改善传感器非线性的自适应分区神经网络方法

Adaptive partition method of improving sensor's nonlin ear using neural networks

提出了一种传感器特性反函数的自适应分区神经网络辨识方法,这种方法根据学习准确度的要求,用自适应方法把传感器特性的反函数分成若干区间,分别用一个BP神经网络去学习辨识进而取代,从而实现传感器特性的线性化。实验结果表明,非线性误差可减小到原来的十分之一。

A method that the inverse function of sensor's characteristic is identified with adaptive partition using neural netw orks.Using this method,the inverse function of sensor's characteristic can be partitioned into many intervals with adaptive method according to requis ition of learning accuracy using BP neural networks,and the interval function is learned and identified in the each interval respectively using a BP neural net work,after that,the each BP neural network replace the corresponding inverse fun ction in the each interval,and the sensor characteristics linearization are real ized.The results of simulations and application have shown that the nonlinear er rors of the sensor are decreased nine tenths using this method.