the fully networked car geneva, 3-4 march 2010 1 spatial diversity for ieee 802.11p v2v safety...
Post on 27-Mar-2015
216 Views
Preview:
TRANSCRIPT
The Fully Networked Car Geneva, 3-4 March 2010
1Spatial Diversity for IEEE 802.11p V2V SafetyBroadcast in a Highway Environment
Nor Fadzilah Abdullah, Robert Piechocki and Angela Doufexi
Centre for Communications Research,University of Bristol, UK.
The Fully Networked Car Geneva, 3-4 March 2010
Need for Vehicular Communication…
o In 2007, the EU recorded ~43,000 deaths and >1.8 million injuries (€160 billion loss)
o Steady growth of car usage and ownership (>200 millions cars in Europe) • congestion built-up, unpredictable journey time• impact on the economy: significant vehicle operating
costs overhead, burden for travellers• impact on the environment: harmful emissions, worsen
air quality
o Allocated bandwidth for C2X serviceso Lowering cost of WiFi and GPS
2
European Road Safety Observatory, Annual statistical report 2007. [Online] http://euroris.swov.nl/safetynet/xed/WP1/2007/SN-1-3-ASR-2007.pdf
The Fully Networked Car Geneva, 3-4 March 2010
Research Contribution
o Vehicular communication requires longer communication range (than 802.11a/g/n), in extreme multipath and high speed environment• Spatial diversity: a low complexity and low cost solution
o Accurate and realistic vehicular communication modelling by means of:• BER curves from detailed PHY simulator specific to
modulation types, vehicular speeds and range of SNR values
• Integration of PHY simulator and realistic mobility model into network simulator
3
The Fully Networked Car Geneva, 3-4 March 2010
4Scenario: Post-crash warning in highway environment
o Realistic mobility traces: 3 lanes bidirectional highway.
o 2 types of traffic density models (Low & High).
o 2 types of ad-hoc V2V safety messages• Emergency message• Periodic message
o Rayleigh channel with 103ns rms delay spread (ETSI channel B)
D. W. Matolak, I. Sen, W. Xiong, and N. T. Yaskoff, “5GHZ Wireless Channel Characterization for Vehicle to Vehicle Communications,” Proceedings of IEEE Military Communications Conference (MILCOM ’05), vol. 5, pp. 3022–3016, Atlatnic City, NJ, USA, Oct 2005.
The Fully Networked Car Geneva, 3-4 March 2010
5Midamble symbol spacing as a function of channel coherence time, data rate, and packet size
Space-time correlation,Midamble spacing chosen: 30 symbols
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
x 10- 3
-0.5
0
0.5
1
Time(s)
Co
rrel
atio
n c
oeffi
cien
t
500 bytes data using IEEE 802.11p QPSK 1/2 (6Mbps)
Jakes PSD (fd=1100Hz)Total pkt duration (84sym=0.672ms)Tc = 0.225ms (28sym)Jakes model PSD (fd=550Hz)Tc = 0.44ms (55sym)
The Fully Networked Car Geneva, 3-4 March 2010
6Frame Structure for Proposed Multi Antenna System with Midamble Channel Estimation
o Channel tracking: reuses long preamble sequence for midamble channel estimation
S. I. Kim, H. S. Oh, and H. K. Choi, "Mid-amble Aided OFDM Performance Analysis in High Mobility Vehicular Channel," IEEE Intelligent Vehicles Symposium, Eindhoven, Netherlands, Jun 2008.
The Fully Networked Car Geneva, 3-4 March 2010
7Need for midamble in fast fading vehicular channel
0 5 10 15 20 25 3010
-3
10-2
10-1
100
SNR (dB)
PER
Midamble vs. Preamble at 50 km/h
SISO: preambleSISO: midambleSTBC 2x2: midambleSTBC 4x4: midamble
o Midamble vs. Preamble at 10 symbols midamble spacing
SISO: Single Input Single OutputSTBC: Space Time Block CodesModulation: QPSK 1/2
Antenna configuration
Rank
SISO 1
STBC 2x2 1
STBC 4x4 3/4
The Fully Networked Car Geneva, 3-4 March 2010
Low density
STBC 2x2 vs. SISO
STBC 4x4 vs. SISO
No priority
60% (200m/125m)
132%(290m/125m)
EDCA 80% (225m/125m)
164%(330m/125m)
8Low Density Traffic: Emergency Message Packet Delivery Ratio in Rayleigh channel
Low density traffic:6 vehicles/km/lane
* EDCA (Enhanced Distributed Channel Access)
0 100 200 300 400 5000
10
20
30
40
50
60
70
80
90
100
Distance (m)
Pack
et
Deliv
ery
Ratio (
%)
Emergency broadcast with Interference: Low Density
SISO, LoD, no prioritySISO, LoD, EDCASTBC 2x2, LoD, no prioritySTBC 2x2, LoD, EDCASTBC 4x4, LoD, no prioritySTBC 4x4, LoD, EDCA
The Fully Networked Car Geneva, 3-4 March 2010
High density
STBC 2x2 vs. SISO
STBC 4x4 vs. SISO
No priority
70% (170m/100m)
65%(165m/100m)
EDCA 50% (195m/130m)
138%(310m/130m)
Low density
STBC 2x2 vs. SISO
STBC 4x4 vs. SISO
No priority
60% (200m/125m)
132%(290m/125m)
EDCA 80% (225m/125m)
164%(330m/125m)
9High Density Traffic: Emergency Message Packet Delivery Ratio in Rayleigh channel
High density
STBC 2x2 vs. SISO
STBC 4x4 vs. SISO
EDCA improvement
15% (195m/170m)
88%(310m/165m)
Low density
STBC 2x2 vs. SISO
STBC 4x4 vs. SISO
EDCA improvement
13% (225m/200m)
14%(330m/290m)
High density traffic:11 vehicles/km/lane
0 50 100 150 200 250 300 350 400 450 5000
10
20
30
40
50
60
70
80
90
100Emergency broadcast with Interference: High Density
Distance (m)
Pack
et
Deliv
ery
Ratio (
%)
SISO, HiD, no prioritySISO, HiD, EDCASTBC 2x2, HiD, no prioritySTBC 2x2, HiD, EDCASTBC 4x4, HiD, no prioritySTBC 4x4, HiD, EDCA
The Fully Networked Car Geneva, 3-4 March 2010
10Conclusion
o Performance of safety broadcast messages, for MIMO-STBC vs. SISO in a vehicular environment has been presented.
o Spatial diversity increase the communication range: 50-80% for STBC 2x2 and 65-164% for STBC 4x4 case.
o Traffic prioritization (EDCA) is efficient in high density scenario and extends the communication range by 15% for STBC 2x2 case and 88% for STBC 4x4.
The Fully Networked Car Geneva, 3-4 March 2010
11Spatial Diversity for IEEE 802.11p V2V SafetyBroadcast in a Highway Environment
Appendix
The Fully Networked Car Geneva, 3-4 March 2010
Vehicular Channel Model
o Differing maximum Doppler shifts for low and high density traffic
o RMS delay spread of 103ns [Matolak, 2005]
12
0 10 20 30 40 50 60 70 80-18
-16
-14
-12
-10
-8
-6
-4
-2
0
Number of OFDM symbols per frame
Norm
aliz
ed p
ow
er
(dB
)
Vehicular time-correlated multipath fading channel at fd=550Hz
path1
path2path3
path4
path5
path6path7
path8
D. W. Matolak, I. Sen, W. Xiong, and N. T. Yaskoff, “5GHZ Wireless Channel Characterization for Vehicle to Vehicle Communications,” Proceedings of IEEE Military Communications Conference (MILCOM ’05), vol. 5, pp. 3022–3016, Atlatnic City, NJ, USA, Oct 2005.
0 10 20 30 40 50 60 70 80-25
-20
-15
-10
-5
0
Number of OFDM symbols per frame
Norm
aliz
ed p
ow
er
(dB
)
Vehicular time-correlated multipath fading channel at fd=1100Hz
path1
path2path3
path4
path5
path6path7
path8
The Fully Networked Car Geneva, 3-4 March 2010
13Physical Layer Simulator Block Diagram
The Fully Networked Car Geneva, 3-4 March 2010
Numerical Analysis Parameters 14
Physical Layer MAC Layer
Tx Frequency: 5.9 GHz Slot Time: 13 us
Bandwidth: 10 MHz OFDM symbol: 8 us
Tx Power: 23 dBm PLCP: 40 us
Receiver threshold: -82 dBm SIFS: 32 us, CWmin: 31
Antenna gain: 0 dBi DIFS: 58 us, BO = 208 us
Antenna height : 1.5 m EDCA: High Priority (EM)
Channel Model: Rayleigh ECWmin: 7
Modulation scheme: QPSK 1/2 AIFS: 58 us, BO = 52 us
Application Layer EDCA: Low priority (PM)
Pkt Generation Rate: 10 pkt/s ECWmin: 31
Packet size: 500 bytes AIFS: 123 us, BO = 208 us
The Fully Networked Car Geneva, 3-4 March 2010
15Midamble symbol spacing as a function of channel coherence time, data rate, and packet size
0 5 10 15 2010
-3
10-2
10-1
100
SNR (dB)
PER
SISO vs MIMO: 100 km/h, with midamble spacing of 30 symbols
SISO: 100 bytesSISO: 500 bytesSISO: 1000 bytesSTBC 2x2: 100 bytesSTBC 2x2: 500 bytesSTBC 2x2: 1000 bytesSTBC 4x4: 100 bytesSTBC 4x4: 500 bytesSTBC 4x4: 1000 bytes
Lower SNR requirement for higher spatial diversity and smaller payload size.• STBC 4x4 reduces maximum data rate
The Fully Networked Car Geneva, 3-4 March 2010
16Midamble symbol spacing as a function of channel coherence time, data rate, and packet size
0 10 20 30 40
10-4
10-3
10-2
10-1
100
SISO, 100 km/h, with midamble spacing of 30 symbols
SNR (dB)
BER
BPSK 1/2BPSK 3/4QPSK 1/2QPSK 3/416QAM 1/216QAM 3/464QAM 2/364QAM 3/4
Higher SNR requirement for higher modulations.
top related