Iterative learning control for multi-agent systems consensus tracking

Abstract

In this paper, under repeatable operation environment, an iterative learning control (ILC) scheme is applied for multi-agent systems (MAS) to perform consensus tracking, where the underline communication graph is assumed to be fixed and directed. Different from many existing consensus schemes for linear agent dynamics, we consider time-varying nonlinear agent models with non-parametric uncertainties. Furthermore, the desired consensus trajectory is only known to a subset of the agents. By virtue of the repetitiveness of tracking task and the learning ability of each agent, the proposed ILC scheme enables all agents to achieve the asymptotic output consensus in the iteration domain and perfect tracking in the time domain simultaneously. Moreover, owing to the associated initial state learning controller, the proposed consensus scheme does not require the identical initial conditions, henceforth, making it more applicable in practice. In the end, an illustrative example is provided to demonstrate the efficacy of the consensus scheme. © 2012 IEEE.