MULTIAGENT PLANNING OF INTERSECTION PASSAGE BY AUTONOMOUS VEHICLES

Abstract

We propose a traffic management system for autonomous vehicles that are agents at the intersection. In contrast to the known solutions based on the usage of semiautonomous control systems in assembly with the control unit, this algorithm is based on the principles of decentralized multiagent control. The best travel plan for intersection passage is produced by means of optimization methods jointly by all agents belonging to a dynamic collaboration of autonomous vehicles. The order of road intersection optimal for a given criterion is determined by the agents in the process of information exchange about themselves and environment. Our experiments show that this protocol can reduce significantly the traffic density as compared to the traditional systems of traffic management. Moreover, the effectiveness of the proposed algorithm increases with increasing density of road traffic. In addition, the absence of the critical object, that is the control unit, in the control system, reduces significantly the effectiveness of possible failures and hacker attacks on the intersection control system.

Keywords

• multiagent system
• autonomous vehicles
• cyber-physical system
• road traffic
• traffic density
• intersection management