受扰多智能体系统的分布式主动抗干扰协调控制研究

Multi-agent systems are usually subject to various kinds of disturbances, including both matched and mismatched disturbances. Disturbance rejection property is one of the key indices of distributed cooperative control methods for multi-agent systems. Due to the superiorities in disturbance attenuation, active anti-disturbance control has been used for distributed cooperative control of disturbed multi-agent systems. However, research on this control problem is still on its initial stage and has several bottlenecks, e.g., the existing control methods are limited to systems with matched disturbances, the convergence rates of the closed-loop systems are not fast enough, and the existing control methods are only suitable for full state-feedback cases, etc. To improve the aforementioned bottlenecks, enhance the disturbance rejection property and speed up the convergence rates of the closed-loop systems, and enlarge the application scope of active anti-disturbance control methods, this project consists of the following research contents for nonlinear multi-agent systems with mismatched disturbances: 1) distributed active anti-disturbance cooperative control for nonlinear multi-agent systems with various mismatched disturbances; 2) distributed finite-time active anti-disturbance cooperative control for nonlinear multi-agent systems with mismatched disturbances; 3) distributed active anti-disturbance cooperative control via output feedback for nonlinear multi-agent systems with mismatched disturbances; 4) the theoretical results are applied to formation control of multi-mobile robot systems and collaborative assembly control of multi-hydraulic manipulator systems, and the proposed control algorithms will be verified on the multi-mobile robot experimental platform and the multi-hydraulic manipulator half-practicality simulation platform, respectively. The research tasks of this project will provide effective and practical theory and methods for high-precision distributed cooperative control for nonlinear systems subject to complicated disturbances, and have great theoretical and practical significance.

多智能体系统中往往含有各种干扰，包括匹配、不匹配干扰。抗干扰性能是协调控制算法的关键性能指标，主动抗干扰控制因在该方面的明显优势，已被初步用于受扰多智能体系统的协调控制。然而，该方面研究存在诸多瓶颈，如已有方法局限于匹配受扰系统、收敛速度不够快、仅适用于全状态反馈情况等。为突破上述瓶颈，提高系统抗干扰性能和收敛速度，并拓宽主动抗干扰控制方法的适用范围，本项目将针对不匹配受扰非线性多智能体系统，展开以下几方面的研究：1) 不同结构特征不匹配受扰情况下的分布式主动抗干扰协调控制；2) 分布式有限时间主动抗干扰协调控制；3) 基于输出反馈的分布式主动抗干扰协调控制；4) 将理论成果应用于多移动机器人编队控制并完成实验验证，应用于多液压机械臂协同装配控制并完成半实物仿真验证。本项目研究将为复杂受扰非线性多智能体系统的高精度分布式协调控制提供切实可行的理论和方法，具有重要的理论和实际意义。

针对受扰多智能体系统的一致性问题，项目负责人提出了“分布式主动抗干扰协调控制”理论和方法框架：针对各受扰智能体设计干扰估计器，然后分布式利用不匹配干扰估计信息，并结合非线性反馈控制方法，形成分布式主动抗干扰协调控制方案。该方案能有效解决不匹配受扰多智能体系统的(输出)一致性问题，且能快速补偿和处理干扰对系统的影响，使闭环系统有强抗干扰性能和鲁棒性。上述成果将为复杂受扰非线性多智能体系统的高精度分布式协调控制提供切实可行的理论和方法，具有重要的理论和实际意义。.项目主要研究内容包括：(1)针对受不同结构特征不匹配干扰影响的二阶、高阶非线性多智能体系统，研究了分布式主动抗干扰协调控制问题；(2)针对不匹配受扰二阶、高阶非线性多智能系统，研究了分布式有限时间主动抗干扰协调控制问题；(3)针对含不可测状态的不匹配受扰二阶、高阶非线性多智能体系统，研究了基于输出反馈的分布式主动抗干扰协调控制问题；(4)在面向实际对象的应用设计方面，针对受扰多高超声速飞行器系统，研究了速度和高度一致性问题。.在项目资助下，负责人共发表录用论文21篇，其中SCI论文13篇，包括发表在控制领域顶级和权威期刊IEEE Trans. on Automatic Control、Automatica、IEEE Trans. on Cybernetics等，EI会议论文8篇。以共同编辑出版Springer编著1部、章节1章。申请发明专利2项。培养博士生5人，硕士生6人。新增国家自然科学基金面上项目1项，航空科学基金1项，博士后科学基金面上、特别资助项目各1项。于2018年入选“东南大学‘至善青年学者’支持计划(A层次)”、“江苏省科协青年科技人才托举工程”，获2018年度山东省高等学校科学技术奖一等奖(排3)，获南京市第十二届自然科学优秀学术论文奖，2017年获SCI期刊JFI杰出审稿人称号。作为副导师指导的硕士生李桂璞获2018年江苏省和东南大学优秀硕士学位论文。