Optimal Data Rate Selection for Vehicle Safety
Communications
ACM Workshop VANET 2008
Daniel Jiang, Qi Chen, Luca Delgrossi
Mercedes-Benz Research and Development North America,
Inc.
Sep 15th 2008
Contents
1. Overview of Data Rate in 802.11p
2. Theory and Methodology
3. Simulation Results and Summary
Data Rate Selection In VANET Communications
54Mbps
The best selection of
data rate depends
on the scenario
Data Rate
choices
For IEEE
802.11p
(Mbps)
27
24
18
11
6
12
1
9
100m
200m
300m
6
4.5
Data rate vs. operating distance for MAC 802.11 a/b/g
•Is there a data rata most suitable for DSRC?
•If so, how do we find out the most suitable one?
3
Choose
Influential Parameters of a DSRC Scenario
Data Rate
Packet Size
Vehicle Density
Transmission Power
Number of Lanes
Message Frequency
Influential Parameters of a DSRC Scenario
Data Rate Supported in IEEE 802.11p
Impact of the Data
Rate Selection
• Higher rate leads to
shorter MAC Frame
transmission duration
• Higher rate makes
MAC frame reception
more difficult and
prone to errors
IEEE802.11p OFDM PHY Parameters
Contents
1. Overview of Data Rate in 802.11p
2. Theory and Methodology
3. Simulation Results and Summary
Methodology of Finding the Most Suitable Data Rate
3-step procedure to figure out the
most suitable data rate:
Challenges:
Step 1 (Make different
parameter combinations)
How to select representative values of
the 6 influential parameters.
Create scenarios with different
combinations of transmission
parameters
Step 2 (Compare combinations
)
How to make sure scenarios with different
parameters can be compared fairly?
Compare the performance of
different combinations
Step 3 (Choose best
combination)
Select the combination with
highest performance
What’s the criterion of highest
performance?
Concept of Communication Density
Data Rate
Data Rate
Packet Size
Vehicle density
Transmission power
Number of Lanes
Packet Size
Communication
Density
Message frequency
4 into 1
Communication
Density
CD = Transmission Power * Message Frequency *
Nr. Lanes * Vehicle Density
•A system is defined by 6 parameters
<Vehicle Density, Number of Lanes,
Message Frequency, CD value,
Packet Size, Data Rate>
•A system can be now defined by
a triplet
<CD value, Packet Size, Data
Rate>
Understanding of Communication Density
•Channel
load at a location is determined by the number of transmission
sensible there and their durations.
•Number of sensible transmissions is related to the number of nodes
around the location and their transmission power and frequency.
(CD is a simple metric value, please see Communication Density paper in
Movenet2007)
•The
duration of a sensible transmission is determined by the MAC frame
size and the data rate being used.
Node B
Node C
T1
Channel Busy
Channel Clear
Node A
t1
t2
CSMA Channel Status of Node A
T2
t
Use CD to Measure the Channel Load
1.
Channel load is determined by a triplet <CD, data rate, message
size>
CD = Transmission Power * Message Frequency * Nr. Lanes * Vehicle Density
2.
3.
Power, Message Frequency and Vehicle Density are inter-changable
CD is additive. Two groups of nodes with CD1 and CD2 are mixed
together, their total CD = CD1 + CD2
How to measure?
Attach a reference group,
Both groups yield the same Channel Load
as long as their CD values are the same
If the reference group
keeps the same
performance, the tested
group shall produce the
same level of load to the
reference group
Transmission Performance Under a Channel Load
Level
1.
Performance of a transmission is determined by a triplet <CD, size,
data rate>, together with the power being used.
Range= 100m
Range= 300m
Tested Group
Reference Group
2.
Performance degrades under high channel load
Range= 500m
Answers to the Challenges
How to select representative values of the 6 influential parameters?
(using Communication Density as a metric to combine 4 parameters)
How to make sure scenarios with different parameters can be compared
fairly?
(use an attached reference group to test if the tested group
generates the same level of channel load, even though its
parameters have been selected in different combinations)
What’s the criterion of highest performance?
(the best transmission performance in the tested group, in terms of
successful reception probability)
Contents
1. Overview of Data Rate in 802.11p
2. Theory and Methodology
3. Simulation Results and Summary
Simulation Settings
Simulator
NS-2 version 2.33
802.11Ext models with
a completely new 802.11 Mac and Phy model
MAC&PHY
Parameters
Configured for 802.11p, including the SINR
decoding requirement for different data rates
RF Model
Rayleigh RF model
Simulator Environment
Simulation Settings (Cont.)
•9
•
•
channel load levels
[6Mbps]
System CD level: (3 options) 200, 400,
800
Packet Size (3 options): 100B, 200B,
500B
Data Rate
Packet Size
[100B, 200B, 500B]
Communication
Density
[200, 400, 800]
Defining Parameter Combinations
•In
each channel load level, parameters are varied in combinations of
<range, message frequency, lanes, vehicle density,data rate, message
size>
•
•
Data rate: (5 options) 3, 4.5, 6, 9, 12 Mbps
Transmission Range: (3 options)100m, 300m, 500m
For each channel
load level with
<CD,size,6Mbps>
System Load
System Load is divided
into 3 groups
Making 15 different combination of
the parameters of the study group
Study Group (60%CD)
Reference Group2 (20%CD)
Reference Group1 (20%CD)
Varying with15 combinations
Using reference group to test the
study group maintains the same
channel load contribution
Transmission Performance in a Given Channel Load
Performance of Study Group
Performance of Reference Groups
Study Group (60%CD)
Reference Group2 (20%CD)
Reference Group1 (20%CD)
Simulation Result
Study Group Range = 100m
Study Group Range = 300m
Study Group Range = 500m
Best combination with 4.5Mbps
Best combination with 4.5Mbps
Best combination with 6Mbps
Performance of reference group1 matches to each other
Performance of reference group 2matches to each other
Channel Load of the System <CD200,100B,6Mbps>
Power Range of the Study Group
100m
300m
500m
Data rate of the best combination
4.5
4.5
6
6
6
4.5
Data rate of the 2nd best
combination
Simulation Result (Cont.)
Channel Load Level
Study <200,
Group 100B,
Power 6M>
<200,
200B,
6M>
<200,
500B,
6M>
<400,
100B,
6M>
<400,
200B,
6M>
<400,
500B,
6M>
<800,
100B,
6M>
<800,
200B,
6M>
<800,
500B,
6M>
100m
4.5
6
6
6
6
6
6
6
6
300m
4.5
6
6
6
6
6
6
6
9
500m
6
6
6
6
6
9
6
9
9
Summary
1.
6Mbps is the most optimum data rate for DSRC communications
across a general spectrum overall channel load and transmission
powers. This result removes one dimension for future research.
2.
Triplet <total CD values, message size, data rate> is an simple but
effective value to indicate channel load level.
3.
Proposed new methodology “COMBINE, COMPARE, CHOOSE”
allows the further optimization study of communications parameters.
Questions and Discussions