基于AGA-LSSVM的电容式位移传感器温度补偿  

Temperature Compensation for Capacitive Displacement Sensor Based on AGA-LSSVM

作　　者：辜申 卢文科[1] 冯阳 左锋[1] GU Shen;LU Wenke;FENG Yang;ZUO Feng(College of Information Science and Technology,Donghua University,Shanghai 201620,China)

机构地区：[1]东华大学信息科学与技术学院,上海201620

出　　处：《东华大学学报（自然科学版）》2020年第5期798-804,共7页Journal of Donghua University(Natural Science)

基　　金：国家自然科学基金资助项目(61274078);中国纺织工业联合会“纺织之光”应用基础研究资助项目(J201608)。

摘　　要：利用自适应遗传算法(adaptive genetic algorithm,AGA)良好的全局寻优能力来优化最小二乘支持向量机(least square support vector machine,LSSVM)的正则化参数γ和核函数参数σ2,构建电容式位移传感器的AGA-LSSVM温度补偿模型,并对其输出位移进行预测,从而实现温度补偿。经计算分析可知,利用AGA-LSSVM算法模型对电容式位移传感器进行温度补偿后,其灵敏度温度系数αs由6.83×10-3℃-1变为2.98×10-4℃-1,相对误差δt由25.58%减小到0.45%。同时将AGA-LSSVM算法模型与LSSVM、遗传算法优化最小二乘支持向量机(GA-LSSVM)以及自适应遗传算法优化支持向量机(AGA-SVM)算法模型的预测结果进行比较,结果表明,利用AGA-LSSVM算法模型对电容式位移传感器进行温度补偿显示出较大的优越性。The adaptive genetic algorithm(AGA)was used to optimize the regularization parameterγand kernel function parameterσ2 of the least square support vector machine(LSSVM),then the AGA-LSSVM temperature compensation algorithm model of capacitive displacement sensor was constructed,and its output displacement was predicted to realize temperature compensation.The calculation and analysis results show that,the temperature sensitivity coefficientαs decreased from 6.83×10-3℃-1 to 2.98×10-4℃-1,and the relative errorδt decreased from 25.58%to 0.45%after temperature compensation of capacitive displacement sensor by AGA-LSSM algorithm model.Meanwhile,the AGA-LSSVM algorithm model was compared with the LSSVM,the genetic algorithm-least square support vector machine(GA-LSSVM)and the adaptive genetic algorithm-support vector machine(AGA-SVM)algorithm model.It indicates that the AGA-LSSVM algorithm model has great advantages for temperature compensation of the capacitive displacement sensor.

关 键 词：电容式位移传感器 温度补偿 灵敏度温度系数 自适应遗传算法 最小二乘支持向量机算法 

分 类 号：TP212.6[自动化与计算机技术—检测技术与自动化装置]