摘要
针对载体位置不受控,姿态受控的漂浮基空间机器人执行机构部分失效故障问题,提出了一种基于有效因子融合的分散滑模神经网络容错控制方案。利用拉格朗日第二类方程建立系统的动力学模型,然后利用分散的思想,将系统分散为若干个子系统,将有效因子融合到子系统的动力学模型中。利用终端滑模进行控制器设计,并利用径向基函数神经网络对系统参数不确定项和未知项进行估计,自适应地补偿了神经网络的估计误差。最后,数值仿真结果表明了提出的分散容错控制器的有效性。
Based on decentralized sliding mode neural network control,a method of decentralized fault-tolerance control for space robots with partial loss of actuator effectiveness was proposed. Based on Lagrange second dynamics equation,the dynamic model of space robots was established. Furthermore,the system was decentralized into some subsystems and then an effective factor was integrated into the dynamic model of each subsystem. Terminal sliding mode techniques was used to design the controller,and with the help of radial basis function(RBF) neural networks,uncertainty terms and interconnection terms were estimated with adaptive compensation of estimation error. Finally,simulations verified the effectiveness of this decentralized fault tolerant controller.
出处
《载人航天》
CSCD
2016年第1期39-44,共6页
Manned Spaceflight
基金
国家自然科学基金资助项目(11372073
11072061)
关键词
漂浮基空间机器人
分散容错控制
有效因子
部分失效故障
径向基函数神经网络
free-floating space robot
decentralized fault-tolerantce control
effective factor
partial loss of actuator effectiveness
RBF neural network