Vehicular AdHoc Networks

1
Vehicular Ad-Hoc Networks

• Martin Mauve, Holger FüßlerUniversity of Mannheim

Hannes HartensteinNEC Europe
2
FleetNet in a Nutshell

• Development of a platform for inter-vehicle
  communication
• Approach
• wireless mobile ad-hoc networking
• Goal
• improve safety and comfort
• Project Duration
• 9/2000 12/2003

Partly funded by
3
Overview

• Project Outline
• Network Characteristics
• Beaconless Position-Based Routing
• Conclusion

4
FleetNet Applications

• Cooperative Driver Assistance
• warnings about dangerous situations
• overtaking assistance
• Decentralized Floating Car Data
• avoiding traffic jams
• route weather forecast
• Entertainment
• Access Network to the Internet

5
Why Ad-Hoc?

• Specifics of Ad-Hoc-Networks
• Well suited for spontaneous local communication
• Direct communication, low latency
• Robust no single point of failure
• ... and it is (almost) free!
• Specifics of Inter-Vehicle Communication
• (almost) unlimited battery power
• high processing speed
• use of external antennas possible
• geographic position available

6
Problem Areas

• Physical
• Medium Access
• Routing
• Application
• Business/Marketing

7
Position-Based Routing

• Idea
• every node knows its geographic position
• every node knows the position of its neighbors
  (by means of beacons)
• source requests position of the destination
• packets are forwarded in the direction of the
  destination
• repair strategy for local optima
• examples face-2, GPSR, Terminodes routing, etc.
• Key benefit for vehicular ad-hoc networks
• no route needs to be constructed maintained
• robustness against topological changes

8
Network Characteristics Evaluation

• Prior Work
• random-waypoint scenario
• use this as input for ns-2
• no consideration of network specifics
• Needed
• realistic movement patterns
• Problem
• generation of realistic movement patterns for
  vehicles is difficult!
• Solution
• cooperate with vehicle manufacturers
  (DaimlerChrysler)!

9
Network Characteristics I
? 20 m ?
? 4000 m ?
? 20 m ?
10
Network Characteristics II
Probability for Reachability
Bidirect. Unidirect.
Distance between Sender and Receiver (m)
11
Comparison DSR/GPSR I
Delivery Rate
Distance between Sender and Receiver (m)
12
Beacons

• Problems caused by using beacons
• outdated information about neighbors
• routing-loops
• unknown neighbors
• overhead for transmission of beacons
• Is it possible to eliminate beacons?

13
Position-Based Routing without Beacons

• So far
• forwarding node knows its neighbors
• ... and selects its successor
• Idea
• broadcast message
• successor selects itself
• Problem
• there should be only one successor
• contention
• suppression

14
Contention

• Every neighbor sets a timer
• the better it is suited, the smaller the runtime
• If the timer expires
• neighbor is successor
• suppresses other neighbors

15
Simple Suppression

• Successor forwards the packet
• All other neighbors cancel their timer
• Problem

16
Area-based Suppression

• Idea
• select an area such that all nodes are in radio
  range of each other
• only nodes in that area compete
• then there will be only one successor
• Problems
• how to determine the area?
• what happens if the are does not contain nodes?

17
Area-Based Suppression Circle
Destination
Forwarder
18
Area-Based Suppression Reuleaux Triangle
19
Conclusion

• Inter-vehicle communication as a promising
  application area for mobile ad-hoc networking
• Encompasses all layers
• Very demanding (dynamics!)
• Many open issues (security? Routing in cities?)
• First prototype experiments this week