摘要
基于奇异摄动将单连杆柔性机械臂动力学模型分解为慢、快变子系统,传统方法分别采用PD控制和最优控制能取得较好控制效果,但负载不确定时,控制效果并不理想。提出对于慢变子系统,采用模糊神经、PD控制相结合的控制方法,对于快变子系统,采用模糊神经、最优控制相结合的控制方法。当负载变化时,采用模糊神经控制器根据实际负载对PD参数及最优控制参数进行调整,达到更优的控制效果。分别采用传统方法及本文提出的改进方法在变负载条件下作了仿真实验,结果表明后者的控制效果明显优于前者。给出了慢、快变子系统模糊神经控制器在变负载条件下训练参数的获取方法。
Based on the singular perturbation approach, the one-link flexible manipulator is decomposed into a slow subsystem and a fast subsystem. The PD control and the optimal control are used in the teaditional method. But the effect is not good when the load varying. A fuzzy-neural and PD controller is designed for the slow subsystem, and a fuzzy-neural and optimal controller is designed to stabilize the fast subsystem. The PD parameter and the optimum controller parameter are adjusted using the fuzzy-neural controller by the varying load, achieving the more superior effect. The simulation results show that the latter surpasses the former. The method of how to obtain training parameter when load varying is given.
出处
《控制工程》
CSCD
北大核心
2009年第6期717-719,786,共4页
Control Engineering of China
基金
国家自然科学基金资助项目(50775002)
关键词
柔性机械臂
模糊神经控制
最优控制
PD控制
变负载
flexible manipulator
fuzzy-neural control
optimal control
PD control
load varying
