非线性伺服电动机的神经网络逆控制

伺服电动机由于存在接触过程的非线性、温漂等非线性因素的影响,很难建立其精确的数学模型,使得基于数学模型的控制困难.针对伺服电动机存在的非线性问题,提出了一种新颖的基于BP神经网络直接逆控制方法.首先,利用BP神经网络建立系统的正向模型(NNI),然后,设计基于神经网络的直接逆控制器(NNC),实现了对伺服电动机的自适应控制.在Lyapunov稳定性分析的基础上,给出了BP算法学习算子的选择方案,保证神经网络权值训练的快速收敛,同时,对训练BP神经网络控制器的专用算法(specialized learning)进行改进,利用NNI的输出求取权值调整的灵敏度函数.数字仿真结果表明提出的控制算法是简单有效的.

Direct Inverse Control of Nonlinear Motor Based on Neural Networks

There are various factors responsible for the nonlinear characteristics of the servo-motor,such nonlinear contact and temperature shift,which cause a great deal of difficulties in establishing the mathematical model of the motor exactly.A direct inverse controller based on BP neural networks has been proposed in this paper to deal with the nonlinearities of the motor.Two neural networks are used in the control system as system identification(NNI) and direct inverse controller(NNC) separately.Upon the analysis of the Lyapunov stability,the rule to choose the learning rate for weights tuning has been achieved.Further,the commonly used specialized learning has been reformed.The output of the NNI has been used to get the sensitivity function during weights tuning instead of motor's, which realizes the control of the system without the exact model.The proposed control method based on the forward neural model has proved to be simple and can reduce the influence of the uncertainty effectively.