Reliable Time-Critical Wireless Communication in … Time-Critical Wireless Communication in...

Reliable Time-CriticalWireless Communication in Vehicular NetworksElisabeth Uhlemann

PhD student course in:

Communications for Cyber-Physical SystemsMälardalen University

2014 11 20

2

The Connected World of Cooperative ITS

Source: CVIS, 2008

What is Cooperative ITS?

● Intelligent Transport System (ITS) is a system whereyou blend transport with information and communication technologies (e.g., GPS, WiFi, 3G, etc.)

● E.g., intelligent bus stop signs tell you when the next bus arrives, variable speed signs

● The end user is only a consumer of information

● Cooperative ITS (C-ITS) is a subgroup of ITS, wherethe end user also provides information and does not only consume information

● E.g., vehicles transmitting position messages wirelessly toneighboring vehicles and the ego vehicle can use receivedinformation to predict an upcoming hazardous event

3

Benefits with C-ITS

Tranarpsbron, E6, Östra Ljungby, Sweden

January 15, 2013 March 31, 2013

Interstate 17, Virginia, US

Around 100 vehicles involved in both accidents.

4Slide courtesy of: Katrin Sjöberg, Volvo

5

Intelligent Transport Systems

Intelligent transport systems (ITS) complemented by wireless vehicular communications currently receive a great deal of attention worldwide:● Safety: reduce traffic accidents and human

injuries by introducing traffic safety applications based on vehicular communications.

● Efficiency: reduce congestion, travel-time and pollution through traffic efficiency applications.

● Value-added services: other types of services may be implemented and offered in order to facilitate system introduction and provide sustainable business and operation models.

Traffic Safety Applications

Extend the driver’s safety margin for detection and prevention of potential accidents:

● Speed reduction and braking● Wrong way driving warning● Intersection support● Reduce crash severity● Hidden object support● Stop light assist● Overtaking vehicle warning● Traffic condition warning● Roadwork warning● Collision risk warning● Blind spot warning ● Emergency vehicle warning● Motorcycle approaching indication● Lane change assistant ● Cooperative glare reduction

Source: Car2Car-CC, 2008

6

RearDetection

Blind Spot

SideCrash

LaneSupport

CollisionMitigation

Extended RearDetection &Blind Spot

CooperativeWarning

Infrastructure BasedWarning

Road Side Equipment (local or remote)

Extended Safe Distance and Speed

Extend the driver’s “Safety

Margin” to detect and prevent potential accidents

Dynamic inter-vehicle and vehicle-to-roadside network extends range of on-board vehicle systems

Lane ChangeAssistant

Source: Safespot Project, 20087

Local Dynamic Map

● In-Vehicle Platform Output Interface

map fromprovider

landmarks for referencing

treeidpos…

com nodes,fusion result

temporaryregional info !

accidentidpos…

fogidposa,b…

congestionidposlengthdir…

egoposvel…

rsuidpostype…

vehidposveltype…

Source: Safespot Project, 20088

9

Wireless Access Technologies

Variable Message Sign

Terrestrial BroadcastRDS, DAB

UMTSWiMAX

Info-Broadcaster

GSM-GPRS

Sat-Comm

COOPERS CVIS SafeSpot SISTER

BroadcastTransmitter

Vehicle-to-Vehicle (M5, IR, MM)

Hot-Spot(Wireless LAN)

GPS, Galileo

Beacon

•CALM-M5

•CALM-IR

•CEN-DSRC

Source: the Coopers, CVIS, Safespot and SISTER Projects, 2008

10

Standardisation – CALM

CommunicationsAccess for

LandMobiles

● Provide user transparent continuous communications

● A set of more than 20 standards

● Formally under ISO TC204 (ITS) WG16 (Communications)

● Base standards available, rest complete 2010

● First implementation within the CVIS project

Source: ISO, 2009

11

ProprietaryProtocol

B

Proprietaryor

DedicatedRadio

Interface

ProprietaryProtocol

C

Proprietaryor

DedicatedRadio

Interface

ProprietaryProtocol

A

Proprietaryor

DedicatedRadio

Interface

App. 1 App. 2 App. 3

ALL ITS

APPLICATIONS

TCP UDP

Stream &

Realtime

Protocols

ISO

DSRC

L7

HTTP/

SMTP

Protocols

MAC

802.11p

WAVE

Init

Hnd-

ovr

Secur

MAC

2G/3G

GSM

Init

Hnd-

ovr

Secur

MAC

IR

Init

Hnd-

ovr

Secur

L2/UDP

…

IPv6 layer

Service

QoS

Interface

Selection

Handover

Interface

QoS

Source: CVIS Project, 2008

Ad Hoc or Cellular

Communications?

V2VR2V

V2V

V2R

Peer-to-peer communication Communicating via an access point

12

Vehicular Ad Hoc Networks

● A Vehicular Ad hoc NETwork (VANET) has some benefits for traffic safety applications as compared to e.g., UMTS or LTE: ● it eliminates the problem of guaranteeing coverage ● it reduces the average communication delay ● no billing between VANET nodes (participation

without account)

13

Ad hoc no central entity.

Standards are a necessity!

● Standards outline how the communication should be performed, i.e., the same language must be spoken by the ones communicating

● Example of standards is WiFi, 3G/LTE, IP

● Protocols are agreements on how the wireless communication should be performed at different abstraction levels

● Standardization on Cooperative ITS has in Europe mainly been done within ETSI Technical Committee on ITS (ETSI TC ITS)

ETSI = European Telecommunications Standards Institute14

Local area network – IEEE LAN

IEEE = Institute of Electrical and Electronics Engineering

IEEE LAN family 802

15

IEEE 802.11 WLAN

WLAN = Wireless Local Area Network802.11

802.11a 802.11b 802.11g

802.11a/b/g PHY amendments

802.11e QoS

16

DSRC in US – IEEE 802.11p

● 75 MHz at 5.850-5.925 GHz Intelligent Transportation Systems Radio Service (ITS-RS)

● IEEE 802.11 CSMA● IEEE 802.11e Quality of Service ● IEEE 802.11a Orthogonal Frequency Divison Multiplexing

17

IEEE 802.11p

● Ratified July 2010

● PHY and MAC amendment

● No support for access points

● Peer-to-peer mode (ad hoc)

● IEEE 802.11a OFDM physical layer

● 3, 4.5, 6, 9, 12, 18, 24 and 27 Mbps

● 5.850-5.925 GHz Intelligent Transportation Systems Radio

Service (ITS-RS)

● 10 MHz channels

● 1 control channel and 6 service channels (WAVE 1609.4)

● European standard (ETSI) – ITS G5

● Worldwide standard (ISO) – CALM M5

18

WAVE

WAVE = IEEE 802.11p, 1609.0, 1609.1, 1609.2, 1609.3, 1609.4 and 1609.5

WAVE = Wireless Access in Vehicular Environment

802.11pPhysical

802.11p 1609.4Medium access

LLC 802.2Logical link

IPv6 Network

WSMP

TCP/ UDP

Transport

1609.31609.2Security

WSMP – WAVE Short Message protocol

HTTP etc

Application 1609.1

1609.5 WAVECommunications Managerstarted during spring 2008.

1609.0 WAVEArchitecture

19

Frequency bands for ITS

● North America● 10 MHz channels at a carrier frequency of 5.850-5.925

GHz (in total a 75 MHz wide band)● Intelligent Transportation Systems Radio Service (ITS-

RS)● WLAN look-a-like

● Europe● 5.795-5.805 GHz Road Transport and Traffic

Telematics (RTTT)● RFID look-a-like

● 5.875-5.905 30 MHz ● Japan

● 715-725 MHz available in 2012 ● WLAN look-a-like

● 5.770-5.850 GHz● RFID look-a-like with TCP/IP support

20

Source: National Telecommunications and Information Administration, http://www.ntia.doc.gov/osmhome/allochrt.html

21

● After less than 5 days and 31 rounds of bids, it was decided which operators got licenses in the 800 MHz band: six blocks of frequencies was auctioned out, but only 3 operators won:

● Hi3G Access, 2 x 10 MHz : Price 431 000 000 SEK

● Net4Mobility, 2 x 10 MHz : Price 469 000 000 SEK● have promised coverage improvements for another 300 000 000 SEK

● behind Net4Mobility we find Tele2 and Telenor

● TeliaSonera, 2 x 10 MHz : Price 854 000 000 SEK

● The reason why the different frequency blocks had different prices is that some blocks require adjustments of the mobile technology. Telia Sonera’s block does not require this.

22

Frequency Auction in Sweden

for 4G at 800 MHz band in 2011

Source: NyTeknik

23

-70-60-50-40-30-20-10

0102030

5450 5500 5550 5600 5650 5700 5750 5800 5850 5900 5950

dB

m/M

Hz

ITS Spectrum in Europe

CE

N D

SR

C

Common Control Channel, Broadcast data, 10MHz@5900 (ch180) (Used by all)

Vehicle-Roadside data, 10MHz@5890 (ch178) (Safespot/C2C-CC, not used initially)

Service Channel: GeoRoute multihopping, 10MHz@5880 (ch176) (Safespot/C2C-CC)

Aux Channel for roadside initiated data, 20MHz @5480+ (ch96) (CVIS/COOPERS)

Blue line represent European spectrum mask for BRAN (conditional use for ITS)Red line represents European spectrum mask for ITS 5.9

V2

V (IS

M b

an

d)

V2

R –

re

str

icte

d e

mis

sio

n

SC

H: V

2R

-r

es

tric

ted

S

CH

: Ge

oR

ou

te M

ultih

op

CC

H

Au

x c

ha

nn

el –

R2

V

ITSBroadband Radio Access (BRAN)

Source: COMeSafety, 2008

Characteristics of VANETs

●Dezentralized network topology● No central controller like an access point or a base station

● Peer-to-peer communication

●Can contain roadside units (RSU)

●Share a common communication channel

●Broadcast● Traditional Automatic Repeat reQuest (ARQ) cannot be used

●The number of participating nodes in a VANET cannot be restricted

24

Reliable, Real-Time Communication

● The wireless communications protocols available today enables either reliable communications with low error rate or time-critical communications with real-time constraints – but not integrated high levels of both:● Voice has real-time requirements

but is relatively error tolerant. ● E-mailing requires reliable

communications but is delay tolerant. ● Control traffic on a fieldbus has

requirements on both reliability and real-time but the fieldbus is wired.

25

Examples of Communication

Requirements

bit error probability: < 0.1 %delay: < 100 ms

data rate: 10-15 kbit/s

bit error probability: < 10-9

delay: < 1 daydata rate: irrelevant

bit error probability: < 10-5

delay: < secondsdata rate: < 80 Mbit/s

Cell phones:

E-mailing:

Mobile broadband:

26

Retransmissions

42

What?

42

Ok

• When humans cannot hear we ask the speaker to repeat

• Computers can do the same.

• The reliability increases (the packet error rate reduces)

• The delay increases

• The deadline limits the number of attempts

27

Identified Problems

● Mobile nodes that moves quickly

● Retransmissions may not be possible

● Broadcasted messages with several recipients

● If 75% got the message should we retransmit or not?

● The lower delay, the longer the time for the driver to react

● We cannot suddenly loose coverage

● Ad hoc network

28

29

Scalability

V2VR2V

V2V

V2R

Medium Access Control

● It two vehicles transmit their position messages at the same time it is difficult for anyone to hear anything…

I’m here

I’m thereOh no!

30

Predictable MAC for Voice

● GSM (2G) uses TDMA ● Predictable

● Fair

● Not scalable

● Centralized (hardware can be added to improve scalability)

● UMTS (3G) uses CDMA● Predictable

● Can be unfair

● Scalable (graceful degradation)

● Centralized (hardware can be added to improve scalability)

31

● Currently the only standard supporting VANET is IEEE 802.11p using CSMA

● Listen-before-talk protocol:When the node wants to send a packet it starts by listening to the channel to determine if it is free. If free, the node sends directly, otherwise the node has to perform a backoff.

● Suitable for lightly loaded networks with bursty (event-driven) data traffic

● Pros:● Supports arbitrary packet sizes

● Low complexity, e.g., no need for synchronization

● Cons: ● Unpredictable

● Unfair (scales badly)

802.11pPHY

802.11p 1609.4MAC

LLC 802.2LLC

IPv6 Net.

WSMPTCP/ UDP

Trans.

HTTP etc

App.1609.1

Carrier Sense Multiple Access withCollision Avoidance (CSMA/CA)

32

Challenges for the MAC Layer

● Predictable access guaranteeing timely channel access delay in decentralized VANET

● The number of nodes participating in a VANET cannot be restricted – fairness and good scalability is needed.

● Interference depends on data collisions which in turn depends on the distance to the nearest concurrent transmitter.

● If a vehicle is not successful in transmitting its periodic update during a time interval, it becomes “invisible” to the surrounding vehicles.

33

Identified Problems

● Unpredictable channel access delay● Periodic messages need to be sent within its time

period

● The random backoff may cause a delay longer than the time period

● Causes packet drops at sending node

● Collisions ● The random backoff time chosen are discrete and

thus nodes may choose the same● For example in 802.11e highest priority have only 4

values {0 µs, 13 µs, 26 µs, 39 µs}

● Two concurrently transmitting nodes may be located very close together

34

35

CSMA/CA in 802.11p

Sensing range=1000 mPacket size = 500 byteReport rate = 10 HzApprox. 230 nodes

Packet drops/Deadline miss ratio

84%

23%

52% Channel access delay

Slide courtesy of: Katrin Sjöberg, Volvo

Overloaded network!

● Cooperative Awareness Messages:

● Hard deadlines but Medium reliability

● Hazard Warning Messages:

● Soft deadlines but High reliability

● A medium access method that is:

● Decentralized (works in ad hoc)

● Self-organizing (works even when nodes move)

● Fair (lets everybody talk at even intervals)

● Predictable (does not cause unexpected delays)

● allows every other cooperative awareness message through and enables repetition of hazard warnings

Desired Solution

36

Picture courtesy of Volvo

Self-Organizing Time Division

Multiple Access (STDMA)

● Already in commercial use: ● Automatic Identification System (AIS)

● VDL mode 4

● Specially designed for periodic messages, LDM is similar to AIS● In overloaded situation “collisions” are scheduled to minimize

interference

● Pros:● Predictable: always channel access regardless of the number of nodes in

the network – no packet drops!● Scalable: The network can be overloaded up to 700% without collapsing –

fair access is provided

● Cons: ● Packet sizes are fixed ● Higher complexity, e.g, needs synchronization through a GNSS

37

Self-Organizing TDMA (Håkan Lans)

NI=nominal increment, SI=selection interval, NSS/NS=nominal slot, NTS=transmission slot

SI SI

NI

….. …..

SI

NI

….. …..

SI

NI

SI

NI

NSNSS

NS NSS NS

Node 1

Node n

.

.

.

.

.

NTSNTS

NTSNTS

NTS

…..NI

…..

…..NI

…..

Choose a free slot within SI. If all slots in SI are occupied we pinch a slot from the node situated furthest away from us.


Channel access delay

300 byte packets, 6 Mbps, 10 Hz,

Channel access delay is not a problem in STDMA!


Interference Distance

Probability that two nodes initiatea transmission at the same time.

2 Hz, 800 byte, 6 Mbps


TxTx

Distance betweenconcurrent transmissions

● In CSMA concurrent transmission takes place randomly in space

● Higher interference level locally for the nodes situated closest to the transmitters

● Receivers cannot decode the packets correctly due to the high interferencein the system

Tx Tx

• In STDMA the position information sent is used for schedulingconcurrent transmission to be as far apart as possible in space

• The nodes closest to the transmitters are ”protected”, and can hopefullydecode the packet correctly

STDMACSMA


42

Research Challenges in

Vehicular Ad Hoc Networks

● Todays' wireless access technologies are not optimized for VANETs. To function properly, the communication needs to be predictable, fair, scalable and reliable.

● Challenges:● Ad hoc network with inherent interference and scalability

problems● Concurrent requirements on real-time and reliability● Traditional performance metrics not directly applicable

● Since communication can never be made 100% robust –the application has to be! However, reliability can be traded off for autonomy: ”Inform – Warn – Advice – Guide – Steer”.

43

Typical Communication Requirements

● Traffic safety applications relying on cooperative driving systems have exceptionally challenging requirements:

● Low-delay, ● reliable, ● time-critical,● scalable, ● wireless communication

is to be provided in a harsh, dynamic environment.

Source: CVIS http://www.cvisproject.org/

Message Types for VANETs


● Event-driven warning messages are broadcasted only in the event of a hazard.

● Requires very high reliability – each warning must be delivered.

● A lower delay gives the driver more time to react.


● Time-triggered position messages are broadcasted periodically.

● Since the messages are repeated periodically and do not signal imminent hazard the requirements reliability are moderate.

● However, to avoid invisible, mute vehicles, not too many consecutive messages can be lost.

● The maximum delay should be related to the message period.

44

Real-Time Communications

● A real-time system interacts with a physical system.

● This means that it is not only the result itself that is important but also when in time it is presented.

● An application with real-time requirements therefore has a deadline to meet.

ttdl

Usefulness of

presented value

ttdl

Usefulness of

presented value

Hard deadline Soft deadline

45





ttdl

Usefulness of

presented value

ttdl

Usefulness of

presented value

46



Hard deadlines Medium reliability



ttdl

Usefulness of

presented value

ttdl

Usefulness of

presented value

47



Hard deadlines Medium reliability



Soft deadlines High reliability

ttdl

Usefulness of

presented value

ttdl

Usefulness of

presented value

48

49

Qualitative Requirements

● So, how low-delay, reliable, and scalable, does the deadline dependent communication need to be?

● Well, that depends…

Source: CVIS http://www.cvisproject.org/

50

Problem #1: What is to be communicated?

●How long are the messages?

●How often is something sent?

●What QoS requirements do the messages have?

●What happens if a message is lost?

●How many consecutive messages can be lost?

Source: SAFESPOT, http://www.safespot-eu.org/

51

Different Realizations

● Hazard Warnings: Information is communicated only in the event of a hazard. Event-driven messaging with requirements on very low delay and very high reliability (the hazard is imminent and the data is critical). However, the system is useful also at moderate penetration rates since it can be implemented as an enhancement of existing sensors.

● Cooperative Awareness: Requires that all vehicles are equipped with a communication system and periodic time-triggered messages are broadcasted. Since the messages are repeated periodically and do not signal imminent hazard the requirements on delay and reliability are moderate. However, in order to avoid a system with invisible, mute vehicles, a very high penetration is likely needed

● Cooperative Autonomous Driving: Not only is this system required to warn or predict and advice – but also to act. Typically both periodic time-triggered and event-driven messages are needed, requiring some sort of service differentiation between the two types of messages, and very likely close to 100% penetration will be necessary.

52

Observation or Evaluation?

● The control loop that is using the communicated data consists of four steps: observation, evaluation, decisionand action.

● The communicated messages could either be raw data (observation) or some level of refined data (evaluation) or even the actual driver intentions.

● Transmitting processed data is likely to require less bandwidth but higher penetration rate such that decisions are made based on the same information.

● When penetration increases, more and more processed data could be transmitted to compensate for the increase in required collective bandwidth.

53

Problem #2: How is the information

used?

● What can the local dynamic map be used for? Warnings? Predictions? Advice?

● How trustworthy is the information?

● How do we make sure that the information indeed helps rather than confuses the driver?

AIDE truck platform based on Volvo FH12Source: AIDE, http://www.aide-eu.org/ Source: CVIS http://www.cvisproject.org/

54

Reliability versus Autonomy

● An issue that greatly influences the requirements of the wireless communication is the selected level of autonomy.

● The more reliable the communication system is, the higher autonomy can be given to the application:

● ”Inform – Warn – Advice – Guide –Steer”

55

Problem #3: Do Cooperative Driving

Systems Offer Improvements?

● How do we know that safety is improved?

● How do we verify that these distributed systems function as intended?

● If something goes wrong – who's fault is it?

● Do people want this?

56

Validation and Verification

● Information theory has provided a scientific foundation for the development of both the internet and cellular phones. But: ● performance limits of decentralized wireless networks are

presently not known, ● a central concept for information theory is capacity,

characterized by the maximum reliable throughput between two mobile terminals

● In real-time communication systems, traditional performance measure throughput is of less importance and the performance measure “deadline miss ratio” is instead used.

● However, even this performance measure needs to be redefined in due to the broadcast nature for both event-driven and time triggered realizations.

● Successful message reception should instead be defined as a function of the number of vehicles in communication range and interest range.

Platooning

• First step towards autonomous driving

• Platoon of vehicles (e.g. trucks) on a highway

• Vehicles follow a leading vehicle with minimum distance

++ reduced fuel consumption

+ increased safety and efficiency

– tighter control needed

• Exchange of status data and hazard warnings between

platoon vehicles

• Relatively stable topology

Fast and reliable inter-vehicle communication

57

Platooning

58

https://www.youtube.com/watch?v=tasa3D1vVTc

https://www.youtube.com/watch?v=tasa3D1vVTc

Challenges of Cooperative Driving

• Ideally

– Always up-to-date status information about every

surrounding vehicle

– Event warning dissemination with minimum delay

• Challenge

– Packet loss over wireless medium

– Packet drops due to limited, shared bandwidth

No channel access

Lost or erroneous packet

DENM

59Slide courtesy of: Annette Böhm, Halmstad University

Performance Measure: CAM

• CAM

• Goal: Every vehicle is informed about the up-to-

date status of its surrounding platoon members at

any given time

• Performance measure: CAM Age

•What is the worst-case inter-arrival time of

CAMs from a specific neighbor?

•Measured from the point in time when a vehicle

receives a CAM from a specific neighbor until

the point in time when the next CAM from that

neighbor is successfully received.

60

Performance Measure: DENM

• DENM

• Goal: Every vehicle in the platoon is warned of

unexpected situations with minimal delay

• Performance measure: DENM Dissemination Delay

• How fast can a hazard warning be spread throughout

the entire platoon?

•Measured from the time when one vehicle in the

platoon detects an event (e.g. the platoon leader

detecting the need to break immediately) until the

point in time when the entire platoon received a

warning about the event

61

MAC Comparison

a) Decentralized– Standard-compliant using 802.11p MAC and send rates

– CAM and DENM coexist

– Dedicated ITS Control Channel

b) Centralized– MAC method, message types and send rates not acc. to

standard

– CAM and DENM separated

– Dedicated Service Channel

CCH

SCH


Proposed Centralized MAC

What is the suitable SF size and CBP/CFP ratio to support

guaranteed channel access for CAMs and reasonable

dissemination delays for DENMs?

• 2 steps

1.Determine CBP size based on knowledge of typical

DENM Dissemination Delay for the platoon length at

hand (e.g. from simulation)

2.Determine CFP size by applying real-time

schedulability analysis


Superframe

Collision-Free PhaseCFP

Contention-Based PhaseCBP

CFP/CBP ratio adapted to current CAM traffic

DENM Dissemination Strategy

• How to stop dissemination of DENMs in a platoon?

• No acknowledgements of successful packet reception, i.e.

no way for a sending node to know who has and has not

received the packet

• Dissemination strategies

• Repeat 1 time

• Repeat 5 times

• Repeat until DENM is received from vehicle situated

further back in the platoon

64

DENM disseminationCAM Exchange

Platoonleader

Control vehicle for centralized channel access control

Slide courtesy of: Annette Böhm, Halmstad University

Delay in Communications

● Channel access delay

● Time from message creation until access is granted to access the channel

● Must be upper bounded

● Depends on the MAC method

● Overall delay

● The time from message creation until the message is successfully received by the destination(s)

● Can never, ever, be upper bounded

● Depends on the channel access delay and packet error rate

● Can be improved by diversity (redundancy)

65

The Wireless Channel

● The resulting error rates are generally higher

● The noise level varies with time, space and frequency

● The pathloss is considerable (and frequency dependent)

● Shadowing and multipath fading affect the error rates

● The medium has inherent broadcast nature

● The link capacity is limited by interference from other users (intentional or non-intentional)

Transmitter

ReceiverLOS

66

Propagation Mechanisms

Slide 67

● Reflection on smooth surfaces

● Transmission through objects

● Scattering on rough surfaces● Diffraction around sharp

edges

● Line-of-sight paths are not necessarily needed for communication

● Increasing frequency gives● more “optical” propagation● smaller antennas● higher path loss

67

If there is a LOS path, the fading h(t) is Rician distributed.

If there is no LOS the fading h(t) is Raylegih distributed

Unknowns:

● The number of paths, N

● The pathloss, alpha, for each path

● The delay, tau, for each path

● If the receiver is mobile each path also has a Doppler shift, phi

Statistical Multipath Model

2 ( )

0

( ) Re ( )c i Di

Nj f t

i i

i

r t s t e

Transmitter

Receiver

68

Multipath Fading

● The path loss depends on the distance between transmitter and receiver.

● The signal is scattered and different parts of the signal travel different distances to reach the receiver.

● At the receiver, the signals interfere with each other, constructively or destructively.

● The result is a fading channel.

0 200 400 600 800 1000 1200 1400 1600 1800 2000-35

-30

-25

-20

-15

-10

-5

0

5

10

time

att

enuation

Transmitter

Receiver

( ) ( ) ( ) ( )r t h t s t n t

69

● If both the receiver and the transmitter moves, everything becomes dependent of time:

● The paths, but also the number of paths changes with time, N(t)

● The pathloss component, alpha, varies with time for each path

● The delay, tau, varies over time for each path

● The Doppler shift, phi, varies with time

( )

2 ( ( )) ( )

0

( ) Re ( ) ( ( ))c i Di

N tj f t t t

i i

i

r t t s t t e

Multipath Fading for Mobile Communications

70

020

4060

80100

0

5

10

15

20-40

-30

-20

-10

0

10

timefrequency

att

enuation

spacefrequency

@time = 1a

tte

nu

ati

on

[d

B]

Channel Characterization

Source: Albin Dunand, Jean-Marc Conrat, France Telecom Research & Development,

COST2100 TD(08)406, Carrier freq. 2.2 GHz, bandwidth 62.5 MHz, downtown Mulhouse, France

0 360°

Diffraction over roof top

Transmission/scattering

through trees

ReflectionStreet waveguide

(guided reflections) Scattering

72

Channel Characterization

Fading

● Constructive and destructive interference of multiple propagation paths gives radically different received signal strength as we move around

● With many propagations paths and no strong a line-of-sight component, the received power changes dramatically by moving about half a wavelength

● The wavelength is83 10

5 cm for 6 GHzc

c c

cf

f f

• At 15 m/s (54 km/h), a wavelength at 6 GHz corresponds to 3ms

• It can be challenging to maintain reliable links in fading

• Remedies: error correcting codes, multiple antennas, wideband transmission…

73

Time-Frequency Response

Slide 74

74

Time and Frequency Selective Fading

020

4060

80100

0

5

10

15

20-40

-30

-20

-10

0

10

timefrequencyatt

enuation

• The fading will vary with time and frequency.

• Time diversity is one way to conquer fading. ARQ is an example of this.

• OFDM and DMT is a way to conquer frequency selective fading.

• Both examples may use the same error correcting code, but the redundancy is spread over time and/or frequencies.

75

Diversity for Increased Rate, Range or Reliability

● If retransmissions are used (time diversity) the range or alternatively the reliability is increased, but so is the jitter.

● Frequency hopping (frequency and time diversity) increases reliability in environments where frequency selective fading occurs, but also increases complexity (cost)

● Multiple-input-multiple-output (MIMO) systems (spatial diversity) gives a higher reliability, higher rate/throughput or longer range – but also increases complexity and has some form factor requirements (i.e., the antennas need to be located at least half a wavelength apart: 6.25 cm for 2.4 GHz and 3 cm for 5 GHz).

● Laptops can have antennas 6.25 cm apart – but most phones cannot.

76

● Would you hear better if you had four ears?● What if both your ears were located on the same side of

the head?

● MIMO gives: ● Increased throughput● Greater range

● Cell size versus transfer rate

77

MIMO – Multiple Multiple Output

1 feet = 0.3 m225 feet = 70 m

Retransmissions for Increased Reliability

42

What?

42

Ok

78

● When humans cannot hear, we ask the speaker to repeat

● Computers can do the same.

● The reliability increases (the packet error rate reduces)

● The delay increases

● A deadline limits the number of attempts

Retransmissions

42What?

He said 42 Ok

• Retransmit at another time

• Retransmit on another frequency

• Retransmit at another place, relaying

020

4060

80100

0

5

10

15

20-40

-30

-20

-10

0

10

timefrequency

att

enuation

79

80

Co-Existing Event-Driven and Time-Triggered

. . . . . . . . . . . . . . .

Last slot of previous hyperframe

First slot in current hyperframe

Beacon, vehicle 1 Beacon, vehicle 2 Beacon, vehicle N

Last slot in

current hyperframe

First slot of

new hyperframe

available slots for event -driven messages

Slide courtesy of: Le Nam Hoang, Halmstad University

Relaying in Platooning

81

One time-slot: only one option 1 -> 12, Pe = 0.2337, after 0.546 ms

Two time-slots: 1 -> 4 -> 12 or 1 -> 9 -> 12; Pe = 0.0265, after 1.092 ms

1 -> 1 -> 12; Pe = 0.0546

Three time-slots: 1 -> 4 -> 9 -> 12; Pe = 2.1 10-4, after 1.638 ms1 -> 1 -> 1 -> 12; Pe = 0.0128

Four time-slots: 1 -> 4 -> 6 -> 9 -> 12; Pe = 2.18 10-7, after 2.184ms

1 -> 1 -> 1 -> 1 -> 12; Pe = 0.003


Minimizing Dissemination Delay in Platooning

82


83

Research Challenges

● To function properly, the communication needs to be predictable, fair, scalable and reliable.

● Challenges:● Decentralized ad hoc networks ● Inherent interference and scalability problems● Concurrent requirements on real-time and reliability● Traditional performance metrics not directly applicable

● Delays in Communications

● Channel access delay● Depends on the MAC method

● To schedule this needs to be upper bounded

● Overall delay● Can never, ever, be upper bounded

● Depends on the channel access delay and packet error rate

● Only probabilistic guarantees can be given

Key References

● Journals:● Larson, J.; Liang, K.-Y.; Johansson, K.H., "A Distributed Framework for Coordinated Heavy-Duty

Vehicle Platooning," IEEE Transactions on Intelligent Transportation Systems, no.99, pp.1-11, 2014.● Bansal, G.; Kenney, J.B., "Controlling Congestion in Safety-Message Transmissions: A Philosophy for

Vehicular DSRC Systems," IEEE Vehicular Technology Magazine, vol.8, no.4, pp.20-26, Dec. 2013.● A. Böhm, M. Jonsson and E. Uhlemann, “'Performance comparison of a platooning application using the

IEEE 802.11p MAC on the control channel and a centralized MAC on a service channel,” in Proc. IEEE International Conference on Wireless and Mobile Computing, Networking and Communications, Lyon, France, October 2013, pp. 545-552.

● Kenney, J.B., "Dedicated Short-Range Communications (DSRC) Standards in the United States," Proceedings of the IEEE , vol.99, no.7, pp.1162-1182, July 2011

● Mecklenbrauker, C.F.; Molisch, A.F.; Karedal, J.; Tufvesson, F.; Paier, A.; Bernado, L.; Zemen, T.; Klemp, O.; Czink, N., "Vehicular Channel Characterization and Its Implications for Wireless System Design and Performance," Proceedings of the IEEE , vol.99, no.7, pp.1189-1212, July 2011

● Alexander, P.; Haley, D.; Grant, A., "Cooperative Intelligent Transport Systems: 5.9-GHz Field Trials," Proceedings of the IEEE , vol.99, no.7, pp.1213-1235, July 2011

● Sepulcre, M.; Mittag, J.; Santi, P.; Hartenstein, H.; Gozalvez, J., "Congestion and Awareness Control in Cooperative Vehicular Systems," Proceedings of the IEEE , vol.99, no.7, pp.1260-1279, July 2011

● Kaul, S.; Gruteser, M.; Rai, V.; Kenney, J., "Minimizing age of information in vehicular networks," 2011 8th Annual IEEE Communications Society Conference on Sensor, Mesh and Ad Hoc Communications and Networks (SECON), 27-30 June 2011, pp.350-358.

● Hong, Kezhu; Kenney, J.B.; Rai, V.; Laberteaux, K.P., "Evaluation of Multi-Channel Schemes for Vehicular Safety Communications," 2010 IEEE 71st Vehicular Technology Conference (VTC 2010-Spring), 16-19 May 2010, pp.1-5.

● Bilstrup, E. Uhlemann, E. G. Ström and U. Bilstrup, “On the Ability of the 802.11p MAC Method and STDMA to Support Real-Time Vehicle-to-Vehicle Communications,” EURASIP Journal on Wireless Communications and Networking, vol. 2009, Article ID 902414, 13 pages, 2009.

● Theses:● Michele Rondinone, “Connectivity-based Routing and Dissemination Protocols for Vehicular Networks,”

PhD Thesis, Universidad Miguel Hernandez de Elche, Spain, November 2013.● Annette Böhm, “Delay-Sensitive Wireless Communication for Cooperative Driving Applications,” PhD

Thesis, Halmstad University, Sweden, June 2013.● Katrin Sjöberg, “Medium Access Control for Vehicular Ad Hoc Networks,” PhD Thesis, Chalmers

University of Technology, Sweden, April 2013.

84

Acknowledgements

● This presentation would not have been possible without lots of input from:

● Katrin Sjöberg, Volvo Group Trucks Technology

● Annette Böhm, Halmstad University

● Le Nam Hoang, Halmstad University

85

Reliable Time-Critical Wireless Communication in … Time-Critical Wireless Communication in...

Documents

Transcript of Reliable Time-Critical Wireless Communication in … Time-Critical Wireless Communication in...