A Multi-UAVs Communication Network Simulation

Platform using OPNET Modeler

Qiang Liu* Honggang Wang Yantao Sun Tingting Han
Laboratory of Transportation Electrical and Computer Laboratory of Transportation Laboratory of Transportation
Data Analysis and Mining Engineering Department Data Analysis and Mining Data Analysis and Mining
Beijing Jiaotong University University of Massachusetts, Beijing Jiaotong University Beijing Jiaotong University
Beijing, China Dartmouth Beijing, China Beijing, China
[email protected] North Dartmouth, United States [email protected] [email protected]
[email protected]

Abstract—Multi-UAVs communication network is becoming a a multi-UAVs network simulation. In Section III, we introduce
research hot topic because of its robustness, flexibility, and the implementation of multi-UAVs network in four OPNET
scalability. On one hand, there is a strong need to validate multi- modeling domains. In Section IV, three simulation experiment
UAV network protocols especially when they become more and cases are shown. Finally, in Section V, the characters and usages
more complex. On the other hand, building real multi-UAVs of the simulation platform are concluded.
networks or hardware testbeds is expensive and lack of scalability.
In this paper, we developed a multi-UAVs communication II. MULTI-UAVS NETWORK SIMULATION FACTORS
network simulation platform using OPNET Modeler. First, the
key factors of multi-UAVs network simulation are studied. Then Before building a multi-UAV network simulation platform,
the implementation of multi-UAVs network simulation platform it is necessary to analyze and summarize some important factors.
is introduced. Thirdly, we performed case studies including A. Network Topology
network performance, 3D visualization demonstration, and real
applications. The results show the simulation software is a During recent years, the UAV network has gradually
potentially effective tool that can support the multi-UAVs evolved from simple large single drone air-ground network to
communication network research. complex miniaturized multi-UAVs air-ground heterogeneous
network. As shown in Figure 1, a typical multi-UAVs network
Keywords—UAV, Network, Simulation, MAC, Routing, consists of ground control stations, terrestrial communication
Modeling, 3D Demonstration, Testbed units, and multi-level fly ad-hoc networks (FANETs).
I. INTRODUCTION Ralays (High Altitude)
An interesting application domain of the Tactile Internet is
the use of multi-Unmanned Aerial Vehicles (multi-UAVs).
Fields of application include the logistics service sector,
inventory tasks in forestry and agriculture and surveillance tasks
in firefighting, especially in complex environments. In the FANET (Low Altitude)
research filed of multi-UAVs networking technology, a number
of ideas, algorithms, and protocols need to be tested, validated
or improved [1][2]. Since the multi-UAVs network has the
characteristics of large scale, various types and complex logic,
the traditional theoretical analysis methods are difficult to meet
the needs, and the hardware testbeds are expensive, difficult to
scale, limited by experimental field, etc.
Ground
Software modeling and simulation technology is an Station Terrestrial Units
important development trend in the field of multi-UAVs
Fig. 1. Multi-UAVs Network Topology
network research because of its strong flexibility, scalability and
interoperability. There exist several popular simulators, such as
¾ Large number of mobile nodes in different areas: The
OPNET, NS-2, NS-3, QualNet and OMNet++. Some
terrestrial communication network can be divided into
comparisons are performed in [3], [4], [5] and [6]. Considering
multiple areas: vehicles, pedestrians, sensors and other
the factors of simulation efficiency, model library, graphical
communication equipment. Although broadband wireless
user interface, documents supporting, network visualization,
communication technology can satisfy most of the
and external interface, we choose OPNET Modeler as the base
communication applications in the same area, those cross-
simulation platform of multi-UAVs communication network.
area communication traffics still need UAVs to relay and
The organization of rest of the paper is as follows: Section forward them.
II, we analyze the important factors need to be used in modeling

978-1-7281-5089-5/20/$31.00 ©2020 IEEE

Authorized licensed use limited to: East Carolina University. Downloaded on July 29,2020 at 12:59:54 UTC from IEEE Xplore. Restrictions apply.
¾ Each single area has a group of low altitude UAVs: These 4) Transport Layer: The application traffic flows have
UAVs fly at hundreds of meters high and they can be in sight different transport requirements such as reliability, timing, and
of all the terrestrial units in the single area. They connect bandwidth. Based on these requirements, we need to configure
each other by forming a Flying Ad-Hoc Network (FANET). different transport layer protocols for different applications. For
This enables multi-hop transmissions between any pair of
example, TCP for file transfer, UDP for voice and RSTP for
terrestrial units in the area using the UAV as a router.
video streaming.
¾ High Altitude Aerial Mobile Backbone: Low altitude 5) Application Layer: Multi-UAVs communication
FANETs might become separated and end up in areas that network can be used in many civil applications due to their low
are geographically isolated. Multiple high altitude UAVs maintenance cost, high-mobility and ability to hover [10]. From
covering the entire area of operation canmaintain line-of- the perspective of network communication, the civil application
sight connectivity with each other. Thus, high altitude UAVs
of multi-UAVs network can be generally related to delivery,
form a mobile backbone that interconnects low altitude
FANETs in different areas and ground stations far away. sensing, monitoring and relaying [11].

B. Protocols Stack C. Mobility

Multi-UAVs network has become a complex system. Before Mobility is one of the most important factors in multi-UAVs
building a multi-UAVs network simulation platform, we need to networks. It reflects the cooperation and collaboration between
examine the key protocols and working parameters. Like other the UAVs in FANETs. A mobility model can characterize the
wireless networks, this information can be organized into OSI movements of UAVs with respect to their location, velocity, and
layered reference protocol stack [1]. direction over a period of time. FANETs mobility models need
to consider the changes of UAVs’ direction and moving speed
'VVROIGZOUT2G_KX *KRO\KX_9KTYOTM3UTOZUXOTM8KRG_OTM hastily at one time. Due to this reason, many researchers
:XGTYVUXZ2G_KX :)6;*68:96 consider a more realistic FANETs mobility model in multiple
9ZGZOI 2)'*32.8
multi-UAVs network scenarios [12].
6XUGIZO\K 5298*9*<,98(':3'4 III. MULTI-UAVS COMMUNICATION NETWORK
4KZ]UXQ2G_KX 5T*KSGTJ '5*<*98 SIMULATION WITH OPNET MODELER
._HXOJ @86:58' In this section, we introduce how to build a typical multi-
-KUMXGVNOI -.--8--*9:8*3*: UAVs simulation platform using OPNET. The overall flow of
)UTZKTZOUT(GYKJ6XUZUIURY 3')'=*,8' the simulation methodology is can be divided into four modeling
domains, which are Network Domain, Node Domain, Process
)UTZKTZOUT,XKK6XUZUIURY :*3',*3'
*GZGROTQ2G_KX Domain, and External System Domain [15].
._HXOJ6XUZUIURY ,33')))3')),3')2*3')

)R[YZKX(GYKJ6XUZUIURY 2+').6+-'9/9
A. Network Domain
6N_YOIGR2G_KX ,XKW[KTI_(GTJ]OJZN*GZG8GZK3UJ[RGZOUT)UJOTM In the network domain, we need to design the network
topology of the UAV network, including the type of
Fig. 2. Multi-UAVs Network Protocol Stack communication entity, the number, geographic location
information (longitude, latitude, and altitude), and attitude
1) Physical Layer: In this layer, it involves a number of information (yaw, pitch, and roll). As shown in Fig. 3, we
items: base frequency, bandwidth, data rate, channel coding, suggest the multi-UAVs topology with three levels.
modulation, duplex mode, radio channel state, beyond line of ¾ Level 1, Ground User Network: dozens of terrestrial units
sight (BLOS) capability, and antenna pattern and gain, etc. are grouped into 2 areas. Each of them is equipped with an
2) Datalink Layer: Due to high mobility and dynamic 802.11 interface. These terrestrial units generate data traffics
topology changes in multi-UAV networks, to design an between the ground station and mobile in a rectangle based
efficient Medium Access Control (MAC) protocol is quite on a FANETs mobility model.
challenging [7]. Typical MAC protocols of multi-UAV ¾ Level 2, Flying Ad-Hoc Network: 12 low-altitude
networks can be classified as shown in Fig. 2. (hundreds of meters) UAVs are deployed to cover the
3) Network Layer: A major challenge in multi-UAVs terrestrial units. The low-altitude UAVs are equipped with
network layer lies in the FANET routing protocol design, which two 802.11 interfaces, one for coverage communication with
is critical in acheiving the low end-to-end delay, robust the terrestrial units and another for communication with the
connectivity, fast convergence, and high delivery ratio [8]. high-altitude UAVs. These UAVs are equipped with
FANET is a subclass of Mobile Ad Hoc Network (MANET). quadcopters that can be hovered at a stationary height.
Therefore, typical routing protocols designed for MANET can ¾ Level 3, Aerial Mobile Backbone Network: 3 high-altitude
be applied in FANET with slightly modification. As shown in (thousands of meters) UAVs can be deployed to provide
Fig. 2, these routing protocols can be categorized into static backbone relay capability. These high-altitude UAVs are
routing, preactive routing, on-demand routing, hybrid routing, equipped with one802.11 interface for communication with
and geographic routing [9]. low-altitude UAVs, one TDMA interface for FANET
communication between the UAVs and one FDM interface
for relay communication with the far-away ground station.

Authorized licensed use limited to: East Carolina University. Downloaded on July 29,2020 at 12:59:54 UTC from IEEE Xplore. Restrictions apply.
 Fig. 3. Typical Network Topology in OPNET GUI
Fig. 5. High Altitude UAV Node Model

B. Node Domain
3) Low Altitude UAV Node Model
In multi-UAVs network simulation scenario, we defined 4 Low altitude UAV node model is designed as a two 802.11
kinds of node models, which are ground station, high altitude interfaces router. The IF-1 is used to connect with high altitude
UAV, low altitude UAV, and terrestrial unit. UAVs and the IF-2 is used to connect with terrestrial units in the
same area. With different BSS-ID configuration, the different
1) Ground Station Node Model
802.11 interfaces can work without interference.
Ground station node model is designed as a OSI protocol
stack server equipped with one FDM wireless interface. Based
on the high gain of the directional antenna, the ground server can
connect with the high altitude UAVs via long-distance wireless
FDM link. The ground station can handle traffic flows to and
from remote UAVs and terrestrial units.

Fig. 6. Low Altitude UAV Node Model

4) Terrestrial Unit Node Model

Terrestrial unit node model is designed as an 802.11 client
with full OSI protocol stack. It can generate traffic flows in the
Fig. 4. Ground Station Node Model
application layer module and run routing protocols in the
network layer module. With the 802.11 interfaces, terrestrial
2) High Altitude UAV Node Model units in the same area can communicate each other directly or
High altitude UAV node model is designed as a router with remotely via multi-hop forwarding. The low altitude UAVs that
e interfaces. The IF-1 is an FDM interface, which is equipped cover the area can provide relay ability for these terrestrial units.
with a PAA directional antenna. The IF-1 can provide high gain
and long transmitting distance between high altitude UAV and
ground station. The IF-2 is a TDMA interface, which can
provide collision-free wireless links between the high altitude
UAVs. The IF-3 is an 802.11 interface, which provide links
between high altitude UAVs and low altitude UAVs.

Authorized licensed use limited to: East Carolina University. Downloaded on July 29,2020 at 12:59:54 UTC from IEEE Xplore. Restrictions apply.
 the terrestrial units. Low altitude UAVs in different areas
should be configured in different BSS-ID to avoid wireless
interference.
¾ Level 3: The high-altitude UAV directly connected with the
ground station should be configured more time slots in order
to get enough link capacity.
3) Network Layer Model
The selection and configuration of the network layer routing
protocol need to consider the hierarchical characteristics of the
multi-UAVs network. Routing protocols need to maintain the
connectivity in the multi-level hierarchical topology, where the
UAVs at the lower level become the members of the next higher
level. In order to reduce the routing overhead and to efficiently
Fig. 7. Terrestrial Unit Node Model use the radio channel resources, different levels can use different
routing protocols.
C. Porcess Domain
¾ Level 1: the topology may change with the terrestrial units’
Process models are mainly used to implement the mobility. So the MANET routing protocol such as AODV or
communication protocols and algorithms, which specify the OLSR is suitable.
behavior of modules that exist in the node domain. As shown in
TABLE I. each node model has their protocol configuration. ¾ Level 2: since the low altitude UAV’s IF-1 is worked as a
member of level 1 network, it should be configured using
TABLE I. PROTOCOL PROCESS MODEL CONFIGURATION AODV or OLSR. Similarly, the IF-2 is configured using
High Low same routing protocol with level-3 network.
Node Ground Terrestrial
Layer Station
Altitude Altitude
Unit ¾ Level 3: the route of the high altitude UAVs are relatively
UAV UAV
Application Server NA NA Client fixed and the changes are slow. The static routing protocol
Transport TCP/UDP NA NA TCP/UDP such as LCAD or MLHR is preferred.
Network AODV [13] AODV OLSR [14] OLSR
IF1. FDM
4) Transport Layer Model
IF1. 802.11 The transport layer protocols are only configured on the
Datalink IF1. FDM IF2. TDMA 802.11
IF2. 802.11 nodes which generates or process data traffic. As shown in
IF3. 802.11
Physical Refer to TABLE II. TABLE III. , the loss and delay tolerance characteristics of the
1) Physical Layer Model application are the main factors for the selections of transport
In physical layer, we need to configure important wireless protocols.
model parameters, including modulation, data rate, base
frequency, bandwidth, and antenna pattern. The propagation 5) Application Layer Model
model is another important issue we should configure. It defines In different scenarios, application traffic of multi-UAVs
the wireless signal path loss as a function of frequency, distance network are asymmetric. We need to analyze the traffic volume
and other conditions. Typical propagation models include Free and fluctuation characteristics and configure the traffic at each
Space Model, Longley Rice Model, HATA Model, CCIR node’s application layer [16].
Model, Walfisch-Ikegami Model, and TIREM (Terrain
TABLE III. MULTI-UAV NETWORK APPPLICATION CHARACTERISTICS
Integrated Rough Earth Model) [15].
Property Delivery Sensing Monitoring Relaying
TABLE II. PHYSICAL LAYER PARAMETERS OF INTERFACES Traffic Depending on
Low Medium High
Volume the scene
Interface Traffic
FDM TDMA 802.11 Low High Low Medium
Parameter Fluctuation
Modulation 64QAM 16QAM QPSK Delay Yes,
Data Rate 10 Mbps 2 Mbps 11 Mbps No Yes, Secs No
Sensitive 100ms
Frequency/ 5.0GHz/ 250MHz/ 2.4GHz/ Data Loss No Loss No Loss Loss Tolerant No Loss
Bandwidth 10MHz 5MHz 20MHz Throughput Low Medium High High
Antenna Pattern Directional Isotopic Isotopic Transport
Propagation Model Free Space Longley-Rice TIREM TCP UDP RSTP TCP
Protocol
2) Datalink Layer Model Mobility High Low Low Stationary
As the node model design, we suggest 3 kinds of datalink
D. External System Domain
protocols.
In the literature, the standalone simulations may not meet the
¾ Level 1: Each group of terrestrial units should be configured actual needs of multi-UAVs network researchers. As shown in
in the same 802.11 BSS-ID. Fig. 8, we set up a distributed multi-UAVs network simulation
¾ Level 2: The low altitude UAVs who cover the same ground system, which consist of 4 subsystems. These subsystems are
group should configure their IF-1 802.11 BSS-ID as same as connected by a data distribution bus (HLA[17] or DDS[18]).

Authorized licensed use limited to: East Carolina University. Downloaded on July 29,2020 at 12:59:54 UTC from IEEE Xplore. Restrictions apply.
 Network Simulation 3D Visualization STIL Application STIL Application
Subsystem Subsystem Source Desination

Data Distribution Bus (HLA or DDS)

Mobility Information Real Traffic Flow (From Source)
Real Traffic Flow (To Destination)

Fig. 8. Distributed UAV Simulation System

IV. SIMULATION EXPERIMENT

Based on the research aims, we can build a series of multi- Fig. 10. Packet Delivery Ratio
UAVs network simulation cases on the simulation platform.
2) Case Study II: 3D Visualization
1) Case Study I: Network Performance Simulation In addition, a 3D network visualization subsystem is
The base simulation scenario topology is shown in Fig. 3, 50 connected to our UAV network simulation platform. In Fig. 11,
terrestrial units are uniformly randomly placed in 2 areas of 5000 we can see the subsystem can demonstrate the UAV entities and
h5000 m2. Within each area, each terrestrial unit randomly wireless communication links in a virtual 3D world. These
chooses a speed (uniformly distributed from 1 to 10 m/s), and features allow UAV network protocol developers to observe
moves to this point with the chosen speed. After that it will stay their design more directly.
there for 30 seconds and then choose a new speed to move. Each
terrestrial unit generates data traffic to the ground station in
constant bit rate mode. The simulation duration is 3600 seconds.
We configure 2 routing protocols, which are AODV and OLSR.
For each of test scenarios, we get 2 results: packet end-to-end
delay and delivery ratio.
(1) Packet End-to-End Delay: Fig. 9 shows the average
packet end-to-end delay of AODV and OLSR with the
progressively heavier routing traffic. With the growth of routing
traffic, OLSR has a much better performance than AODV has.
This is because of the proactive nature of OLSR, it can always
find shorter routing paths when the network topology changes
and routing traffic increases.

Fig. 11. 3D Network Visualization Subsystem

3) Case Study III: Real Application Testbed

Instead of simulations, may be more realistic to consider a
hybrid approach in which only the lower layers (physical,
datalink and network layers) are simulated and all the upper
layers (from transport to application) are executed on a
dedicated hosts. In our UAV network simulation platform, we
can emulate application traffic flow using the SITL (System-In-
The-Loop) technology [15]. As shown in Fig. 12, the video
stream traffic can be delivered from the ground station node to
Fig. 9. Packet End-to-End Delay any UAV or terrestrial unit node in simulation.
(2) Packet Delivery Ratio: As shown in Fig. 10, we can see
that OLSR also outcomes AODV in this aspect. With the growth
of routing traffic, MAC layer will undertake more loads and
SITL SITL
consequently more and more packets will be accumulated in
MAC layer packet queue. If the congested delay exceeds the
MAC delay limitation, these packets will be dropped. Compared Ground UAV Communication UAV (LA)
Station Network based on OPNET
with OLSR, AODV has much lower routing overhead, the
AODV can greatly reduce the possibility of dropping packets Fig. 12. Real Application Testbet Subsystem
due to network congestion.

Authorized licensed use limited to: East Carolina University. Downloaded on July 29,2020 at 12:59:54 UTC from IEEE Xplore. Restrictions apply.
 As shown in Fig. 13, with different UAVs network ACKNOWLEDGMENT
configurations, the video transmission effects can be observed This work was supported by the National Natural Science
under different packet loss ratios and further optimize the UAVs Foundation of China (No.61572220) and partly supported by my
network protocol configurations. visiting appointment at UMass Dartmouth.
REFERENCES
[1] Lav Gupta, Raj Jain, and Gabor Vaszkun, “Survey of Important Issues in
UAV Communication Networks,” IEEE COMMUNICATIONS
SURVEYS & TUTORIALS, VOL. 18, NO. 2, 2016
[2] Ozgur Koray Sahingoz, “Networking Models in Flying Ad-Hoc Networks
(FANETs): Concepts and Challenges,” Journal of Intelligent & Robotic
Systems, April 2014, Volume 74, Issue 1–2, pp 513–527
[3] Aarja Kaur, Dr. Jyoteesh Malhotra, “A Survey of Network Simulation
Tools,” CiiT International Journal of Wireless Communication, 2015, Vol
7, No 6
[4] Arvind T , “A Comparative Study of Various Network Simulation Tools,”
International Journal of Computer Science & Engineering Technology
(A) Packet loss 1% (IJCSET), August 2016, Vol. 7 No. 08
[5] Clifton M. Durham, Todd R. Andel, Kenneth M. Hopkinson, and Stuart
H. Kurkowski, “Evaluation of an OPNET Model for Unmanned Aerial
Vehicle (UAV) Networks,” Proceedings of the 2009 Spring Simulation
Multiconference, Article No. 66
[6] David Cavin, Yoav Sasson, and André Schiper, “On the Accuracy of
MANET Simulators,” Proceedings of the second ACM international
workshop on Principles of mobile computing, Pages 38 – 43
[7] Arafat, Muhammad Yeasir Moh, Sangman, “MAC Protocols for UAV
Networks : A Comparative Study,” KISM Spring Conference 2019, At
Chungju, South Korea
[8] Omar Sami Oubbati, Abderrahmane Lakas, Fen Zhou, Mesut Günes, and
Mohamed Bachir Yagoub, “A survey on position-based routing protocols
for Flying Ad hoc Networks (FANETs),” Vehicular Communications,
Volume 10, October 2017, Pages 29-56
(B) Packet loss 5% [9] Daniel Lihui Gu, Guangyu Pei, Henry Ly, Mario Gerla, and Xiaoyan
Hong, “Hierarchical Routing for Multi-Layer Ad-Hoc Wireless Networks
with UAVs,” Proceedings - IEEE Military Communications Conference
MILCOM, Volume 1, 2000, Pages 310-314
[10] Samira Hayat, Ev¸sen Yanmaz, and Raheeb Muzaffar, “Survey on
Unmanned Aerial Vehicle Networks for Civil Applications: A
Communications Viewpoint,” IEEE Commuincations Surveys &
Tutorials, Vol. 18, No. 4, 2016
[11] Mohammad Mozaffari, Walid Saad, Mehdi Bennis, Young-Han Nam,
and M´erouane Debbah, “A Tutorial on UAVs for Wireless Networks:
Applications, Challenges, and Open Problems,” IEEE Communications
Surveys & Tutorials, Vol. 21, Issue: 3, 2019
[12] Junfei Xie, Yan Wan, Jae H. Kim, Shengli Fu, and Kamesh Namuduri,
“A Survey and Analysis of Mobility Models for Airborne Networks,”
IEEE Communications Surveys & Tutorials, Vol. 16, No. 3, 2014
(C) Packet loss 10% [13] C. Perkins, “Ad hoc On-Demand Distance Vector (AODV) Routing,”
RFC 3561, July 2003
Fig. 13. Video Transmission Effect under Different Packet Loss Ratios
[14] T. Clausen, “Optimized Link State Routing Protocol (OLSR),” RFC 3626,
October 2003
V. CONCLUSION
[15] OPNET Technologies Inc., “OPNET Modeler Help Release 16.0,” 2010
In this paper, we developed an UAV communication [16] H. Shakhatreh1, A. Sawalmeh, A. Al-Fuqaha, Z. Dou, E. Almaita, I.
network simulation platform using OPNET Modeler. We expect Khalil, N. S. Othman, A. Khreishah, and M. Guizani, “Unmanned Aerial
to provide a powerful and comprehensive multi-functional tool Vehicles (UAVs): A Survey on Civil Applications and Key Research
that can support following functionalities: Challenges,” IEEE Access, Vol. 7, April 2019
[17] IEEE Standard 1516-2020, “IEEE Standard for Modeling and Simulation
z Performance Simulation: to study various UAV network (M&S) High Level Architecture (HLA) -- Framework and Rules,”
performances under different protocols and configurations. https://standards.ieee.org/standard/1516-2010.html, 2010-03-25
[18] DDS Foundation, “Data Distribution Service,” https://www.dds-
z 3D Visualization: to study how to visualize multi-UAVs foundation.org/, March 2015
network works in a virtual 3D world.
z Real Application Testbed: to study how real application
traffic performed in a highly controlled, reproducible
multi-UAVs network environment.

Authorized licensed use limited to: East Carolina University. Downloaded on July 29,2020 at 12:59:54 UTC from IEEE Xplore. Restrictions apply.