 [Link] Laser Prototypes
Concept Laser +44 (0)28 9070 6940 [Link]
+49 95 719 49 238 [Link] Laser Lines
EOS +44 (0)1295 672500 [Link]
+49 89 893 36 134 [Link] Performance Engineered Solutions
ExOne +44 (0)114 321 6375 [Link]
+49 82 174 760 [Link] Renishaw
Spires-Tec +44 (0)1785 815651 [Link]
+49 157 551 47671 -
USA
Italy 3D Systems
CRP +1 978 494 8200 [Link]
+39 059 821 135 [Link] / [Link] CRP USA
ProtoService +1 704 660 0258 [Link]
+39 052 540 1281 [Link] AD Cytec
Sinthesi Engineering +1 973 357 3100 [Link]
+39 331 556 557 [Link] ExOne
+1 724 863 9663 [Link]
SWEDEN Incodema
Arcam AB +1 607 227 7070 [Link]
+46 31 710 32 00 [Link] Paramount Industries
+1 215 757 9611 [Link]
UK Stratasys
3DDC +1 952 937 3000 [Link]
+44 (0)1234 391894 [Link]

to-machine geometry can be achieved. are beginning to show their potential Acknowledgements
Aerospace alloys such as 6xxx or 7xxx for manufacturing in low to medium The author would like to thank Dr Martin
series materials are not widely used, as volumes. However, it is important that McMahon, Sue Burnip and Lynn Painter
they are more difficult to process and such components make sense for the of 3T RPD, Daniel Hund of Concept
their low cost and ease of machining manufacturing technology. Value such Laser, Dan Fleetcroft of Performance
reduces the drivers to develop such as reduced weight, improved design Engineered Solutions, Kevin Lambourne
materials when compared with titanium flexibility and manufacturing agility and Keith Haynes of Graphite Additive
or nickel alloy families. must often be added to justify the usual Manufacturing, Pieter Vos and Stijn De
increase in manufacturing cost, for metal Rijck of Materialise, Francesca Cuoghi
Conclusions parts in particular. of CRP, Andy Cheney of Magna Parva,
Both plastic and metal AM technologies All these factors would appear to be Stuart Jackson of EOS, Ian Campbell
have matured substantially in the an ideal fit for many components in and Mike Kelly of Innovate 2 Make and
past few years, from fundamentally the unmanned systems industry, where the project partners of the Innovate
laboratory-derived machines capable manufacture of only a few high-value UK-supported LIGHT project for their
of producing one-off prototypes to items may be required, and weight, invaluable insight and assistance in the
reasonably industrialised processes that space or function are at a premium. research for this article.

74 Spring 2015 | Unmanned Systems Technology

UUV fleet of craft known as gliders, from
the National Oceanography Centre

Torpedo run
Torpedo-shaped autonomous underwater vehicles
are taking over the oceans. Nick Flaherty
reports on developments in the technology

S
ubmarine designers across surveying the sea floor, and it has also ships, aircraft and helicopters that already
the world have for some evolved to allow fleets of smaller systems launch torpedoes, giving operators a
time been using unmanned to be created that can be connected lot more flexibility in how they store and
underwater systems for together using underwater wireless links. launch the systems. It also creates less
military applications such as It is not just the size of these systems drag when they’re in the water which,
mine clearance, exploiting the benefits that has changed but their shape, with new and more efficient propulsion
of their smaller size and cost and from remotely piloted submersibles systems, allows them to be used for
autonomy to give a greater range than in the 1990s to systems that are fully missions lasting for months at a time.
remotely operated craft. In recent years autonomous and that look like torpedoes. Then, at the end of a mission, the craft’s
though the technology has crossed over This new shape or ‘form factor’ makes it programming tells it to return to a base
into commercial applications such as easier to transport and launch them from or support vessel to recharge its batteries

76 Spring 2015 | Unmanned Systems Technology

 Marine UUVs | Insight

Launching the Kongsberg SeaGlider

The new shape

makes it easier
to transport and
launch them from
ships, aircraft and
helicopters that
already launch
torpedos, giving
more flexibility

and download the data that has been

collected.
However, communications can be a
problem as the vessels have to surface
in order to connect to an operator for
location data or to upload images.
Being able to stay underwater for longer,
store images and return to a base ship
or location therefore changes the way
surveys and other undersea applications
are designed.
With this in mind a new £4m Centre
for Maritime Intelligent Systems (CMIS)
in Portsdown Technology Park in
Southampton, UK, is developing technology
for autonomous underwater vehicles Integration Testbed (MUVIT) into the CMIS of unmanned AUVs could improve the
(AUVs) and unmanned underwater vehicles environment for the rapid development accuracy of their positioning by sharing
(UUVs), which can be autonomous but can of manned and unmanned concepts information over a hydro-acoustic data
also be remotely controlled. into sea-trialled systems using in-the- link that uses sound rather than radio
The centre is setting up a simulation loop testing of hardware and software. waves. This is ideal for underwater, where
system called a synthetic environment to Blue Bear’s autonomy and automation traditional radio systems and the satellite
develop proof-of-concept designs of new software is already used on AUVs, and navigation network don’t work and an
vessels as part of a maritime autonomous the company sees it being used for autonomous system has to use an inertial
systems demonstrator programme. underwater designs at CMIS. navigation system.
Alongside £1m of funding from the UK’s The technology has been demonstrated
Ministry of Defence, there is also £1m Swarming using AUVs as a model, and Plextek is
available for small companies to develop It’s not just new craft that are being now investigating how the linked vessels
autonomous technologies in collaboration developed though. A new technique can share information among themselves
with the centre. One of these, Blue called Ghia, developed by UK to maintain highly accurate estimates
Bear, has integrated its Multiple Vehicle consultancy Plextek, shows how a swarm of position and time. To estimate

Unmanned Systems Technology | Spring 2015 77

Insight | Marine UUVs

Kongsberg’s Hugin autonomous underwater vehicle,

which is targeted at subsea surveys down to 4500 m

If multiple AUVs
were operating
then only one
would need to
resurface to
improve the time
and position
estimates for all
the nearby AUVs

years, with a 3D accelerometer within

four years.
Another approach for handling fleets
of AUVs is being developed by SeeByte,
in Edinburgh, which is working with the
UK’s National Oceanography Centre
(NOC) to develop control software for
their position relative to a known starting condition of the electronic components in the Marine Autonomous and Robotic
point, AUVs use gyroscopes and doppler an autonomous system. Systems (MARS) group in the NOC.
velocity logs that keep a record of the By sharing information, the AUVs MARS has recently acquired a large fleet
speed and direction of the craft that is reduce error and drift while minimising of torpedo-style AUVs that are powered
independent of the onboard sensors, their dependence on the GNSS network. If by ocean currents. Commonly called
but to do so they need highly accurate multiple AUVs were operating, they would gliders, they are made by Teledyne and
clocks. To have an accurate position, be able to surface for a GNSS fix less Kongsberg (Norway’s major defence
AUVs typically have to surface at regular often, as only one would need to resurface contractor), but each type has its own
intervals in order to obtain a fix of their to improve the time and position estimates interface and processing tools.
position from a GNSS satellite such as for all the other nearby AUVs. SeeByte is modifying its SeeTrack
GPS. This reduces the area an AUV can This approach could use quantum software suite to provide a common
survey during a given mission. clock technology to provide more interface for all the gliders, regardless
Plextek’s investigations have shown accurate timing and navigation systems of the manufacturer. By combining all
that an AUV can substantially improve its for underwater systems. To this end, the mission plans and monitoring in a
position and time estimate by combining researchers at the UK’s National Physical single workstation, the NOC will be able
its own time estimates with those Laboratory are using hollow fibres filled to manage larger fleets of gliders without
from a number of nearby craft, via an with caesium atoms to create miniature putting additional strain on the operators,
appropriate Kalman filter, an algorithm atomic clocks that can provide high and data from these missions will be
that looks at a series of measurements accuracy for navigation systems in consolidated on a shared database,
over time to produce better estimates of small, light packages with low power. A allowing data to be easily analysed.
an unknown variable than those based commercial chipscale one-dimensional The Kongsberg SeaGlider used in
on a single independent measurement. (1D) quantum accelerometer for inertial the NOC fleet uses a novel propulsion
Variants of it are used for monitoring the navigation is expected in the next two system. Rather than an electrically

78 Spring 2015 | Unmanned Systems Technology

Cost effective
composite solutions
Founded in 2008 our highly skilled team of composite
specialists have over 65 years experience working at
the top end of global motorsport.

Following substantial recent investment and relocation

to our new composite facility, we now have real-time
capacity to expand into the Unmanned sector.

• Composite Prototyping
• Composite Repair
• Composite Manufacture
• Composite Training
• Trackside support

Please contact us with your requirement

and we’ll be delighted to help.

Composite Technologies Ltd Plot 2, Oakfield Industrial

Estate, Eynsham, Oxon, OX29 4TH, UK
T: +44 (0)1865 880887 E: info@[Link]
[Link]
driven propeller, it uses wings and small Kongsberg’s SeaGlider uses wings and small of 1000-4500 m using a more traditional
changes in buoyancy to achieve forward motion
changes in buoyancy to achieve forward electric motor and propeller, with different
motion; its pitch and roll are controlled payloads for applications such as high-
using adjustable ballast that is actually resolution, high-speed seabed mapping,
the vehicle’s battery, which powers the imaging and site inspection as well as
sensor systems. pipeline and subsea structure inspection.
It moves through the water in a It can also be used for search operations.
sawtooth-like pattern and surfaces The Hugin comes in three variants –
regularly to determine its position. the 1000, 3000 and 4500. The 1000’s
Navigation is achieved using a larger diameter of 75 cm compared to
combination of GNSS fixes while it is the 30 cm of the SeaGlider, as well as its
on the surface and internal sensors that modular design, allows larger sensors
monitor its heading, depth and attitude to be used such as a high-resolution
during dives. The propulsion uses very The propulsion interferometric synthetic aperture sonar
little energy so that it can be deployed that can be used at depths of 1000 m.
for up to ten months at a time, compared uses very little The vessel also supports a multi-beam
with a day or two with other AUVs. While echo sounder, sidescan sonar, still image
its top speed is low, at half-a-knot (25 energy so that it camera with HD resolution and 10x
cm/s), its long endurance allows it to zoom, a turbidity sensor to measure the
travel thousands of kilometres in a single can be deployed ‘cloudiness’ and quality of the water and
deployment. At 1.8 m long and weighing a methane sensor for detecting leaks
52 kg for the standard version, it can for up to ten from pipelines. A key characteristic of
carry a 2 kg payload and be deployed
easily from small vessels to work at months at a time, the Hugin payload system is that all its
sensors can operate together, and the
depths down to 1000 m.
compared with a data is time-stamped at source to allow
accurate location-tagging of the data.
Payloads
The SeaGlider system is modular,
day or two with Payload options
and can accommodate many kinds of
sensors depending on the application.
other AUVs The vehicle body is modular and made
from carbon fibre laminate material
It can be used for physical, chemical covering a composite known as syntactic
or biological oceanography and foam to manage the buoyancy, and this
environmental monitoring, but with is deliberately extremely stiff to make sure
its long mission time it can also be operators around the world through a there is accurate orientation between
used as a data gateway to receive relationship with iRobot in Massachusetts. the payload sensors and the navigation
information from other systems, and act It is now produced by Kongsberg at its system. The Hugin 1000 consists of three
as a navigation aid. With its low-impact Lynnwood facility in the US, not far from main sections to make it easier to adapt
propulsion system it is also being used to the University of Washington, under a new the vehicle to different payloads and
actively and passively monitor animals in technology licence. As well as building battery configurations, while the Hugin
the sea via the noises they make, and to new SeaGlider vehicles, the centre 3000 and 4500 have an even larger
profile the different ocean currents. A 2 m will also refurbish or upgrade existing diameter, of 1 m, and a single payload
version called Ogive doubles the payload systems, and plans to start shipping the bay to make it easier to integrate larger
to 4 kg, so it can support systems for new SeaGliders in the first half of 2015. sensors. The pressure containers used to
profiling currents with echo sounders and Gliders aren’t the only AUVs to use the protect the sensors are made from either
dissolved oxygen sensors. torpedo shape. Kongsberg also supports titanium or sea-water resistant aluminium,
SeaGlider was originally developed in two other AUV platforms – Hugin, which depending on depth requirements.
the 1990s at the University of Washington was developed in-house; and Remus, The Hugins are also equipped with a
in Seattle with funding from the US which was acquired when Kongsberg hydro-acoustic comms link that provides
Office of Naval Research and, as of 2012, bought a US company called Hydroid, good connections in the range of tens
more than 190 Seaglider AUVs have both using traditional propulsion systems. to a few thousand bits per second
been delivered to research institutes, Hugin is being targeted at remote over several kilometres with carrier
naval organisations and commercial subsea surveys as it can travel to depths frequencies of around 25 kHz, although

80 Spring 2015 | Unmanned Systems Technology

 Marine UUVs | Insight

changes depending on the data from

its sensors and a fixed set to rules that
determine what actions to take as a result
of that data. It can be used to solve a
specific task within an area of up to 20
sq km and up to 500 m depth, travelling
at speeds of up to 20 knots. To do this
requires obstacle avoidance algorithms
and highly accurate positioning, so the
AUV62 uses dual sonar systems on its
sides and high-resolution sonar to fill the
gaps, giving a positioning accuracy of
better than 4 cm2.

Modular design
It can also communicate with its parent
vessel either while it’s surfaced, using a
wireless local network or satellite comms
link, or submerged via a hydro-acoustic
link. There is also a nose-mounted
camera for object identification as it
travels around, and it can capture data
for 3D images of objects for commercial
imaging applications.
Saab’s AUV62 is fully autonomous
The AUV62 varies from 4 to 10 m
long depending on the payload, and is
53 cm in diameter. Its modular design,
data rates of up to 25 kbit/s are possible both the electric propulsion system and with swappable payloads, enables
with shorter links. This allows a tracking the AUV’s sensors down to depths of it to be used for missions such as
vessel to receive payload data from the 4500 m. The standard rechargeable seafloor profiling, detecting and tracking
AUV in real time, and can be used as a lithium-ion battery supports a mission subsurface installations such as cables
quality measure to make sure a survey is time of 75 hours at 4 knots, and is made and pipelines, and environmental surveys
covering all the necessary regions of the from commercial battery cells with built-in All of this is a long way from the early
sea floor, for example. safety systems to prevent overcharging submersibles, and demonstrates the
Offshore services group DOF Subsea, or fast discharging. This is used down growing importance of commercial
in Norway, for example is using to 1000 m, but for the 3000 and 4500 applications for AUV systems, both within
Hugin for tasks such as geophysical versions of Hugin, a non-rechargeable the craft and the system interfaces. While
surveys, inspecting subsea facilities aluminium oxygen semi-fuel cell (ALHP mine clearance is still an important
and environmental inspections, as well FC) is used. This is a specially developed application driving r&d, increasing
as quick inspection of pipelines, using derivative of a fuel cell that generates the range and depth of an AUV with
sidescan sonar, multi-beam sonar and power from aluminium and oxygen smaller form factors and more efficient
high-resolution still and video cameras. stored in the cell, and the largest of these propulsion is allowing the technology
Software developed over the past three can support 100 hours of use at 4 knots to be used for pipeline monitoring and
years by Kongsberg allows Hugin to and down to 4500 m. oceanographic tasks. This is increasing
follow as close as 3 m directly above Saab, in Sweden, also uses the torpedo the volume of systems in the market
the pipeline, something that has not shape and traditional propulsion for its and providing more momentum for AUV
been possible previously with standard latest autonomous underwater craft, the technology development.
positioning sensors. This gives more AUV62, which starts shipping later in 2015. With new technology centres such as the
effective monitoring of the pipeline It is fully autonomous and operates in one CMIS being set up specifically to develop
without needing to have an operator of two modes – it can be programmed autonomous craft for marine applications,
guiding the AUV. to follow a route of predefined waypoints, the focus is increasing on AUV
Battery technology is vital for powering or in a rule-based mode where the route technologies for the next few years.

Unmanned Systems Technology | Spring 2015 81

PS | Wankel engines

“ Now,
here’s a
thing ”
O
ne of my strongest Over the years a number of
memories of Le Mans
1990 [writes Ian Bamsey]
At the moment, motorcycle manufacturers looked
seriously at the Wankel concept, but it
is of the post-race
voluntary work I undertook
around the was Norton primarily which exploited it
on road and track. Sadly, the company
at the behest of Mazda’s PR team. Let’s
face it, who better qualified than members
world there then fell apart, and although it was
revived in 2008, the current incarnation
of the media to take time out to help them
dispose of all that surplus-to-requirements
are at least five has no apparent interest in a Wankel-
type engine. Others have carried the
celebratory champagne?
No such luck the following year though,
companies concept forward though.
In fact, at the moment, around the
as the historic win for the rotary engine
meant that all the reserve bottles were
offering rotary world there are at least five companies
offering rotary engines specifically for
(quite rightly) rushed over to the pits,
to reward those whose hard work had
engines for use in UAVs, the creation of some of
which was directly inspired by Norton’s
made it possible.
That triumph by Mazda remains the use in UAVs experiences. In a UAV, power-to-weight
ratio is a crucial consideration, and this
pinnacle of achievement of the Wankel- is where the Wankel concept scores,
type engine in motor racing. On the together with its multi-fuel potential
road, the Japanese company kept the and smooth running, as UAVs tend to
concept alive for the best part of 50 lone championing of the rotary engine be bristling with electronics that need
years, but in recent times it was defeated was a drop in the ocean. Who knows protection from vibration.
by emissions regulations – or at least by what would have been the case had that The point is that Felix Wankel’s very
the understandable failure of that engine development scenario been reversed? clever innovation hasn’t gone away. It has
type to meet changing requirements The rotary engine has certain inherent found its niche, energetically branching
without a huge amount of additional advantages, not least that it is far simpler, out into the world of unmanned systems
development cost. smaller and lighter than a conventional technology. It follows then that, in due
Understandable because the four-stroke of comparable output. The course and in future issues of this
conventional four-stroke internal comparison with two-strokes is less magazine, you will be learning about
combustion engine is in its second clear-cut, yet the rotary engine has been the state of the art in UAV rotary engine
century of evolution, with the vast majority impressive in the world of motorcycle technology. Of course, that technology
of the resources of a massively powerful racing, highlighted by Norton using one hasn’t stood still since the days of Mazda
automotive powertrain world having long to win the Senior TT on the Isle of Man and Norton, and modern developments
been dedicated to it. By contrast, Mazda’s in 1992. are fascinating, as we will reveal.

82 Spring 2015 | Unmanned Systems Technology

Subscribe
today.
The a
ll-seei
agicnt
State-o ng e e
Driver rt vision sensoy
f-the-a
o -s t rokeAVm werpla
tw p o
less 4 rs for u
Hirth’s Inside a hi-te
ch U nmann
M IRA’s
W D ed cra
temuses
autono ft
ng sloyps
mous L
a t i Danie and Ro
e r
e opalysing the deve
e nt iss lso v e
the ne n diese
m
Inside r
l -tim
Rea An w gene
rati
l
on

Scion
LAU N
ISSU H
C SA-40
Winte E
r ’14
Autono
mous h 0elicopte
r techn
ology

in ISSU
Pengu ossier Sprin E 2
g ’15
UAV D
aled
rm reve
reaking platfo
cord-b
UST 02

of a re
: SPRI NG
2015
Secrets
co m
t-media.
[Link]
MB ER 2014 [Link]
UE : NO
VE
stechno
LAUNCH
ISS dsystem 2014
20:53
manne 01/10/
[Link] UK £15,
USA $3
0, EU RO
PE e22

PE e22
0, EU RO
USA $3
UK £15,
[Link]
manne
dsystem
[Link]
3 stechno
T Cover [Link]
01 US
m [Link]
ghpowe
rmedia
.com

4 ways to subscribe Subscription prices

to guarantee a personal copy 1 year subscription: The next 4 issues
1) Online at [Link] UK £60 £50
2) Telephone +44 (0)1934 713957 Europe £72 £60 (~72e)
3) Fax to +44 (0)208 497 2102 USA/Canada £75 £62.50 ($100)
4) Email chris@[Link] Rest of World £78 £65

To enquire about advertising with us, email simon@[Link]

High Power Media Ltd, Whitfield House, Cheddar Road, Wedmore, Somerset, BS28 4EJ, UK.
Tel: +44 (0)1934 713957 Fax: +44 (0)208 497 2102 E-mail: chris@[Link] [Link]