Event-Triggered Iterative Learning Control of Regular Time-Varying Multi-Agent Systems Under Switching Topology

An event-triggered iterative learning control algorithm is proposed to address the consensus problem of regular time-varying multi-agent systems under switching topology, while considering the insufficient resource space of the system and the output saturation constraint phenomenon. Firstly, the algorithm utilizes the pseudo partial derivative estimates and output estimation errors to design an output observer to overcome the output constrained in the communication network. Secondly, the output estimation error of the observer and the trigger function are used to design the event trigger condition, and when the trigger function value satisfies the event trigger condition, the state values of the agents are updated; otherwise, the state values of the agents will remain unchanged. The gain error of the output observer is used as a variable to design the deadband controller function to avoid the Zeno phenomenon effectively. Then, the control algorithm utilizes the pseudo partial derivative estimation value to adjust the proportion of consistency error in real time, thereby continuously correcting the control input. Under the condition that both the pseudo partial derivative estimation and observer output estimation errors are bounded, the control algorithm proposed in this paper can enable the system to fully track the desired trajectory without the need for real-time updates of state information. Finally, the effectiveness of the proposed control algorithm is further verified by simulation cases.