基于RBF神经网络的传感器非线性误差校正方法  

Nonlinear errors correcting method of sensors based on RBF neural network

作　　者：侯立群[1] 张智娟[1] 仝卫国[1] 

机构地区：[1]华北电力大学自动化系,河北保定071003

出　　处：《传感器技术》2004年第3期43-45,共3页Journal of Transducer Technology

基　　金：:华北电力大学青年教师基金资助项目(09320015).

摘　　要：介绍了利用人工神经网络进行传感器非线性误差校正的原理。提出了传感器非线性误差校正的径向基函数(RBF)神经网络方法,并与采用BP神经网络校正非线性误差进行了比较。最后给出了一个仿真实验,实验结果表明:采用RBF神经网络可以明显提高网络收敛速度,大大减小传感器非线性误差,校正效果优于BP神经网络。The principles for correcting the nonlinear errors of the sensors with a neural network are introduced. The method of radial basis function neural network (RBFNN) is given to correct the nonlinear errors of the sensors. A BP neural network has been developed to solve the same problem for comparison. The experimental results show that network learning speed can be sped up markedly and nonlinear errors of the sensors can be greatly reduced by using RBFNN. RBFNN is quite effective and superior to BPNN in correcting nonlinear errors of the sensors.

关 键 词：RBF神经网络 传感器 非线性误差校正 径向基函数 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置] TP183[自动化与计算机技术—控制科学与工程]