hybrid robust model predictive based controller for a class of multi-agent aerial dynamic systems

The decentralized control of a multi-agent system with leader-follower consensus is investigated. the system is formulated in graph theory, and a general configuration for l-f formation is proposed. the goal for the formation is defined to track the predefined trajectory in the presence of high-frequency noise. the controller for the system is proposed on the basis of a model predictive-based controller. different scenarios for a multi-agent system are considered, which lead to the linearization of the plant. meanwhile, external structured disturbances are considered in the system. the novelty of the present paper addresses the gap between optimal controllers and robust controllers. the robustness of optimal controllers is not verified in the optimality of mpc controllers. thus, a tube mpc theory is proposed to increase the robustness of the interacting noise system. consequently, the optimal controller maintains robust throughout the existence of external disturbances and high frequency noises. meanwhile, the closed-loop multi-agent response is investigated in the presence of external bounded disturbances. next, the hybrid controller is designed for the formation. the switches take place between mpc and tube-mpc controllers for each agent. on the other hand, hard constraints on control input and its variations and soft constraint on graph structures and topology of the multi-agent system are submitted. at length, the stability proof is considered for the closed loop multi-agent system. finally, the simulation results demonstrate the formation results and the proposed controller can also satisfactorily deal with the high-frequency noise with hard and soft constraints.