A Markov model of safety message broadcasting for vehicular networks

Some safety applications in vehicular ad-hoc networks (VANETs) require the dissemination of safety information to all nearby vehicles in a broadcast fashion. Each vehicle should periodically broadcast its state information up to a safety range around itself to avoid likely collisions. This causes a congested channel in dense areas especially in multi-lane roads and leads to significant performance reduction. In this paper, using a Markov model, we analytically derive the percentage of channel utilization as well as the packet transmission rate based on the contention window size, carrier sense range, density of vehicles and packet generation rate. Unlike the previous models, the devised Markov model enables us to derive the probability of packet obsolescence before broadcasting from the probability mass function of service delay. Also, we can evaluate the performance of tracking applications for large and small contention window sizes. The extensive simulations carried out confirm the accuracy of our model.