基于递阶遗传RBF网络的MEMS陀螺温度补偿  

Temperature Compensation for MEMS Gyroscope Based on Hierarchical Genetic Algorithm and RBF Neural Network

作　　者：陈殿生[1] 邵志浩[1] 雷旭升[1] 王田苗[1] 

机构地区：[1]北京航空航天大学,北京100191

出　　处：《中国机械工程》2009年第17期2063-2066,共4页China Mechanical Engineering

基　　金：国家863高技术研究发展计划资助项目(2006AA04Z206);国际科技合作资助项目(2008DFR70100)

摘　　要：针对温度变化对MEMS陀螺的零位偏移和标度因数影响较大的问题,提出了基于径向基函数(RBF)神经网络的温度补偿方案,采用二级递阶遗传算法(HGA)对RBF网络的网络拓扑结构和网络参数进行了优化。基于陀螺当前状态信息对MEMS陀螺输出信号进行补偿,可提高测量精度。实验证明,该方法可以有效减小温度对陀螺输出的影响,在温度变化的情况下,角速度误差仍控制在±0.5°/s以内,大大提高了MEMS陀螺的输出精度。The bias and scale factor of MEMS gyroscope are often influenced by temperature. To get the high precision attitude information, this paper proposed a temperature compensation scheme based on RBFNN. A new hierarchical genetic algorithm was also used to optimize the network construction and parameters. The effectiveness of the proposed scheme was shown by a series of tests. The error of the angular speed is controlled within ±0.5°/s under real conditions.

关 键 词：MEMS陀螺 RBF网络 遗传算法 温度补偿 误差模型 

分 类 号：V241.59[航空宇航科学与技术—飞行器设计]