基于RANSAC算法的三轴磁通门传感器误差校正  

Error Correction of Three⁃axis Fluxgate Sensor Based on RANSAC Algorithm

作　　者：罗建刚 罗骋 李海兵[1,2] 刘静晓 李海虎[1] LUO Jian⁃gang;LUO Cheng;LI Hai⁃bing;LIU Jing⁃xiao;LI Hai⁃hu(Qingdao Marine Science and Technology Center,Qingdao 266237;Beijing Institute of Aerospace Control Devices,Beijing 100039)

机构地区：[1]青岛海洋科技中心,青岛266237 [2]北京航天控制仪器研究所,北京100039

出　　处：《导航与控制》2023年第1期75-82,共8页Navigation and Control

基　　金：崂山实验室科技创新项目(编号:LSKJ202200100);科技部重点研发计划(编号:2022YFC3103401);山东省重大科技创新工程专项(编号:2018SDKJ0201)

摘　　要：针对三轴磁通门传感器存在转向误差影响其测量精度而传统误差校正算法在非理想条件下应用效果不佳的问题,在椭球拟合算法的基础上提出了基于随机抽样一致性(Random Sampling Consensus,RANSAC)算法的三轴磁通门传感器误差校正方法,该方法可以有效剔除样本数据中的噪声和畸变数据点,最终估算出更加准确的误差参数。仿真分析和外场实验表明,即使在样本数据中含有噪声或畸变数据点时,该方法仍然可以估算出更加准确的误差参数。相比传统椭球拟合算法,仿真分析中传感器误差参数的估算精度提升了1~2个数量级,外场实验中实测数据磁总场的峰峰值和均方根误差的改善率分别提高了10.2%和32.1%。Aiming at the problem that the diversionary error of three⁃axis fluxgate sensor affects its measurement ac⁃curacy,and the traditional error correction algorithm is not well applied under non⁃ideal conditions,based on the ellipsoid fitting algorithm,a three⁃axis fluxgate sensor error correction method based on random sampling consensus(RANSAC)algo⁃rithm is proposed.This method can effectively eliminate the noise and distortion data points in the sample data,and finally estimates more accurate error parameters.Simulation analysis and field experiments show that this method can still estimate more accurate error parameters even when the sample data contains noise or distorted data points.Compared with the tradi⁃tional ellipsoid fitting algorithm,the estimation accuracy of sensor error parameters in the simulation analysis is improved by 1~2 orders of magnitude,and the improvement rates of peak to peak and root mean square error of the measured data total magnetic field in the field experiment are increased by 10.2%and 32.1%respectively.

关 键 词：三轴磁通门传感器 椭球拟合 随机抽样一致性算法 误差校正 

分 类 号：P318.61[天文地球—固体地球物理学]