Adaptive learning control for group consensus tracking of discrete nonlinear multiagent systems

Abstract

In this article, we explore the group output tracking consensus problem for discrete‐time strict‐feedback ‐order nonlinear multiagent systems that run repeatedly on finite time . A novel distributed adaptive iterative learning group consensus protocol is designed, which consists of two main components. The first component is based on time‐varying neural networks, which is used to approximate the unknown nonlinear function in the ‐step ahead predictor. In general, not all followers can access the information regarding the leader, which complicates the design of iterative learning protocols for MASs. Therefore, the second component of the protocol addresses this challenge by treating the leader's output as a time‐varying parameter and designing a time‐varying auxiliary term to compensate the leader's output information. Parameter updating laws and initial state learning laws are also proposed via the cooperative‐competitive relationship between the agents. We demonstrate the group consensus with sufficient small errors can be achieved at time , as the number of iterations proceed to infinity. Then, the results are extended to the case of multisubgroups and multileaders. Finally, two simulations validate the findings of this article.