基于RBF 神经网络的传感器静态误差综合校正方法

以一受环境温度和电源波动影响的压力传感器为例,说明了具体实现方法和校正效果.并与采用BP神经网络进行误差校正的方法进行了比较.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,大大减小传感器静态误差,校正效果优于BP神经网络.

The Static Errors Comprehensive Correcting Method of Sensors Based on Radial Basis Function Neural Network

The Radial Basis Function Neural Network (RBFNN) method is addressed to correct sensors static errors comprehensively. To illustrate achieve method and correcting effect, an example of a pressure transducer that affected by environment temperature and power source fluctuation is presented. A BP Neural Network has been developed to solve the same problem for comparison. The experimental results show that using RBFNN can speed up networks learning speed markedly and reduce sensors static errors greatly and RBFNN is quite effective and superior to BPNN in correcting sensors static errors.