Congestion management in vehicular ad-hoc networks

Abstract

We study the performance of the adaptive PI (Proportional Integral) rate controller [HoYa07] to deal with the problem of vehicle traffic congestion and data traffic congestion in vehicular ad-hoc networks. Derived from classical control theory, the adaptive PI rate controller performs active queue management by controlling the source rate based on the instantaneous queue length of the congested node. Using simulations we have demonstrated the capability of the adaptive PI rate controller to deal with vehicle traffic and data traffic congestion in vehicular ad-hoc networks. For data traffic congestion management, we have evaluated the performance of the algorithm in both static and mobile network scenarios. The controller dominated the combined performance of TCP NEWRENO and RED (Random Early Detection) queue with a positive influence on source throughput. For vehicle traffic management we have studied the congestion condition that arises when vehicle traffic volume exceeds the capacity of the road. The adaptive PI rate algorithm controlled the traffic volume in the road segment of interest to a preconfigured target threshold additionally improving the vehicle speed. Various other performance measures have been evaluated.