行波超声波电机单神经元自适应控制

Single Neuron Adaptive Control for Ultrasonic Motor

由于超声波电机具有非线性、时变性、强耦合等特点,增加了它的控制难度,采用PID等常规控制方法难以满足较高精度的控制需求。单神经元作为一种神经网络应用形式,具有自学习和自适应能力,而且易于实现。设计了基于低频PWM方法的单神经元自适应PID转速控制策略,构建了超声波电机自适应闭环控制装置。在一定程度上解决传统PID调节器不易在线实时整定参数、难于对复杂控制对象进行有效控制的不足。实验证明,该控制系统具有响应迅速、适应性强等优点。

As the ultrasonic motor running with non-linear, time-varying, strong coupling, increases the difficulty to control, using the con uentional P2D control method can＇t meet its high-precision eontro requirements. Single neuron as one form of neural network applications, it has self-learning, adaptive ability, simple structure and easy implementation. Used single neuron to adjust PID parameters, synthesize the merits of single neuron and PID. This paper used single neuron adaptive PID control strategy for speed control, constructed ultrasonic motor adaptive closed-loop control device. To some extent, the deficiencies had been resolved for traditional PID parameters not easy online to regulate and difficult to control complex objects effectively. Experiments show that the control system has rapid response, adaptability, and higher accuracy and stability.