Quantized Iterative Learning Control for Consensus of Nonlinear Impulsive Multi-Agent Systems With Inter-Channel Encoding-Decoding Mechanisms and Packet Dropouts

Abstract

This paper addresses the consensus problem of impulsive multi-agent systems (MASs) under bandwidth constraints and packet dropouts, and a distributed quantized iterative learning control (ILC) scheme based on intermittent update principles is designed. Unlike traditional packet dropouts models in MASs where dropouts are characterized by simultaneous across the entire network, we have established a novel framework featuring inter-channel independent packet dropouts processes. Based on this framework, a distributed learning control algorithm that leverages the quantized signals from encoding and decoding mechanisms, incorporates an inter-channel quantizer, and accounts for the packet dropouts characteristic. Using mathematical tools such as the Lyapunov equation, impulsive Gronwall inequality, Cauchy inequality, and the fixed-point theorem, we rigorously derive the consensus result of tracking error in the sense of expectation. Furthermore, this paper presents the connections between the scaling sequence, the learning step size, and the packet dropouts rate, avoiding the divergence of the quantized values in the iterative process. Finally, the effectiveness of the theoretical results is verified by the simulation of the multi-pendulum networked system.