摘要
在利用磁惯性传感器组合解算姿态角的过程中,磁干扰对俯仰角和横滚角的估计精度产生显著影响。为此,提出一种基于扩展卡尔曼滤波的磁干扰解耦姿态估计算法(DMI-EKF)。首先,在乘性扩展卡尔曼滤波框架内,将四元数分解为估计四元数和误差四元数,并将误差四元数作为状态量优化。接着,通过深入分析磁力计干扰信息对俯仰角和横滚角的影响机制,利用惯性传感器数据计算出的四元数提取水平磁矢量,结合加速度信息构建观测方程,实现磁干扰分离。动态实验表明:与传统EKF算法相比,所提算法的俯仰角和横滚角解算精度分别提升44.5%和50.2%,显著增强复杂环境下姿态估计的可靠性与准确性。
In the process of estimating attitude angles using magnetic inertial sensor fusion,magnetic interference significantly affects the estimation accuracy of pitch and roll angles.Therefore,decoupling magnetic interference in attitude estimation algorithm with extended Kalman filter(DMI-EKF)is proposed.Firstly,within the multiplicative extended Kalman filter framework,the quaternion is decomposed into estimated quaternion and error quaternion,and the error quaternion is optimized as a state variable.Then,by deeply analyzing the influence mechanism of magnetometer interference information on the pitch angle and roll angle,the horizontal magnetic vector is extracted from inertial sensor-derived quaternion.The observation equation is formulated by integrating acceleration data with the horizontal magnetic vector,enabling effective magnetic interference decoupling.Dynamic experiments show that compared with the traditional EKF algorithm,the calculation accuracy of the pitch angle and roll angle of the proposed algorithm is improved by 44.5%and 50.2%respectively,significantly enhancing the reliability and precision of attitude estimation in complex environments.
作者
乔美英
杜衡
韩昊天
邱运强
QIAO Meiying;Du Heng;Han Haotian;Qiu Yunqiang(College of Electrical Engineering and Automation,Henan Polytechnic University,Jiaozuo 454000,China;International Joint Laboratory for Intelligent Equipment Direct Drive Technology and Control,Jiaozuo 454000,China)
出处
《中国惯性技术学报》
北大核心
2025年第7期680-687,697,共9页
Journal of Chinese Inertial Technology
基金
国家自然科学基金(U1404510)
河南省科技攻关项目(222102220076)
河南省自然科学基金资助项目(No.232300421152)。
关键词
四元数
扩展卡尔曼滤波
姿态估计
磁惯性传感器
quaternion
extended Kalman filter
attitude estimation
magnetic inertial sensor
