电液伺服系统的神经网络并行控制  

作　　者：武交峰 

机构地区：[1]广东技术师范学院天河学院,广东广州510540

出　　处：《中国西部科技》2009年第8期17-19,共3页Science and Technology of West China

摘　　要：文章针对电液伺服系统的非线性、不确定性和参数时变的特点,提出一种基于神经网络的并行自学习控制算法。该控制策略以系统动态误差和给定信号量作为小脑模型神经网络控制器的激励信号,并与单神经元自适应PID控制器相结合构成系统的复合控制。仿真结果表明,该并行控制算法较常规PID控制具有更快的响应特性和良好的动态特性,对模型参数变化和负载扰动表现出更强的适应性和鲁棒性。A parallel self-learning control strategy based on neural network,with respect to the complex nonlinearities and uncertainties as well as parameters time-variation characteristic of electro-hydraulic servo system is presented.Combining with the single neuron adaptive PID controller, the multiplex control strategy uses the dynamic errors and given signals of the system as input parameters to the Cerebellar Model Articulation Controller（CMAC）.By comparison with the conventional PID controller,the simulation results show that the proposed parallel control scheme is of better adaptability and higher robustness while the mathematic model changing or the external load disturbancing.

关 键 词：小脑模型关节控制器 单神经元PID控制器 电液伺服系统 神经网络控制 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TP271.31[自动化与计算机技术—控制科学与工程]