FA优化BP神经网络的MEMS陀螺仪温度漂移补偿  

Temperature Drift Compensation of MEMS Gyroscope Based on FA-Optimized BP Neural Network

作　　者：郭震 刘颖 于福华 Guo Zhen;Liu Ying;Yu Fuhua(School of Automation,Xi'an University of Posts&Telecom m unications,Xi'an 710121,China)

机构地区：[1]西安邮电大学自动化学院

出　　处：《微纳电子技术》2019年第10期817-821,827,共6页Micronanoelectronic Technology

基　　金：国家青年基金资助项目(51405387);陕西省青年基金资助项目(2016JQ5051);陕西省教育厅科学研究计划立项资助项目(12JK0791)

摘　　要：微电子机械系统(MEMS)陀螺仪输出易受环境温度的影响,产生温度漂移,测量精度降低,为解决这个问题,提出一种萤火虫算法(FA)优化BP神经网络的温度漂移补偿方法,在传统的BP神经网络中,存在易陷于局部极值的问题可能降低建模精度甚至导致建模失败,而此方法可以避免这个问题。首先在全温区(-40℃^+70℃)选取7个温度点进行测试,接下来采用该方法建立MEMS陀螺仪温度漂移模型并进行实际验证,验证结果表明该方法可以明显降低MEMS陀螺仪温度漂移,且相比于传统BP神经网络,其补偿效果也有较大幅度提升。The output of micro-electromechanical system(MEMS)gyroscope can be influenced by ambient temperature,resulting in the temperature drift which reduces its measurement accuracy.In order to solve above problems,based on firefly algorithm(FA)optimization BP neural network,a new method of temperature drift compensation was proposed.For traditional BP neural network,the problem of being trapped in local extremum may lower the modeling accuracy and even lead to modeling failure,and the method can avoid this problem.Seven temperature points for testing were firstly selected from the full temperature range(-40 ^+70 ℃).The MEMS gyroscope temperature drift model was established by the method and was verified.The verification results show that this method can significantly reduce the temperature drift of the MEMS gyroscope.Compared by traditional BP neural network,its compensation effect also has greatly improvement.

关 键 词：微电子机械系统(MEMS) 陀螺仪 萤火虫算法(FA) BP神经网络 温度漂移补偿 

分 类 号：V241.5[航空宇航科学与技术—飞行器设计] TH703[机械工程—仪器科学与技术]