超声波电机迭代学习自适应控制策略

Iterative Learning-adaptive Control Strategy of Ultrasonic Motor

电机控制策略通常都以被控电机的数学模型为其设计的必要基础。用于控制器设计的电机模型与实际电机之间必然存在偏差。在电机控制系统的使用过程中,考虑到实际电机的特性会随着时间的推移而变化,电机模型与实际电机特性之间的偏差会逐渐增大,导致控制性能偏离设计要求。本文以超声波电机转速控制系统为研究对象,通过在控制系统中增设可调增益,给出一种基于迭代学习思想来在线校正模型偏差的方法。不同模型偏差程度情况下的实验结果表明了所述方法的有效性。在迭代的学习过程中,可变增益不断变化以减小实际输出与模型输出之间的偏差,从而使电机控制系统的实际性能符合设计要求。

The control strategies of motor are usually based on the mathematical model of the controlled motor. There will be a deviation between the motor’s model used for controller design and the actual motor. In the process of using the motor control system, considering that the characteristics of the actual motor will change over time, the deviation between the motor’s model and the actual characteristics will gradually increase, resulting in the deviation of the control performance from the design requirements. In this paper, the speed control system of ultrasonic motor was taken as the research object. By adding adjustable gain in the control system, a method to correct the model deviation online based on the idea of iterative learning was proposed. The experimental results under different degrees of model deviation show the effectiveness of the method. In the iterative learning process, the adjustable gain changes constantly to reduce the deviation between the actual output and the model output, so that the actual performance of the motor control system meets the design requirements.