摘要
针对硅压阻式传感器灵敏度和零点温度漂移大、硬件补偿电路效果不佳的问题,提出最小二乘支持向量机方法对其温度漂移进行补偿。首先分析了经硬件补偿后的硅压阻式传感器的温度漂移特性,在整个检测范围内选取均匀分布的温度、压力数据作为模型输入,经预处理后对输出数值进行训练,并运用网格搜索法和交叉确认法优化模型的惩罚因子和正则化参数,建立了传感器温度补偿模型。实验结果表明,基于最小二乘支持向量机的温度补偿算法在0-100℃温度范围内把传感器输出综合精度从3.2%FS提高到0.25%FS,进一步提高了传感器的精度和温度使用范围,具有较高的实用价值。
Aiming at the problems of the large sensitivity and zero temperature drift of the silicon piezoresistive sen-sor using hardware compensation,a new method based on the least square-support vector（LS-SVM）is adopted toconstruct the temperature compensation model. In this paper,the temperature drift property of the silicon piezore-sistive sensors which have been compensated by hardware are analyzed,and data with the uniform distribution ofthe whole temperature and pressure range is selected to be as the model input and is preprocessed to train the out-put data. The kernel and regularization parameter of the model is optimized by using the grid searching method andthe cross-validation method,and a temperature compensation model of the sensor is established. Experiment resultsshow the comprehensive accuracy is improved from 3.2% up to 0.25% in the temperature range of 0-100 ℃ by tem-perature compensation model based on LS-SVM,this method further improves the accuracy and temperature appli-cation range and can be put to good use.
出处
《传感技术学报》
CAS
CSCD
北大核心
2016年第4期500-505,共6页
Chinese Journal of Sensors and Actuators
基金
浙江省科技厅公益技术研究社发项目(2014C33114)
关键词
硅压阻式传感器
温度漂移
温度补偿
最小二乘支持向量机
silicon piezoresistive sensor
temperature drift
temperature compensation
least square-support vector machine（LS-SVM）
