An Improved Multi-Agent Model-Free Adaptive Iterative Learning Consensus Control under Data Dropouts

Abstract

Aiming at the problem of multi-agent consensus tracking under output data dropout, a model-free adaptive iterative learning control scheme for multi-agents, and a data compensation method are proposed. The phenomenon of data dropout is described as a Bernoulli sequence with the known probability, and a compensation algorithm for data dropout is proposed, that is, using the known output data, the estimated value of the pseudo gradient and the control input difference to compensate for the lost data. Then, the convergence analysis of the algorithm is given for the compensation algorithm proposed. The effectiveness and superiority of the algorithm is verified through the simulation of multi -agent system with the fixed topology.