Second-Order Linear Multi-Agent Formation Control Based on Fuzzy Logic System Approximator and Actor-Critic Reinforcement Learning

In this paper, an optimized leader-follower formation control for second-order linear multi-agent systems using the actor-critic reinforcement learning is proposed. An optimal control problem is formulated and solved using a combined actor- critic reinforcement learning and fuzzy logic systems(FLSs) approximator. Actor FLSs are constructed for performing control behavior and critic FLSs are responsible for evaluating the control effectiveness, respectively. Through theoretical analysis, we show that the desired optimal performance can be achieved. Finally, simulation examples are given to demonstrate the effectiveness of the proposed method.