基于多智能体系统的点状化仿生机翼建模与控制

This project mainly studies the modeling and control of spotted bionic wings based on multi-agent systems. The purpose of this project is to achieve theoretical analysis methods and experimental verification for ascent, hover and flight of the bionic wings. Main contents include the following three parts: (1) Spotted model of the wings is established in the framework of topology graph theory, in which the joints are regarded as nodes and the bones are regarded as paths. The correspondence between coordinated vibration of the nodes and aerodynamic lift is given, and furthermore, the performance analysis of the wings can be developed. (2) Based on the spotted model, dynamic behavior control problem is considered. To achieve rapid bionic matching, finite-time coordinated vibration control methods are investigated with amplitude restrictions. Besides, switching topologies are introduced into the control problem of bionic wings, for which invariance principles are developed to achieve switching between different wing structures. (3) The theoretical model and control methods with dynamic behavior are verified by simulation experiments. We expect the achievements of this project can provide theoretical analysis method and technical support for spotted bionic wings, which is of great significance both in theoretical and practical applications.

本项目拟采取多智能体系统的分析方法研究仿生机翼的建模与控制问题。目的是给出仿生机翼的举升、悬停和推进三种运动形式的理论分析和实验验证。主要研究内容包括以下三点：(1)建立机翼的点状化模型，依据拓扑图理论，分别将关节和连接关节的骨架看作节点和边，给出多节点协调振动与空气升力之间的对应关系，分析其运动形式。(2) 针对点状化模型，研究其动态行为要求下的控制问题，拟提出满足振幅要求的节点有限时间协调振动控制方法，实现机翼动态的快速仿生拟合，并将切换拓扑图的概念引入到仿生机翼的控制中去，提出适合切换系统的不变集分析方法，从而实现突发状况下不同机翼类型的快速切换。(3)对仿生机翼的理论模型和动态行为要求下的控制方法进行仿真与实验验证。本项目的成果期望为点状化仿生机翼的研究提供理论分析方法和有效的技术支撑，具有重要的科学意义和应用价值。

仿生飞行是当前自动控制、人工智能领域的热点研究问题之一，具有广阔的应用前景。本项目结合多智能体系统协同控制理论与流体力学数值计算方法，重点研究了流动气体中多智能体系统的建模与控制问题。在空气中，受控个体之间不仅仅是受到简单的几何拓扑约束，而且会受到气体反作用力的影响。针对这一特点，提出了粘性多智能体系统的建模方法，揭示了固体运动与气体运动之间的相互作用机理。提出了曲率约束下的协同轨迹优化方法，建立优化指标评价体系，确定了路径最优解。基于时变气体扰动，给出了分布式系统鲁棒性分析方法。本项目研究成果为处理复杂分布式流固耦合多智能体系统的建模与控制问题提供了新思路和新方法。在项目资助下，负责人发表录用SCI论文8篇，包括控制领域权威期刊Automatica、IEEE Trans. on Cybernatics、IEEE Trans. on Systems, Man, and Cybernetics: Systems等。获IEEE人工生命与机器人国际会议最佳论文奖。