基于RBF网络的滑模变结构控制在无刷直流电机伺服系统中的应用

无刷直流电动机的参数强耦合、高度非线性特性增加了对其速度控制的难度，针对这一特点，设计了一种神经网络滑模变结构速度控制器。将无刷直流电机速度伺服系统分成名义模型和不确定模型，采用状态反馈方法对名义模型进行控制，以RBF神经网络为滑模动态补偿器对不确定系统进行控制。该方法不仪具有变结构控制的抗参数摄动、抗干扰以及速度快等优点，神经网络控制的加入还有效地减弱了单纯滑模变结构控制所带来的“抖振”现象。Matlab仿真结果表明，该控制器具有良好的控制性能和鲁棒性能。

Application of Sliding Mode Variable Structure Control Using an RBF Neural Network in the Servo System of Brushless DC Motor

The brushless DC motor has characters of high non-linearity and closed linkage for parameters, which make its speed control difficuh. Directed to this condition, a sliding mode variable structure controller using an RBF neural network is designed. The speed servo system is composed of nominal model and uncertain model with state feedback controller designed for nominal model and uncertain model compensated by sliding controller using an RBF neural network. It has such advantages as strong ability of resisting parameter distortion, rejecting chattering and fast response. Because of neural network control, it also effectively decreases the chattering of only using a sliding mode variable structure control system. The matlab simulation shows this controller has good performance and robustness.