Simulation evaluations of cooperative relative positioning around intersections

Position information plays an important role in support system for safe driving, where inter-vehicle distance is used to prevent collision accidents from happening. In urban areas, however, multipath propagation of GPS signals degrades the accuracy of absolute positions and the deduced relative position. On the other hand, simply removing reflected signals might lead to a shortage of satellites in fixing positions. In our previous work, we suggested cooperative relative positioning to improve the accuracy of relative positions of adjacent vehicles so as to prevent collision accidents from happening. In this paper, we focus on intersections, which are risky areas for vehicles where accidents occur with a higher probability than other areas, then investigate whether our scheme can help to improve the accuracy of relative positions at intersections. More specifically, we evaluate multipath errors due to reflection/diffraction and analyze their correlation between receivers around intersections. We find that (i) Our system helps to improve the accuracy of relative positions of vehicles running on the same road, (ii) The spatial correlation of positioning signals decreases as vehicles get nearer to intersections. As a result of the initial step to focus on intersections, only relying on our scheme is insufficient and other measures become necessary as well in order to get accurate relative positions.