异构不确定二阶非线性多智能体系统事件触发状态趋同  

Event-triggered State Consensus of Heterogeneous Uncertain Second-order Nonlinear Multi-agent Systems

作　　者：李振兴 庄娇娇 杨成东 邱建龙 曹进德 LI Zhen-Xing;ZHUANG Jiao-Jiao;YANG Cheng-Dong;QIU Jian-Long;CAO Jin-De(School of Automation and Electrical Engineering,Linyi University,Linyi 276000;School of Mechanical and Vehicle Engineering,Linyi University,Linyi 276000;School of Information Science and Engineering,Linyi University,Linyi 276000;School of Mathematics,Southeast University,Nanjing 211189)

机构地区：[1]临沂大学自动化与电气工程学院,临沂276000 [2]临沂大学机械与车辆工程学院,临沂276000 [3]临沂大学信息科学与工程学院,临沂276000 [4]东南大学数学学院,南京211189

出　　处：《自动化学报》2025年第4期804-812,共9页Acta Automatica Sinica

基　　金：山东省青创团队计划(2022KJ114);山东省自然科学基金(ZR2021Q E299,ZR2022MF222);云南省科技厅重大科技专项计划(202302AD 080006)资助。

摘　　要：研究异构不确定二阶非线性多智能体系统事件触发状态趋同控制问题.首先,为每个智能体设计参数观测器,用以估计不确定参数,这些观测器可渐近估计不确定参数.其次,为每个跟随智能体设计分布式参数观测器,渐近估计领导智能体不确定参数,每个智能体利用邻居智能体触发时刻的采样值估计其邻居智能体的状态.基于估计的参数和邻居状态,提出完全不依赖智能体间连续信息传输的事件触发趋同算法.同时,证明在所给算法的作用下,多智能体系统能够达到状态趋同且不存在芝诺现象.最后,给出一个多单摆系统,用以验证事件触发趋同算法.仿真实验结果表明,跟随智能体的位置和速度可以渐近跟踪领导智能体的位置和速度,并且整个多智能体系统平均每秒触发8.825次.对比仿真实验显示,所提出的事件触发趋同算法可有效减少事件触发次数.This paper studies event-triggered state consensus of heterogeneous uncertain second-order nonlinear multi-agent systems.Firstly,to deal with the uncertain parameter of each agent dynamics,parameter observers are designed for each agent to asymptotically estimate uncertain parameters,these observers can estimate exact values of uncertain parameters.Secondly,distributed parameter observers are designed for each follower agent to asymptotically estimate uncertain parameters of leader agent,and each agent use its neighbors'sampled information at the trigger instant to estimate its neighbors'states.Based on estimated parameters and neighbors'states,an eventtriggered consensus algorithm is proposed,which does not rely on any continuous information transmission among agents.Furthermore,it is proved that multi-agent systems can reach state consensus under the proposed algorithm,and there is no Zeno-behavior.Finally,a multi-pendulum system is given to verify event-triggered consensus algorithm.The simulation experiment results show that the position and velocity of follower agents can track that of leader agent asymptotically,and the average trigger number of overall multi-agent systems is 8.825 per second,comparison simulation experiment shows that the proposed event-triggered consensus algorithm can effectively reduce event-triggered number.

关 键 词：多智能体系统 事件触发 不确定二阶非线性系统 状态趋同 异构系统 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TP273[自动化与计算机技术—控制科学与工程]