摘要：无刷直流电机以其高功率密度、高可靠性等优点而广泛应用于各种场合。电机驱动器对控制性能有很大的影响，驱动器采用常规的PID算法难以获得满意的控制效果。为了获得良好的控制性能，本文提出了模糊算法与径向基函数（Radial Basis Function，RBF）神经网络相结合的控制策略。设计了一个四层的神经网络，调整模糊逻辑控制器的输入和输出参数，以参考速度与反馈速度的偏差作为控制的输入，参考电流值作为控制的输出，使得控制系统具备自适应能力。通过仿真和实验对常规PID控制和模糊RBF神经网络算法分别进行了比较，结果表明所提出的方法具有更好的静态和动态响应性能，具有较强的鲁棒性。
关键词：无刷直流电机；径向基函数；神经网络；模糊算法

Abstract: Brushless DC motor (BLDCM) is widely used in many industrial fields because of its high power density and high reliability. The drive has great influence on the motor control performance. It is difficult to obtain satisfactory control characteristics using conventional PID controller. In order to obtain good control performance, the control strategy combined of fuzzy method and basis function radial (RBF) neural network was proposed. A four layer neural network was designed. The input and output parameters of fuzzy logic controller were adjusted. Reference speed and feedback speed were set as input, the output was reference current. The control system has adaptive capacity. Simulation and experiment of conventional PID control method and fuzzy RBF neural network method were carried out respectively. The comparisons show that the proposed method has better static and dynamic response performance, it also has stronger robustness.
Key words: BLDCM；Radial Basis Function(RBF)；Neural Network；Fuzzy control