The aim of this paper is to propose a collaborative perception system in order to avoid potential vehicle collisions. We present a basic system where vehicles ensure their own localization using a global positioning system with a Kalman filtering process. This information and the velocity data are broadcasted via a wireless network to its neighbours. Moreover, some of the vehicles are equipped with telemetric sensors, thus they can supply a direct perception of their surrounding environment and update the knowledge given by the network. If a potential collision is detected, the vehicles are brought under speed control. We analyse the results and the limits of this system. All results are validated via an extended perception simulator.
Collaborative perception for collision avoidance
2004
6 Seiten, 13 Quellen
Conference paper
English