基于扩张状态观测器的异构多智能体编队控制  

Heterogeneous Multi-agent Formation Control Based on Expanded State Observer

作　　者：陈世东 庹先国 韩强 CHEN Shidong;TUO Xianguo;HAN Qiang(School of Automation and Information Engineering,Sichuan University of Science&Engineering,Yibin 644000,China;Artificial Intelligence Key Laboratory of Sichuan Province,Yibin 644000,China)

机构地区：[1]四川轻化工大学自动化与信息工程学院,四川宜宾644000 [2]人工智能四川省重点实验室,四川宜宾644000

出　　处：《四川轻化工大学学报(自然科学版)》2025年第1期39-46,共8页Journal of Sichuan University of Science & Engineering(Natural Science Edition)

基　　金：国家自然科学基金项目(61902268);四川省科技厅项目(21ZDYF4052,2020YFH0124,2021YFSY0060);四川轻化工大学创新计划项目(cx2023193,cx2023195,cx2023198)。

摘　　要：针对复杂环境下异构多智能体编队协同控制问题,提出了一种基于扩张状态观测器(ESO)的异构多智能体编队协同控制方法。该异构智能体编队主要由无人地面车辆(UGV)以及无人飞行器(UAV)组成。首先,建立单个UGV动力学模型以及单个UAV动力学模型,在此基础上构建了异构多智能体编队动力学模型。其次,针对异构编队系统存在的未知扰动和模型参数不确定性,设计了一种新的扩张状态观测器,该观测器在传统扩张状态观测器的基础上引入积分项来估计模型参数不确定和系统未知扰动。基于积分扩张状态观测器(IESO),设计了动态编队控制协议,并给出误差在有限时间内收敛的证明。最后,与基于传统扩张状态观测器设计的控制方法仿真相比,所提出方法在X方向上的误差在7 s左右已趋于零,Y方向上的误差在6 s左右已趋于零,仿真结果表明所提方法观测精度更高,收敛速度更快,抗干扰能力更强。To address the problem of heterogeneous multi-agent formation cooperative control in complex environments,a cooperative control method for heterogeneous multi-agent formation based on the expansion state observer(ESO)is proposed,which consists of unmanned ground vehicles(UGV)and unmanned aerial vehicles(UAV).First,a dynamic model for a single UGV and a single UAV is established,upon which the dynamic model for the heterogeneous multi-agent formation is constructed.Next,to address the unknown disturbances and model parameter uncertainties present in the heterogeneous formation system,a novel ESO is designed,which introduces an integral term into the traditional ESO to estimate model parameter uncertainties and unknown system disturbances.Based on the integral expansion state observer(IESO),a dynamic formation control protocol is designed,and the convergence of the error within a finite time is proven.Finally,compared with the control method designed based on the traditional ESO,the proposed method shows that the error in the X direction tends to zero in about 7 s and the error in the Y direction tends to zero in about 6 s.Simulation results indicate that the proposed method has higher observation accuracy,faster convergence speed,and stronger anti-interference capability.

关 键 词：扩张状态观测器 异构多智能体 编队控制 UAV/UGV编队 

分 类 号：V279[航空宇航科学与技术—飞行器设计] U463[机械工程—车辆工程]