基于径向基函数神经网络的空间漂浮机械臂装配控制  

Assembly control of a space floating manipulator based on radial basis function neural network

作　　者：刘育强[1,2] 魏庆生 李浩然 魏承[1] 赵阳[1] LIU Yuqiang;WEI Qingsheng;LI Haoran;WEI Cheng;ZHAO Yang(School of Astronautics,Harbin Institute of Technology,Harbin 150001,China;Beijing Institute of Spacecraft System Engineering,Beijing 100094,China)

机构地区：[1]哈尔滨工业大学航天学院,黑龙江哈尔滨150001 [2]北京空间飞行器总体设计部,北京100094

出　　处：《哈尔滨工程大学学报》2023年第5期831-836,共6页Journal of Harbin Engineering University

基　　金：国家自然科学基金项目(11772102)。

摘　　要：针对空间漂浮机械臂在轨装配中动力学模型失准、待装配模块质量特性未知情况下的精确控制问题,本文提出了一种基于径向基函数神经网络补偿控制的装配方法。给出空间漂浮装配机械臂的构型设计及动力学模型描述,设计了一种基于径向基函数神经网络原理的补偿控制率,进行了稳定性与收敛性分析,推导出基于径向基函数神经网络神经网络的空间漂浮基座机械臂装配控制自适应律。不同工况下的在轨装配仿真案例表明:基于径向基函数神经网络原理补偿的控制方法能够控制机械臂末端执行器在未知特性大负载作用下快速准确地跟踪期望轨迹,而不进行补偿时控制器性能下降严重、甚至失效。所提出装配控制方法具有较强的工程实用价值,能为未来空间装配任务提供有效借鉴。A compensation control method based on a radial basis function neural network(RBFNN)is proposed.RBFNN focuses on the problem of the accurate control of space floating manipulators in orbit assembly when the dynamic model is inaccurate and the quality characteristics of the module to be assembled are unknown.First,the configuration design and dynamic model description of the space assembly manipulator are given.Then,a compensation control rate based on the principle of RBFNN is designed,and stability and convergence are analyzed.Finally,the RBFNN-based adaptive assembly control law of space floating base manipulators is derived.Simulation cases of in-orbit assembly under different working conditions show that the control method based on the RBFNN compensation principle can control the manipulator end effector to track the desired trajectory quickly and accurately under the action of heavy load with unknown characteristics.However,in the absence of compensation,the performance of the controller declines seriously or even fails.The method proposed in this paper has high engineering practical value and can provide an effective reference for future space assembly tasks.

关 键 词：空间机械臂 空间应用 动力学模型 控制系统分析 控制系统稳定性 神经网络 装配技术 仿真分析 径向基函数 

分 类 号：V47[航空宇航科学与技术—飞行器设计]