Vehicular Networking for Intelligent and Autonomous Traffic Management

With the increasing number of vehicles on the roads, traffic congestion has become a frustrating daily routine. Valuable time and human energy are wasted everyday by vehicles stuck in traffic. In addition, traffic congestion causes excessive fuel consumption and high dose of pollution, which add to the negative economic impact on the nation and health risk on the citizens. As means for mitigating the effect, many agencies provide live traffic reports using data gathered from cameras and sensors on road, and also go to the extent of suggesting alternatives for congested routes. However, such an approach does not help when the majority of the vehicles follow the same suggestions and the alternate route gets congested too.
This project promotes a novel approach for mitigating congestion to the best possible extent using autonomous and adaptive data exchange between vehicles. The proposed approach attempts to balance the load on the roads and prevent

congestion, unlike other techniques which only respond after it takes place. Every vehicle chooses a route to its destination and generates a set of optional routes. By leveraging the VANET technology, the system on each vehicle queries other vehicles in its neighborhood on whether they are travelling on any of the edges that it is using or the edges on its optional routes. Based on the overall data collected from as many vehicles as possible, at every specific unit of time, the system would make a decision to continue on the same route or suggest an alternate route. The combined effect of the entire set of vehicles would impact the traffic pattern and allow smooth and faster travel for everyone.
The challenges in VANETs, namely, being potentially large scale, highly dynamic, and subject to intermittent connectivity, actually turn out to be advantageous in our approach. Our system uses real-time data from neighboring vehicles to make a route decision. Moreover, the routes which others vehicles take at every unit of time influence a vehicle's own decision, implying that the system takes advantage of the highly dynamic vehicular network to make a fine-grained analysis and yield more optimized travel route. This technique also prevents congestion and saves time for all other travelers. The same data when shared with traffic signals would help in improving the vehicle throughput by allowing high traffic flow on roads and accommodating increased vehicle density.
Recent Publication
  1. S. Gupte and M. Younis, "Vehicular Networking for Intelligent and Autonomous Traffic Management," in the Proceedings of the IEEE International Conference on Communications (ICC'12), Ottawa, Canada, June 2012 (to appear).
  2. M. Eltoweissy, S. Olariu, and M. Younis, "Towards Autonomous Vehicular Clouds," in the Proceedings of the 2nd International Conference on Ad Hoc Networks (ADHOCNETS 2010), Victoria, BC, Canada, August 2010.