Motion strategy by intelligent vehicles-agents fleet in unfriendly environment

Abstract

This article considers the territory monitoring problems by air vehicles. Vehicles are considered as intelligent agents. A specific feature consists in the inherent antagonism of the motion environment, which is a common situation in practice. Three major conditions of monitoring are formulated, namely, (1) the necessity of repeated solution of the monitoring tasks with varying routes in each cycle, (2) the necessity of online communication among vehicles under their complete independence in decision-making and (3) the possibility of task failure by some vehicles due to constraints imposed by an unfriendly environment. We introduce a group control strategy for a fleet of vehicles performing monitoring. All vehicles-agents receive a given route from a leading agent or calculate and correct the route in the autonomous mode. The efficiency of the suggested approach is demonstrated by monitoring of an emergency situation, viz., a fire in a forest zone approaching a critical object (a nuclear power plant).