Improving the Car GPS accuracy using V2V and V2I Communications

Abstract

This paper addresses the problem of vehicular localization on the road and proposes a stochastic solution that leverages vehicle-to-vehicle communication as well as the knowledge that vehicles acquire regarding their approximate locations. Such knowledge is inferred from generated GPS readings together with distance measurements calculated using the beacons broadcasted periodically by other neighboring vehicles. Furthermore, the proposed solution methodology also adopts the locations of stationary RoadSide Units (RSUs) as fixed reference points that help in determining the locations of vehicles whenever these vehicles navigate through the RSUs’ coverage ranges. It is shown here that the additional position measurements received from neighboring vehicles lead to a remarkably accurate estimate of the position of a certain target vehicle. An analytical framework is established in this paper with the objective of formulating the target vehicle’s position estimate problem using particle filters. The validity, reliability and accuracy of the presented mathematical formulae are verified through extensive simulations using a combination of the Network Simulator (NS-3) and the Simulation for Urban MObility (SUMO) that were used to generate realistic vehicular mobility traces.