基于PSO-LSSVM模型的扩散硅压力传感器的温度补偿  

Temperature Compensation for Diffused-silicon Pressure Sensor Based on PSO-LSSVM

作　　者：杨婷 卢文科[1] 左锋[1] 

机构地区：[1]东华大学信息科学与技术学院,上海201620

出　　处：《仪表技术与传感器》2017年第12期25-29,共5页Instrument Technique and Sensor

摘　　要：针对扩散硅压力传感器温度漂移的问题,文中提出了一种基于粒子群算法优化最小二乘支持向量机(PSO-LSSVM)的温度补偿模型。通过对扩散硅压力传感器做二维标定实验,利用AD590集成温度传感器监测实验环境温度,建立PSO-LSSVM模型。最小二乘支持向量机的惩罚因子和核函数的选取会直接影响到模型的预测精度,PSO-LSSVM模型利用粒子群算法优化最小二乘支持向量机模型的惩罚因子和核函数的参数,改善了传统的最小二乘支持向量机模型对参数选取耗时耗力且未必找到全局最优解的缺陷。实验结果表明,经该模型补偿后的零点温度系数和灵敏度温度系数都减小了一个数量级,且预测值与标定值的均方误差的数量级达到10^(-6),实现了温度补偿并改善了预测精度。For the temperature drift of the silicon pressure sensor,an least squares support vector machine（ LSSVM） temperature compensation model was put forward,which was based on particle swarm optimization（ PSO）. Through the two-dimensional experiment on silicon pressure sensor,a PSO-LSSVM model was built and the integrated temperature sensor AD590 was taken advantage to monitor the environment temperature.Because the penalty factor and kernel parameter of LSSVM can affect the prediction precision directly,the optimal penalty factor and kernel parameter can be selected by PSO,which improves the defect of time consuming parameter selection.Experiment results show that after the PSO-LSSVM model,both the zero temperature coefficient and the sensitivity coefficient decrease in an order of magnitude and the mean square error of predicted values and the calibration values is up to 10^（-6）.PSO-LSSVM model realizes the temperature compensation and improves the prediction precision.

关 键 词：扩散硅压力传感器 温度补偿 粒子群算法 最小二乘支持向量机 

分 类 号：TH765[机械工程—仪器科学与技术] TP212[机械工程—精密仪器及机械]