Active Fault-tolerant Control for Multi-agent System in Dynamic Tracking

When carrying out dynamic tracking mission, especially criminal tracking or security patrol, the multi-agent system will face problems such as terrain constraint, communication fading and single-agent failure, which leads to avalanche-type effect to the whole system. An active fault-tolerant control method is proposed to guarantee task execution capability of the multi-agent system, which is a combination of a consistent time-varying formation tracking control and differential evolution-based self-reconstruction for control instructions. And a simulation platform of the multi-agent system is built to test and verify the proposed theory. The results show that the multi-agent system with this active fault-tolerant control strategy can tack a dynamic target even facing communication failure.