Interface selection in hybrid V2V communications: A hierarchical approach

Abstract

The emergence of new types of applications for connected cars is expected to increase data traffic in vehicular networks. A promising solution is to concurrently use multiple types of radios, dynamically selecting network interface(s) that are likely to achieve the highest performance (i.e., hybrid networking). In this paper, we propose an intelligent interface selection mechanism, tailored to hybrid vehicle-to-vehicle communications. We introduce a novel approach of hierarchical decision making, in which interface selection by vehicles is loosely controlled by a remote central server. The server provides vehicles with recommended interface selection strategy, which is optimized based on the statistical knowledge about road and network conditions. While the vehicles basically follow the recommended strategy, they are allowed to adjust it if they detect actual channel conditions deviating from the statistics on the server. The adjustments are carefully made based on the expected channel load information provided by the server, so that the changes of strategy do not harm the communication performance in neighboring geographical regions. Simulation results show that our solution can significantly reduce the end-to-end latency of packet delivery, compared to conventional selection mechanisms.