Lecture #24

Bluetooth & CAN

18-348 Embedded System Engineering
Milda Zizyte
Wednesday, 13-April-2016
(With slides contributed by Chris Szilagyi)

Electrical &Computer
ENGINEERING
© Copyright 2006-2015, Philip Koopman, All Rights Reserved

http://www.sailmagazine.com/
cup-watch/wings-next-
generation
Where Are We Now?
 Where we’ve been:
• RTOS
• SW Quality

 Where we’re going today:

• Bluetooth
• CAN

 Where we’re going next:

• Test #2 on Wed April 20, 2016
– Covers all lectures after Test #1

• Final projects

Preview
 Controller Area Network (CAN)
• Important automotive network protocol
• Bit dominance
• Binary countdown

 Bluetooth
• Wireless protocol
• Gaussian Frequency Shift Keying
• Frequency Hopping Spread Spectrum

4
Embedded Communications
 So far this semester mostly just one microcontroller

 What if we want to connect two (or more) computers?

• Earlier in the semester you used a serial bus
• Lab: Communication between module and PC
• Gets complicated as soon as we connect more than two nodes on a shared wire
– How do you know which node transmits next?

 Look at two common embedded network protocols

• Wired - Controller Area Network
• Wireless - Bluetooth

Wired Embedded Network:

Controller Area Network
(CAN)
CAN Is Central To Automotive Networks

[Leen02]
7

Controller Area Network (CAN) Protocol

 Originally developed as automotive protocol
• Used in essentially all cars made today
• Real time, high reliability communications among:
– Controllers
– Sensors
– Actuators
– Human interfaces
• Up to 1 Mbit/s data rate (for a 40 meter wiring harness)
– Often run at 250kbps or 500kbps to keep wiring costs down (faster is expensive)

 The way it is usually used in embedded control networks is:

• Each node performs some tasks in its control loop
• Each node send periodic state data, constantly updating everyone else
– E.g., “here is the engine speed in RPM” sent at 10 times per second
• Messages may also indicate infrequent events

8
CAN History and Applications
 Developed and published in 1980s by Robert Bosch GmbH
• Originally developed to deal with the complexities of automotive networks
• Today almost every automobile has at least one CAN network

 Also used in many other distributed embedded control systems

• Trains
• Ships
• Industrial automation
• Robotics
• Data center power backup systems (coordinates batteries & inverters)

 One of the most wide-spread embedded network protocols

• Anything used in automobiles is forced to be ultra-inexpensive
• So automotive components are attractive for other applications

CAN – Example Network Topology

 Typically set of nodes and topology is fixed at design time

 We know exactly what nodes are connected to the network
[Siemens] 10
CAN - Physical Layer (Remember this?)
NOISE-FREE OUTPUT SIGNAL
 Differential Drivers
1 1
UNCERTAIN VALUE
0
 Send both Data and GND
0

Inverse Data values on a EXAMPLE NOISE

2-wire bus

GND
 Receiver subtracts two
voltages
INVERSE BUS DATA
• Eliminates common mode
voltage bias
• Leaves any noise that
affects lines differently GND
BUS DATA

RECEIVED DIFFERENTIAL VALUE

UNCERTAIND VALUE

GND 11

Non-Return to Zero Encoding (Remember this?)

 Example: Send a Zero as LO; send One as HI
• Worst case can have all zero or all one in a message – no edges in data
• Simplest solution is to limit data length to perhaps 8 bits … BUT
– CAN uses “bit stuffing” instead of fixed byte size
– Extra reverse polarity “stuff” bit is inserted after 5 values the same in a row
• Bandwidth is one edge per bit
Simple NRZ Bit Encoding

SYMBOL SYMBOL SYMBOL SYMBOL

ONE ZERO SYNC END

H L L H

PHYSICAL PHYSICAL PHYSICAL PHYSICAL

BIT BIT BIT BIT

Simple NRZ Encoding Example: 1101 0001

END SYNC ONE ONE ZERO ONE ZERO ZERO ZERO ONE END SYNC …
PREVIOUS SUBSEQUENT
MESSAGE MESSAGE
L H H L H L L L H H
12
Non-Return to Zero Encoding (Remember this?)
 Example: Send a Zero as LO; send One as HI
• Worst case can have all zero or all one in a message – no edges in data
• Simplest solution is to limit data length to perhaps 8 bits … BUT
– CAN uses “bit stuffing” instead of fixed byte size
– Extra reverse polarity “stuff” bit is inserted after 5 values the same in a row
• Bandwidth is one edge per bit

Bit stuff example: 1111 1111

13

CAN - Media Access Strategy

 All of these nodes are trying to communicate on the same wire
• At some point we’re going to get contention on the bus

 What if multiple nodes try to send messages simultaneously?

• If two nodes attempt to broadcast at the same time, get a collision

Node #1 Message Message

LUCKY
Node #2 Message Message
UNLUCKY LUCKY
(Collision)

 Ethernet uses random backoffs then retransmits

• But, you can get unlucky and collide repeatedly
• There are many tricks (see 18-649), but then it is something other than Ethernet
• Need a collision-free protocol to make real time deadline guarantees!
 CAN uses a more elegant solution - Bit dominance
• Collision free protocol for real time control – need bound on worst case time
14
CAN – Bit Dominance
 CAN uses the idea of recessive and dominant bits
• Also called a “Wired OR” design
• Bus floats high via resistor unless a transmitter pulls it down (down=dominant)
• (Other bus wire in differential transmission floats low and transmitter pulls up)
 High is “recessive” value
• Sending a “1” can’t override the value seen on the bus
 Low is “dominant” value
• Sending a “0” forces the bus low no matter what another node is sending

Resistor

ON OFF OFF

[Siemens]
15

CAN – Binary Countdown

 Use bit dominance to decide who gets the bus
• Based on message header

 Binary Countdown operation:

• Each node uses unique identifier for each message type
• All nodes wishing to transmit compete for the channel by
transmitting a binary signal based on their identification value
• A node drops out the competition if it detects a dominant state
while transmitting a passive state
– If it transmits “1” and sees “0” on the bus, it knows it lost the arbitration
• Thus, the node with the lowest identification value wins
• Contention is resolved without consuming any extra bandwidth

16
Example: Binary Countdown
Value seen
on bus
0 0 0 1 0 0 0 Node 8 message

Values that each node attempts to transmit:

Node 8
ID=0001000 0 0 0 1 0 0 0 8 has the bus

Node 9
ID=0001001 0 0 0 1 0 0 1
Node 10
ID=0001010 0 0 0 1 0 1 9 drops out

Node 12
ID=0001100 0 0 0 1 1 10 drops out

Node 17
ID=0010001 0 0 1 12 drops out

Time
17 drops out (stops competing for the bus) 17

CAN - General Message Format

HEADER DATA ERROR DETECTION
19 or 39 bits 0 to 64 bits 15 bits
 Header
• Application sets any desired value in 11-bit or 29-bit identifier
– A few other fields in the header too (e.g., 4 bits for length of data field)
• Identifier determines global priority (which message gets on bus first?)
• Headers often chosen to reflect source, destination, and/or message type
– But are entirely application-dependent
– Broadcast messages – all receivers can receive every message by default
 Data
• Application- or high-level-standard defined data fields
• CAN Data field size is variable (0 to 8 bytes, defined via length in header)
 Error detection
• Detects corrupted data (uses a 15-bit CRC):
– All 15-bit or shorter burst errors (groups of flipped bits clumped together)
– All 5-bit errors regardless of where they occur (except for a design bug; see 18-649)
18
CAN in Real Time Systems
 Why not just use Ethernet and TCP/IP?
• Collision avoidance uses random backoff and retry strategy
• Worst case message delivery is infinite!
• OK for best effort, but not real time
• Message overhead is huge – what if you just want to send “switch X is ON”?
– 18(min. Ethernet "header" fields) + 6(padding) + 20(IPv4) + 20(TCP) = 64 bytes
(http://stackoverflow.com/questions/1846077/size-of-empty-udp-and-tcp-packet)

 CAN message format is optimized for real time control

• Bus arbitration uses low overhead and is constant
• Bit dominance enables global prioritization of messages
• Can work efficiently on inexpensive 250K, 500K, or 1Mbit/sec networks

 CAN allows real time schedulability

• Non-preemptive prioritized messages

 Can predict worst case latency for message delivery

• As long as the system isn’t overscheduled
• The math looks the same as the non-preemptive prioritized tasking equation!
19

CAN Tradeoffs
 Advantages
• Very low overhead (minimum size is only 44 bit message with 11 bit ID)
• High throughput under light loads
• Global prioritization possible, basically for free
• Arbitration is part of the message - low overhead
 Disadvantages
• Requires bit dominance
– Restricts physical layer choices
• Unfair access - node with a high priority can "hog" the network
– Can be reduced in severity by using Message type instead of Node # for header
– Can, in principle, use a bus guardian to limit duty cycle of each node
– You have to pay attention to real time (e.g., Deadline Monotonic Scheduling)
• Poor latency for low priority nodes
– Starvation is possible

 Typical performance:
100 bit messages @ 1 Mbit/sec  10,000 msgs/sec
20
 Wireless Embedded Network:
Bluetooth

Bluetooth
 Open wireless protocol standard
• Published by Bluetooth Special
Interest Group (SIG)
• http://www.bluetooth.org

[Wikipedia]

 Provides short range wireless communication

• 10 to 100 meter range
• 1 Mbit/s (basic mode) up to 3 Mbit/s (enhanced data rate mode)
• Primarily intended for voice and data transfer
• Eliminates wires and cables between both stationary and mobile devices
• Ad-hoc networks and synchronization of personal devices
• Intended to be cheap (< $5 per node) to allow ubiquitous computing
– This is still pretty expensive in embedded

22
Bluetooth - History and Applications
 Harald Blåtand (Bluetooth in English)
• Viking and king of Denmark between 940 and 981
• One of his skills was to make people talk to each other
• During his rule Denmark and Norway were united
… what they don’t tell you is that his murder was arranged by his son, who
coveted the throne

 Goal: Bluetooth standard lets devices communicate and work together

• PCs and peripherals [Wikipedia]

• Synchronization of PDAs and other

personal devices
• Cellphones and wireless headsets
• Tethering (relay Internet connection
from cell phone to laptop)

23

Bluetooth - Network Topology

 Ad-hoc network
• Any nodes that come within range can form small network called a “piconet”
– Nodes can dynamically join and leave the ad-hoc network
– Nodes can belong to multiple piconets
• 1 master node and up to 7 slave nodes in a piconet
• Elects a new master if master leaves network

 Scatternets
• If a node is part of two different piconets, can share information across them
24
Bluetooth - Physical Layer
 Bluetooth uses RF as a wireless medium
• Uses the unlicensed 2.4 GHz Industrial-Scientific-Medical (ISM) radio
frequency band bit 0 bit 1 bit 2 bit 3 bit 4
• Why is ISM at 2.4 GHz?


Bits are sent using frequency
modulation (FM)

Gaussian Frequency Shift Keying

• Different frequencies to
indicate a “1” or “0”
• “Gaussian” just means it uses a
{
+ Data

Carrier

smoothing filter on the data first

to remove sharp edges =
[Wikipedia] Modulated Signal 25

Bluetooth - Media Access Strategy

 Get collisions if two devices broadcast on same frequency at same time
• Need to prevent two nodes in piconet from causing collision
• Need to prevent collisions caused by outside RF interference

 Bluetooth uses master-slave approach within the piconet

• Master controls which slave gets to broadcast next
• Slaves wait to be given their turn

 Frequency Hopping Spread Spectrum

• Constantly hop among psuedo-random set of frequencies
• Reduce collisions caused by outside RF interference

26
Bluetooth - Master Slave
 Master broadcasts every other message
• Sends data and declares who goes next; master relays data to other nodes
 Time Division Duplex
• Time divided into slots
• All slave nodes synchronize to master node’s clock (start of message)
• Use a different frequency each time slot

“Node 1: Go” “Node 2: Go” “Here is some data for node 3; [Bluetooth Core Spec 3.0]
Node 3: Go”

“Node 1 Response” “Node 2 Response” “Node 3 Response” 27

Bluetooth - Frequency Hopping Spread Spectrum

 Master defines a pseudo-random series of frequencies
• Signal hops among frequencies to avoid interference
• Tolerate the occasional collision; data lost but protocol recovers automatically
 Other devices on the 2.4 GHz ISM band
• Garage door openers, baby monitors, microwave ovens
• (2.4 GHz is a bit of a mess with different protocols using it different ways)

f Piconet 1 Piconet 1
4

f Piconet 1 Collision
3
Baby Monitor Piconet 1
f
2
Frequencies

f
1
f Piconet 1
0
Time
28
Bluetooth - General Message Format

Access Code Header Payload

68 – 72 bits 54 bits 0 – 2745 bits

 Access Code
• Want to keep messages from two piconets separate if they happen to use same
frequency pattern
• Master node defines an access code for its piconet
• Any message must first have a valid access code
• Includes some serious error detection/error correction

 Header
• Contains destination or source address, packet type, flow control, and error
checking code

 Payload
• Application data, often voice or other media
• Includes 16-bit CRC to catch corrupted data 29

Bluetooth Tradeoffs
 Advantages
• No wires!
• Ad-hoc networking
– Nodes can join and leave whenever they want
– Overlapping/interacting/hierarchical networks (scatternets)

 Disadvantages
• No way to distribute power (everything has to have a battery or power plug)
• Geometry may introduce standing waves/fading
• Interference from other RF emitters (EMI = ElectroMagnetic Interference)
• Limited spectrum space
• In general, unsuitable for use in critical applications that aren’t fail-safe!

 Also, cost
• Bluetooth cheap, but still working on being super- “cheap”
• Has to be able to beat a piece of copper and a plastic connector
30
CAN/Bluetooth Comparison
 Both:
• Relatively efficient for relatively short messages
• Optimized for single-hop network to connect physically close devices
• Give predictable and bounded worst-case message delays

 CAN:
• Wired network for critical control applications
• Assumes fixed network configuration known at design time
• Prioritized message sequence supports rate monotonic scheduling
• Optimized for extremely short data payloads (e.g., 1-byte payloads)

 Bluetooth:
• Wireless network for consumer electronics data exchange
• Higher overhead, but longer maximum message size for bulk data transfer
• Round-robin message sequencing for fairness

31

Other Wireless Alternatives

 ZigBee – IEEE 802.15.4
• Lightweight, low-rate data protocol
• Desgined as cheaper alternative than Bluetooth

 Custom wireless protocols

• For example, lighting systems, remote automotive entry devices
• For battery-powered equipment RF can take most of the energy budget
• FCC regulates frequencies and how many bits you can send
– Often only a few msec of data at a low data rate
– Serious limitations on automatic devices (i.e., no “button press”)

 Infrared – IrDA
• Infrared Data Association
– http://en.wikipedia.org/wiki/Infrared_Data_Association
• TV remotes … but also data transfers up to 1 Gbit/Sec with GigaIR
• Good for indoors within same room; no RF interference from WiFi
– Wii uses Bluetooth for data communications, IR for pointing
32
Review
 Controller Area Network (CAN)
• Bit dominance and binary countdown

 Bluetooth
• Wireless protocol
• Frequency hopping to minimize effects of collisions

 Comparisons between CAN and Bluetooth

33