Asynchronous Distributed Model Predictive Control for Multi-Agent Systems

作     者：Mi, Xiaoxiao Liao, Yuanjiang Zeng, Hongzheng 

作者机构：Ningbo Univ Technol Robot Inst Ningbo 315211 Peoples R China Civil Aviat Flight Univ China Key Lab Flight Tech & Flight Safety Civil Aviat Adm China Mianyang 621052 Peoples R China 

出 版 物：《IEEE ACCESS》 (IEEE Access)

年 卷 期：2025年第13卷

页      面：9592-9605页

核心收录：

基　　金：Scientific Research Fund of Zhejiang Provincial Education Department [Y202250282] Open Fund of Key Laboratory of Flight Techniques and Flight Safety, Civil Aviation Administration of China (CAAC) [FZ2022KF17] 

主　　题：Multi-agent systems Optimization Formation control Vectors Predictive control Clocks Trajectory Synchronization Sufficient conditions Stability criteria Asynchronous sampling cooperative control distributed model predictive control formation control multi-agent system 

摘      要：This paper primarily focuses on the asynchronous cooperation control problem of multi-agent systems with external disturbances. Here, all agents take control actions, such as sampling measurements, computing control inputs, and exchanging information based on their individual clocks rather than a global synchronized clock. Owing to the asynchronous nature, the information of each agent received by all its neighbors at the same time instant may be different and deviate from what this agent predicts, potentially resulting in poor decision-making and even instability. To ensure stability, dynamic compatibility constraints related to all the uncertain deviations are established and incorporated into the optimization problem. Additionally, the constraint restriction method is employed to satisfy physical constraints and optimize performance in the presence of external perturbations. Then, an asynchronous distributed model predictive control with a dual-mode strategy is developed and sufficient conditions on the design parameters are constructed to ensure recursive feasibility and system stability. Finally, simulation experiments on a multiple-vehicle system with asynchronous control actions are carried out to validate the theoretical results.