Abstract: The speed of Ultrasonic Motor（USM） has strong nonlinearity which varies with the frequency of two-phase driving voltages，so it is difficult to give an accurate dynamic mathematic model by theory and to control accurately.Based on system identification，a nonlinear Hammerstein model of traveling-wave USM for speed control is built，and the perturbed parameters value bounds of the model is measured.Because the parameter perturbed within a large range，two-step generalized predictive control（TSGPC） for USM is proposed，the main control parameters of the controller are tuned，which provides important dependencies for the speed control experiment of USM.Then the closed-loop stability theoretical analysis and speed control simulation of USM TSGPC are conducted.The theoretical analysis and simulation results show that the TSGPC can track the change of setting value excellently even though the parameter perturbed within a large range.

Key words: ultrasonic motor, Hammerstein model, parameter perturbation, generalized predictive control

摘要： 针对超声电机速度与驱动频率具有很强的非线性，理论建模困难，难以实现高精确控制的问题，采用辨识的方法建立了旋转行波超声电机的Hammerstein模型，并测量了该模型的参数摄动；针对电机参数在大范围内摄动，提出采用两步法广义预测控制方法，整定了广义预测控制算法的参数，并对超声电机两步法广义预测控制进行了稳定性分析和速度控制仿真。理论分析和仿真结果都表明，超声电机Hammerstein模型参数在大范围内摄动时，两步法广义预测控制器能够准确地跟踪速度设定值的变化，证明了超声电机广义预测控制的可行性，为超声电机高精度控制提供了控制模型和控制方法。

关键词: 超声电机, Hammerstein模型, 参数摄动, 广义预测控制