Geo-based Backend Dissemination for Safety-relevant V2X Applications

Abstract:
Safety-relevant V2X applications can be enhanced using communication, such that even objects out of the vehicle’s sensor range are known by the vehicle. This communication congests the communication channel, which is relieved by reducing the dissemination of these messages. However, this decreases the performance of safety-relevant V2X applications, as the knowledge of the vehicles is limited. In this work, we increase the awareness range of vehicles by distributing information over larger distances without increasing the channel load. Our approach is based on the assumption, that only those messages should be exchanged, which are of value for the application. Thus, we develop a utility function for a V2X Intersection Collision Warning application which aims at capturing the relevance of a location-message for a specific vehicle. For this application, the relevance is defined as the probability that two vehicles meet at the intersection. Based on this utility, location-messages are exchanged stochastically. In the evaluation, we show that our approach reduces the channel utilization by up to 68% in the downlink channel, while similarly increasing the awareness range of vehicles. Additionally, the update frequency of vehicles meeting at the intersection is comparable to the baseline dissemination strategy.