An adaptive automobile control system using scheduling by imprecise computation and multiagent-based traffic information exchange and its experimental evaluation

Recently, it is expected that traffic congestion in expressways is dissolved, and that safe and smooth traffic is realized. In this paper we propose an automobile control system based on imprecise computation and multiagents. The automobile control system consists of sensors, mobile stations and automobiles. Each mobile station gathers traffic information through sensors in an expressway, exchanges the traffic information with other mobile stations through a network, and gives driving directions to every automobile in the expressway. Also we evaluate effects on traffic information exchange in the automobile control of the proposed system by simulation experiments. Simulation results show that the traffic information exchange in the proposed system improves performance, that is, the average speed of automobiles and the number of automobiles that arrive at their destinations.