漂浮基空间机器人的智能融合自适应控制器设计  

Design of intelligent integrated adaptive controllers for free-floating space robotics

作　　者：张文辉[1,2] 齐乃明[2] 马静[3] 肖阿阳[2] 

机构地区：[1]徐州师范大学机电工程学院,江苏徐州221116 [2]哈尔滨工业大学航天学院,黑龙江哈尔滨150001 [3]东北农业大学工程学院,黑龙江哈尔滨150001

出　　处：《华中科技大学学报（自然科学版）》2011年第12期63-67,共5页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：中国航天科技联合技术创新基金资助项目(CAST-HIT09C01)

摘　　要：提出了一种智能融合自适应控制策略用于带外部干扰及模型不确定性的漂浮基空间机器人系统.首先建立不确定空间机器人动力学模型,利用径向基神经网络在逼近域内来补偿模型中出现的未知非线性部分,为保证权值及网络参数在线调节,采用线性化技术将非线性的RBF网络部分线性化,其高阶项量及逼近误差通过自适应鲁棒控制器消除,无须预先估计系统的不确定性程度和外部干扰,包括网络权值和基函数宽度及中心在内的所有参数均能在线实时调整,从而提高了控制精度.该控制器在神经网络控制器的暂时失效的情况下也能保证系统鲁棒性,基于李雅普诺夫理论证明了整个闭环系统信号一致最终有界(UUB).仿真结果表明该智能融合控制器能够达到很好的控制精度.Free-floating space robot with uncertainty of model and external disturbances was considered.An intelligent integrated adaptive control strategy for space robot base on neural network was brought forward.Firstly,space robot dynamic model should be established,and then neural network was used to approximate nonlinear function.Non-linear RBF（radial basis function） neural network was part linear.High rank term and approximate errors were eliminated by adaptive robust controller.So it dispensed with estimates uncertainty of model and exterior disturbance.The network weights and the parameters could be tuned on-line,with no off-line learning phase required.When outside mutation make to invalid neutral network,the controller that can guarantee good robust and the stability of closed-loop system by variable structure controller with good dynamic capability.Based on the Lyapunov theory,the whole closed-loop system was proved as the uniformly ultimately bounded（UUB）.Simulation results show that the control strategy that integrated merits of each controller can achieve higher precision,and has important engineering value to uncertainty space robots.

关 键 词：空间机器人 神经网络 智能控制器 线性化 自适应 鲁棒控制 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]