基于GSA-BP神经网络的压力传感器温度补偿  

Temperature Compensation of Pressure Sensor by GSA-BP Neural Network

作　　者：黄世震[1] 林淑玲[1] 

机构地区：[1]福州大学,福建省微电子集成电路重点实验室,福州350002

出　　处：《电子器件》2013年第5期680-684,共5页Chinese Journal of Electron Devices

摘　　要：针对压力传感器输出的温度非线性特性,采用遗传模拟退火算法优化的BP神经网络(GSA-BPNN)进行误差补偿,并与标准BP神经网络补偿误差进行了比较。通过MATLAB仿真验证,实验结果表明:采用GSA-BP神经网络不易陷入局部极小值,大大减小传感器非线性误差,误差精度小于0.1%,补偿结果优于标准BP神经网络。In accrodance with the nonlinear ambient temperature output characteristics of pressure sensor. Genetic simulated annealing algorithm to optimize BP neural network（ GSA-BPNN） is proposed for error compensation. A BP neural network has been developed to solve the same problem for comparison. Through the MATLAB simulation,the experimental results show that network not easy to fall into local minimum and errors of the pressure sensors can be greatly reduced by using GSA-BP,the accuracy is less than 0. 1%,the GSA-BPNN is superior to BPNN in correction errors of the pressure sensors.

关 键 词：压力传感器 温度补偿 神经网络 遗传模拟退火算法 MATLAB 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]