Transcript Researchers: Farzad Farnoud, Le Zhang, Benham Hassanabadi

VANET Collision-Avoidance and Platooning
using reliable MAC on DSRC
Researchers:
Farzad Farnoud, Le Zhang,
Behnam Hassanabadi, Christine Shea
Project Code:
F202
Supervisor:
Dr. Shahrokh Valaee
Motivation and Background:
In 2002 there were approximately 228,000 injuries and 3,000 deaths caused by motor vehicle accidents in Canada. The leading cause of these accidents is driver error, particularly slow driver reaction time. In a high-speed
highway scenario, this slow reaction time can often lead to catastrophic multi-car pileups. In an effort to reduce the number of vehicular accidents on the road, intelligent transportation systems (ITS) are being developed. One
promising intelligent transportation system uses wireless communication technologies to bring wireless access to the vehicular environment (WAVE). The Dedicated Short Range Communication (DSRC) standard is a wireless
protocol that is in development to allow vehicle to vehicle (V2V) and vehicle to roadside (V2R) communication. This technology will allow vehicles to communicate with one another, and will create Vehicle Ad Hoc Networks
(VANETs) on the road. VANETs will allow cars to send safety messages amongst one another to indicate the presence of accidents and other hazards. In order for these safety applications to run effectively, it is necessary to
have a highly reliable Medium Access Control (MAC) layer, such that vital safety messages aren’t lost. In this project, we have built a simple VANET test bed using developmental DSRC devices from MARK IV Industries as the
physical layer. We have then developed two safety-relevant prototype applications on this test-bed, which include Collision Avoidance and Vehicle Platooning. Finally we present the theoretical basis for a reliable VANET
MAC scheme, as well as simulations.
Robotic Application Test Bed
Set-up:
A robotic test bed was developed to run and test our collisionavoidance and vehicle platooning applications. The design
consists of the following:
 two ER1 Robotics System test vehicles
 two Windows XP laptops (attached to each vehicle)
Physical Layer: DSRC
The Cricket technology provided us with a location
sensing subsystem. The cricket motes use RF and
ultrasound technology to simulate an indoor GPS
system. The two cricket transmitters were mounted
on the ceiling against a wall, and the cricket
receivers were placed on the vehicles. Each test
vehicle was then able to determine its location.
Dedicated Short-Range Communication (DSRC) was introduced in
2003 as a means to facilitate V2V and V2R communication. In
this project, we have collaborated with MARK IV industries, and
they have supplied us with experimental DSRC devices, which
provide the physical layer of our test bed.
The Logitech game pads were used to steer the test
vehicles and turn on platooning mode.
 two wireless Logitech game pads
 two MARK IV DSRC units
 two Cricket mote transmitters (mounted on the ceiling)
Fig. 3 – MARK IV DSRC Device
 two Cricket mote receivers (attached to each vehicle)
Collision-Avoidance
Application
Fig. 1 – A Cricket mote and Logitech Game pad
Vehicle Platooning
Application
Once the test-bed was functional, we
developed a collision-avoidance safety
application. The software was written in
C++, and the basic algorithm is depicted
on the right.
Another safety application that was developed and
tested was vehicle platooning. Vehicle platooning is
an intelligent cruise-control where one vehicle
follows another vehicle at a safe distance while
avoiding a crash. For our test bed, software was
developed to cause a “platooning” vehicle to move
to the location of the other test vehicle, but not hit it.
The vehicles transmit their current
position to one another using DSRC,
and if a collision is predicted, the
vehicles motion is disabled in that
direction.
Fig. 2 – Collision
Avoidance Algorithm
Reliable Medium Access Control Scheme
Theory:
VANET broadcasting occurs under high-mobility with harsh channel conditions. In addition, most VANETs will
operate under the developmental IEEE 802.11p (WAVE) standard. The WAVE standard uses 802.11 technology,
which generates further issues, including a lack of acknowledgments (ACKs) and the Hidden Terminal problem.
All of these issues result in an unreliable MAC, which will allow vital safety messages to be lost.
Fig. 4 – The complete robotic test bed
Simulations:
For the simulations, a Rician fading channel model was used. For N vehicles within
the reception range, each vehicle independently decides to broadcast its location with
probability μp . The probability of success was simulated for each of the SFR, SPR,
and OOC repetition schemes. In addition, QoS levels were simulated by randomly
removing a 1 (repetition) from the lower priority vehicles’ codewords.
The proposed scheme uses a repetition-based broadcast using Optical
Orthogonal Codes. Each frame is divided into L timeslots, and each vehicle
is given a binary code, which represents the repetition pattern. For a given
binary code, a 1 indicates a repetition (the packet should be rebroadcast)
and a 0 indicates that no broadcast should be made.
We examine three different repetition-based schemes, with a focus on Optical Orthogonal Codes. The first
scheme, Synchronous Fixed Retransmission (SFR), retransmits a packet w times per frame, where the w
repetition slots are chosen randomly from the L available timeslots. In the next scheme, Synchronous pPersistent Retransmission (SPR), each of the L timeslots decides to broadcast the packet with probability p, and
decides to remain idle with probability 1-p. The final scheme uses Optical Orthogonal Codes (OOC) to decide the
repetition pattern. OOC’s are desirable because they have a small cross-correlation. For any two codewords x
and y in an optical orthogonal code C, with length L, the maximum cross correlation λ, will be:
Fig. 5: Simulation results for μp =0.3 and N = 61; left and right plots show Probability of
Success vs. distance from the receiver for (L,w) = (64,6) and (94,8) respectively
L
x, y   xi yi   x, y  C
i 1
Given a constant weight (the number of 1’s), w, it is possible to generate OOC codes with a specific maximum
correlation. We can therefore choose this maximum correlation to be 1, and create a code with only one
interfering slot amongst any two codewords. Each of these codewords can then be assigned to a different
vehicle in the VANET. For a specific frame, any two vehicles will only have one interfering retransmission, thus
greatly increasing the probability of successful transmission. A sample code is displayed below:
1100001
0010101
0110010
1000110
0101100
0001011
1011000
Fig. 6: Simulation of different QoS Levels for μp =0.3 and N = 61; left and right plots
show probability of success and delay vs. distance, respectively, for L = 64, whigh = 6,
and wlow =4