不确定Euler-Lagrange多智能体系统经验回放自适应蜂拥控制  

Adaptive flocking algorithm for uncertain Euler-Lagrange multi-agent systems with experience replay

作　　者：王希铭 孙金生[1] 李志韬 吴梓杏 WANG Xi-ming;SUN Jin-sheng;LI Zhi-tao;WU Zi-xing(School of Automation,Nanjing University of Science and Technology,Nanjing Jiangsu 210094,China)

机构地区：[1]南京理工大学自动化学院,江苏南京210094

出　　处：《控制理论与应用》2022年第9期1699-1706,共8页Control Theory & Applications

摘　　要：针对具有可参数化线性回归的不确定项的Euler-Lagrange多智能体系统,提出了一种基于经验回放的自适应蜂拥控制算法.在系统模型中的不确定项可以被分解为已知的回归矩阵和未知的回归参数的情况下,该算法通过在线辨识未知参数,降低了传统自适应蜂拥控制算法中估计参数收敛对持续激励条件的要求,可以有效地提高蜂拥系统的性能.利用设计的滤波器,在获得估计参数量与实际参数的误差信息的同时,可以避免使用系统状态的导数信息.本文设计的自适应律不仅保证系统达成蜂拥控制的目标,还通过记录不同时刻的误差信息,使得系统在满足间断激励的情况下,保证估计参数收敛于实际值.通过LaSalle不变集理论对算法进行了分析,给出了理论证明.仿真验证了该算法的有效性.An adaptive flocking algorithm with experience replay is proposed for networked uncertain Euler-Lagrange multi-agent systems.Under the assumption that the uncertain terms can be decoupled into a known regression matrix and an unknown regression vector,the proposed algorithm can significantly improve the performance of the flocking system by identifying the unknown regression vectors online without satisfying the persistent excitation(PE)condition,which is required for the convergence of the estimated regression vectors in traditional adaptive flocking algorithms.The adaptive law designed in this paper not only guarantees the system to achieve the goal of flocking control,but also guarantees the convergence of the estimated parameters to the actual value by recording the error information at different times when the system meets the interval excitation condition.Finally,the algorithm is analyzed and proved by LaSalle’s invariant set theory.Simulation results show the effectiveness of the proposed algorithm.

关 键 词：蜂拥控制 多智能体系统 自适应控制 欧拉拉格朗日系统 经验回放 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]