Analysis of actuator faults in a cooperative team consensus of unmanned systems

Abstract

This work presents results on performance analysis of a team of agents in the presence of team members faults for three types of faults, i.e. loss of effectiveness, float, and lock-in-place faults. The team goal is to accomplish a cohesive motion in a modified leader-follower architecture using a semi-decentralized optimal control. This controller that is recently proposed by the authors is designed based on minimization of individual cost functions using local information. It is shown that a loss of effectiveness (LOE) fault does not deteriorate the stability or the consensus seeking goal of the team and would only result in a different transient behavior, e.g. a change in the agent's convergence rate, without a change in the consensus value. On the other hand if the fault in one or more of the agents is of the float type, either in the leader or the followers, the team does not maintain its consensus anymore but stability of the team can be guaranteed. Moreover, the leader and the healthy followers adapt themselves to the followers changes when a float fault occurs in one of the agents. Finally, the behavior of the team in the presence of a lock-in-place (LIP) fault is also discussed. Simulation results are provided to demonstrate the performance of the team in the presence of the above faults.