基于迭代学习的多智能体系统协同编队控制  

Cooperative Formation Control of Multi-agent Systems Based on Iterative Learning

作　　者：顾立雯 王玉龙 马浪 GU Li-wen;WANG Yu-long;MA Lang(School of Mechatronic Engineering and Automation,Shanghai University,Shanghai 200444,China)

机构地区：[1]上海大学机电工程与自动化学院,上海200444

出　　处：《控制工程》2021年第11期2178-2184,共7页Control Engineering of China

基　　金：国家重点研发计划项目课题(2020YFB1708202);科技部“新一代人工智能”重大项目课题(2018AAA0102804);国家自然科学基金项目(61873335,61833011);上海市高等学校特聘教授(东方学者)项目;上海市自然科学基金项目(20ZR1420200);国家111引智基地项目(D18003)。

摘　　要：针对离散形式的智能体动态模型,为减少外部环境干扰的影响,提出一种基于迭代学习的多智能体编队协同控制方法。首先,在考虑外部环境干扰的情况下,建立了多智能体协同编队控制系统离散时间模型,并在通信拓扑设计的基础上,将多智能体系统协同编队控制问题转化成跟踪控制问题。其次,针对每个智能体设计了形式统一的基于迭代学习算法的编队控制协议,并利用范数理论证明了算法的收敛性。随着迭代次数的增加,该算法能够使离散多智能体系统的输出在有限时间内跟踪期望轨迹,从而使得多智能体系统整体实现期望的编队运动。最后,与传统无模型自适应控制方法进行比较,验证了该设计方法的有效性和优越性。For a discrete-time agent dynamic model, in order to reduce the influence of external environment disturbance, an iterative learning-based cooperative control method for multi-agent formation is proposed in this paper. Firstly, the discrete-time model of a multi-agent cooperative formation control system is established in the case of considering external environment disturbance. Based on the designed communication topology,the cooperative formation control problem of the multi-agent system is transformed into a tracking control problem. Secondly, a unified formation control protocol based on an iterative learning algorithm is designed for every agent. The convergence of the algorithm is proved by norm theory. With the increase of the number of iterations, the algorithm enables the output of the discrete-time multi-agent system to track the expected trajectory in finite time, so that the whole multi-agent system can achieve the desired formation motion. Finally,compared with the traditional model-free adaptive control method, the effectiveness and superiority of the design method are verified.

关 键 词：多智能体系统 协同编队控制 迭代学习算法 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]