Quantized model-free adaptive iterative learning bipartite consensus tracking for unknown nonlinear multi-agent systems

Abstract

This paper considers the data quantization problem for a class of unknown nonaffine nonlinear discrete-time multi-agent systems (MASs) under repetitive operations to achieve bipartite consensus tracking. Here, a quantized distributed model-free adaptive iterative learning bipartite consensus control (QDMFAILBC) approach is proposed based on the dynamic linearization technology, algebraic graph theory, and sector-bound methods. The proposed approach doesn’t require each agent’s dynamics knowledge and only uses the input/output data of MASs, where the data is coded by the logarithmic quantizer before being transmitted. Moreover, we consider both cooperative and competitive relationships among agents. We rigorously prove the stability of the proposed scheme and analyze the effects of data quantization. Meanwhile, we demonstrate that data quantization does not affect the stability of MASs, and bipartite consensus tracking errors can converge to zero with the processing of the proposed scheme, although the data quantization slows the convergence rate. Furthermore, the results are extended to switching topologies, and three simulation studies further validate the effectiveness of the designed method.