Consensus iterative learning control for multi-agent systems with non-repetitive perturbations

For an especial multi-agent system (MAS) characterized by non-repetitive parameter perturbations, this brief proposes a composite iterative learning control (ILC) algorithm to achieve complete trajectory tracking of the system. Firstly, employing 2-D system theory, the ILC system model consisting of an uncertain multi-agent system and ILC law is transformed into an typical uncertain 2-D Roesser system. Secondly, a controller with a real-time state information feedback term was designed to this system, featuring characteristics of feedback on the time axis and feedforward on the iteration axis, thus constituting a composite ILC scheme with a control effect in two directions for the ILC system. Subsequently, utilizing Lyapunov stability principles and LMI methods, some sufficient conditions are obtained for robust stability of the system under the action of iterative learning control, and feasible solutions for the learning gain matrix are derived.