神经网络在超声波马达控制中的应用现状

Present Status of Neural Network in the Control of Ultrasonic Motor

超声波马达作为一种新型马达具有显著的优点,但由于其输出特性的严重非线性及其数学模型的不精确性,难以实现精确控制,因此影响了超声波马达的实际应用。发达国家及地区已将能对非线性系统实现有效控制的神经网络理论应用到超声波马达的控制中并取得了较好的效果。目前境内有关这方面的研究内容仍是一个空白。在查阅大量文献的基础上较全面的综述了神经网络理论在超声波马达控制研究中的现状。主要介绍了目前所应用的神经网络控制形式、结构、算法及其控制效果,并结合神经网络相关发展做出进一步展望。

The ultrasonic motor (USM) is a newly developed motor with excellent performance and many useful features. But it is difficult to derive mathematical model precisely because of its heavy nonlinearity. Therefore, precise control of USM is also difficult. While the neural network (NN) can be applied to control the nonlinear system with good property, so the neural network is also been used in the USM's control technology. Based on a good number of references and research results, the current status of control technology with NN on USM is described. It mainly includes the neural network control structures, the network topology, algorithm and experimental results. In the end, the prospects of the control technology on USM are discussed from the aspect of interrelated NN's developments.