基于补偿模糊神经网络的BLDCM伺服控制

为了实现无刷直流电机(BLDCM)位置伺服系统的高精度位置跟踪控制，针对系统多变量、非线性、强耦合、时变的特点，提出了一种基于补偿模糊神经网络控制器(CFNNC)的设计方法.该控制器将补偿模糊逻辑和神经网络相结合，引入了模糊神经元，使网络既能适当调整输入、输出模糊隶属函数，又能借助于补偿逻辑算法动态地优化模糊推理，大大提高了网络的容错性、稳定性和训练速度.仿真和在DSP控制系统上的实验结果表明，采用补偿模糊神经网络控制器，系统响应快、精度高、鲁棒性强，动态特性明显优于传统PID控制.

Brushless DC Motor Servo Control Based on Compensation Fuzzy Neural Network

In order to implement high precision position tracking controlling for the brushless DC motor(BLDCM), a CFNNC (compensation fuzzy neural network controller) algorithm was proposed based on the multivariable, nonlinearity, strong coupling, time-variable characteristics of position servo system. The compensative fuzzy logic and neural network were combined in the proposed algorithm, which could not only adjust the input and output of fuzzy membership functions, but also optimize the fuzzy inference dynamically by using the logic compensation algorithm. The fault tolerance, stability and working speed of the network were improved greatly due to the introduction of fuzzy neuron. The simulation and experiment results of DSP-based control system illustrated that this method has rapid response, high precision and robustness, and its dynamic characteristic was much better than that of traditional PID controller.