基于MPSO-BP的超声波电动机Hammerstein模型建模

Modeling of Ultrasonic Motor Based on BP Neural Network Hammerstein Model via Modified Particle Swarm Optimization

超声波电动机由于随着驱动条件不同具有严重非线性以及负载依赖特性,因此其建模成为一件困难并且具有挑战性的工作。文中提出基于改进的粒子群优化BP(MPSO-BP)的超声波电动机Hammerstein模型建模方法,该方法中静态非线性部分采用MPSO-BP神经网络建模,动态线性部分采用一阶传递函数建模。仿真结果显示该方法建立的模型和实验获取的模型较吻合。

Ultrasonic motor have heavy nonlinear and load dependent characteristics which vary with driving conditions,therefore,modeling of USM is a difficult and challenging task. Modeling of USM based on BP neural network Hammerstein model via modified particle swarm optimization was presented. The nonlinear static part is BP neural network model via modified particle swarm optimization（ MPSO） and linear dynamic part is a 1- order transfer function model. The simulation results show the good matching between developed model and experimental measurements.