EKF-GPSR: An Extended Kalman Filter for Efficient Routing in Vehicular Networks

Fast advances in wireless technologies and the automotive industry have propelled the proliferation of vehicular ad hoc networks (VANETs). These latter aim to increase road safety and reduce traffic accidents, as well as to provide drivers and passengers with entertainment features. Although, the amalgamated VANETs applications, still face challenging requirements to achieve reliable communication between vehicles or/and vehicles and infrastructure. This is mainly due to their high mobility. Hence, the location information of the vehicles quickly becomes outdated and false. Therefore efficient mechanisms are needed to keep it up to date. The simplest solution consists of increasing the frequency of Hello messages exchanged between vehicles. Unfortunately, this will largely occupy the communication channel and, consequently, may lead to a lot of collisions. In this paper, we propose a novel Extended Kalman Filter Greedy Perimeter Stateless Routing protocol (EKF-GPSR), to address this issue. EKF-GPSR uses a stochastic prediction model based on an EKF to obtain the user information when transmitting data instead of using information from the last beacon exchanged messages which is most likely outdated. The simulation results in Omnet++ and Sumo showed that our proposed protocol EKF-GPSR outperforms the GPSR protocol in terms of packet delivery rate, delay, and throughput.