AHRS姿态估计的迭代扩展卡尔曼滤波方法  

An Iterative Extended Kalman Filter Method for AHRS Orientation Estimation

作　　者：史宜巧 赵辉[2] SHI Yi-qiao;ZHAO Hui(School of Intelligent Manufacturing,Jiangsu Vocational College of Electronics and Information Technology,Jiangsu Huaian 223003,China;School of Information and Electrical Engineering,Hebei Engineering University,Hebei Handan 056038,China)

机构地区：[1]江苏电子信息职业学院智能制造学院,江苏淮安223003 [2]河北工程大学信息与电气工程学院,河北邯郸056038

出　　处：《机械设计与制造》2022年第4期270-274,共5页Machinery Design & Manufacture

基　　金：江苏省“333工程”科研资助项目(BRA2019304)。

摘　　要：针对现有滤波方法在低成本航姿参考系统(AHRS)姿态估计应用中存在准确性不足的问题,本文提出一种非线性滤波求解的姿态估计方法。根据四元数姿态表示原理与传感器测量输出模型构建了基于AHRS系统直接形式姿态估计的非线性状态空间模型,采用迭代扩展卡尔曼滤波方法进行滤波求解,实现了对姿态四元数与传感器偏差的实时估计。通过MPU9150型MEMS惯性测量单元的实测数据与ABB工业机器人同步输出的参考姿态对本文算法进行了验证,并与现有基于非线性滤波与互补滤波的姿态估计方法以及商用姿态测量单元的结果进行了对比。结果表明,本文算法有效提高了姿态估计准确性与估计精度,具有较好的性能。To improve the estimate performance for orientation estimation in attitude and heading reference system(AHRS),we proposed a nonlinear filter based approach in it.The state space model based on direct form of orientation estimation for AHRS is derived using the principle of quaternion representation of orientation and sensor output models,the derived model is solved by using iterated extended Kalman filter,which realizes the real-time estimation of quaternion represented orientation and sensor bias.The proposed method is tested by a real dataset containing raw IMU data from MPU9150 and simultaneous reference orientation quaternion output from ABB robot,and is also compared with several filter-based algorithm and complementary filter and commercial orientation sensor.The results suggest that our method is favorable in terms of accuracy and precision in the estimation of orientation.

关 键 词：航姿参考系统(AHRS) 姿态估计 四元数 非线性滤波 

分 类 号：TH16[机械工程—机械制造及自动化] V448.22[航空宇航科学与技术—飞行器设计]