摘要
针对不确定自由漂浮柔性空间机器人系统,采用模糊CMAC神经网络自学习控制策略来解决轨迹跟踪控制问题.首先建立漂浮基空间机器人的动力学方程,然后利用具有快速学习能力的模糊CMAC神经网络来逼近非线性柔性臂的逆动力学模型.网络参数采用改进的有监督的Hebb学习规则进行自适应在线调整,并通过关联搜索进行自学习和自组织,其误差代价函数由PID控制器提供.仿真结果表明,这种模糊CMAC逆模PID控制器能够达到较高的控制精度,具有一定的工程应用价值.
Considering the uncertainty of free floating adaptable space robot systems (FSRS) , eerebellar model ar ticulation controller(GMAC) neutral network selflearning control strategies are used to solve the trajectory tracking control problems of the inverse model control algorithm. Firstly, a nonlinearity dynamics equation of flexible space ma nipulator is established. The controller based on fuzzy CMAC neutral network is used for effectively learning how to com pensate inversemodel, and fuzzy CMAC network parameters that could be adaptively adjusted online by improved super visory Hebb learning rifles. Error function is provided via proportional integration differential (PID) controller. The con troller improved control accuracy and asymptotic convergence of tracking error, The simulation results illustrate the pres ented controller system has engineering value.
出处
《智能系统学报》
北大核心
2012年第5期457-461,共5页
CAAI Transactions on Intelligent Systems
基金
国家自然科学基金资助项目(61171189)