Abstract: To address the issues of strong nonlinearity, time-varying parameters, and external disturbances in hub motor control systems, traditional proportional-integral-derivative (PID) control exhibits slow dynamic response and poor disturbance resistance. To improve control performance, this study proposes an adaptive control strategy based on a wavelet neural network (WNN). A feedforward neural network structure using wavelet functions as activation functions is designed and integrated with a PID controller to construct a WNN-PID controller. This structure utilizes the timefrequency localization characteristics of wavelet transform to extract system error features online and adaptively adjusts PID parameters through the online learning capability of the neural network, thereby achieving precise control of the motor speed. Simulation results demonstrate that, compared to traditional PID control, the proposed WNN-PID method significantly improves dynamic response speed, load disturbance resistance, and control accuracy. It effectively enhances the overall performance of hub motor under complex working conditions and provides a valuable reference for the development of high-performance motor control systems.

Key words: hub motor; wavelet neural network; adaptive control; parameter tuning

摘要： 针对轮毂电机控制系统存在的非线性、参数时变与外部干扰等问题，传统比例-积分- 微分（PID）控制动态响应慢、抗扰能力较差。为提高控制性能，研究提出一种基于小波神经 网络（WNN）的自适应控制策略。通过设计以小波函数为激活函数的前馈网络，并将其与 PID 控制器结合，构建 WNN-PID 控制器。该结构利用小波变换提取误差特征，借助神经网络在 线学习能力自适应调整 PID 参数，实现电机转速的精确控制。仿真结果表明，相较于传统 PID， 所提方法在动态响应速度、抗负载扰动能力和控制精度方面均具有显著提升，有效增强了轮 毂电机在复杂工况下的综合性能，对高性能电机控制系统开发具有参考意义。

关键词: 轮毂电机；小波神经网络；自适应控制；参数整定