Output Consensus of Heterogeneous Multi-Agent Systems with Mismatched Uncertainties and Measurement Noises: An ADRC Approach

Mengling Li,Ze-Hao Wu,Feiqi Deng,Zhi-Liang Zhao
2023-03-03
Abstract:In this paper, the practical output consensus problem for heterogeneous high-order leader-follower multi-agent systems under directed communication topology containing a directed spanning tree and subject to large-scale mismatched disturbances, mismatched uncertainties, and measurement noises is addressed. By introducing a reversible state transformation without changing the output, the actual total disturbance affecting output performance of each agent and matched with the control input of the transformed system is extracted and estimated by extended state observers. Then, the control protocols based on estimates of extended state observers, are designed by combing the output feedback control ones to obtain output consensus and feedforward compensators to attenuating the total disturbance of each agent actively. It is shown with a rigorous proof that the outputs of all followers can track practically the output of the leader, and all the states of the leader-follower multi-agent systems are bounded. Some numerical simulations are performed to verify the validity of the control protocols and theoretical result.
Optimization and Control
What problem does this paper attempt to address?
The paper aims to address the practical output consensus problem of heterogeneous multi-agent systems (MASs) under directed communication topology, facing large-scale mismatched disturbances, uncertainties, and measurement noise. Specifically: - **Research Object**: The paper focuses on high-order leader-follower heterogeneous multi-agent systems. - **Problems Faced**: These systems are affected by mismatched uncertainties, mismatched disturbances, and measurement noise. - **Solution**: By introducing reversible state transformation and using an extended state observer (ESO) to estimate the actual total disturbance, an active disturbance rejection control (ADRC) protocol based on ESO estimation is designed to achieve output consensus and disturbance suppression. - **Theoretical Basis**: The paper demonstrates through rigorous theoretical proof that the output of all followers can track the output of the leader, and all states are bounded. In summary, the paper proposes a new ADRC method to solve the output consensus problem of heterogeneous multi-agent systems under complex environmental conditions.