超声电机非线性多步预测自校正转速控制

Self-tuning Nonlinear Generalized Predictive Speed Control of Ultrasonic Motors

基于电机动态响应过程的实验数据,建立了以电机驱动频率为输入、转速为输出的超声电机驱动系统哈默斯坦(Hammerstein)非线性模型,用粒子群算法辨识模型参数。在该模型的基础上,求取模型非线性部分的逆,使超声电机的主要非线性特性得到在线补偿;随后,设计了超声电机非线性多步预测自校正转速控制策略,并给出了控制器参数的整定方法。该策略采用滚动预测、优化及在线自校正方法,以应对超声电机的未建模非线性。扰动情况下的实验,表明了所提控制策略的有效性及鲁棒性。

Based on the tested data of ultrasonic motor＇s dynamic response, the non-linear Hammerstein model of an ultrasonic motor driving system is proposed. The model＇s parameters are identified using the particle swarm optimization method. The input variable of the model is driving frequency of motors, the output is the rotating speed. Using this Hammerstein model, the inversion formula of the nonlinear part can be directly obtained. Therefore, the motor＇s main nonlinearities depicted by the nonlinear part of Hammerstein model can be compensated online. Thereafter, the self-tuning nonlinear generalized predictive speed control strategy of ultrasonic motors is proposed. The tuning method of controller＇s parameters is also discussed in detail. This control strategy utilizes roll forecast, optimization and online self-tuning methods to deal with the unmodeled nonlinearity of ultrasonic motors. The experiments with disturbance indicate the validity and robustness of the proposed strategy.