Recursive-Gradient-Based Data-Driven Consensus Control for Multiagent Systems

Abstract

The paper presents an improved recursive-gradient-based model-free adaptive control (IRGBMFAC) method which can increase the convergence rates of the loss function and improve the convergence performance of the system compared with the basic MFAC. Then IRGBMFAC is introduced into the consensus control of multi-agent systems. Due to the external environment, communication failure, data packet loss and other problems occur randomly, the communication topology of multiagent system may change inevitable. Here, the fixed topology and the switched topology are both considered. Finally, the effectiveness and feasibility of the IRGBMFAC have been demonstrated by the simulation examples.