Event-based distributed cooperative learning control for discrete-time strict-feedback multi-agent systems

Abstract

This paper proposes an event-based distributed cooperative learning (EBDCL) control method for a type of discrete-time strict-feedback multi-agent systems. The multi-agent systems have the same nonlinear dynamics, but different tracking control tasks. A new estimate-error-based neural network (NN) update law is designed to avoid the $n$-step delay disadvantage of combining the classical tracking-error-based NN update law with event-based communication in previous works. The stability of the system is proved using the Lyapunov stability theorem and error transformation method. Moreover, all the NN weights are proved to converge to small neighborhoods of their ideal values in a limited domain along the union trajectories of all agents, which can be stored as experience knowledge. Hereafter, the experience knowledge is reused to design the experience-based control scheme, which improves the control performance. Finally, simulation results are presented to demonstrate the efficacy of the EBDCL method.