一种基于降维对偶四元数的多源导航系统信息融合方法  

Multi-source navigation system information fusion method based on reduced-dimensional dual quaternion

作　　者：夏琳琳[1] 肖建磊 徐迅 李鑫颖 XIA Linlin;XIAO Jianlei;XU Xun;LI Xinying(School of Automation Engineering,Northeast Electric Power University,Jilin 132012,China;Institute for Superconducting and Electronic Materials,University of Wollongong,Wollongong 2500,Australia)

机构地区：[1]东北电力大学自动化工程学院,吉林132012 [2]卧龙岗大学ISEM,澳大利亚卧龙岗2500

出　　处：《中国惯性技术学报》2019年第6期738-745,共8页Journal of Chinese Inertial Technology

基　　金：吉林省科技厅自然科学基金项目(20170101125JC);吉林市科技局杰出青年人才培养专项(20190104128)

摘　　要：以陆用载体为研究背景,提出了一种基于降维对偶四元数的多源导航系统信息融合方法。以SINS作为公共参考系统,在分别获得推力、引力及位置对偶四元数误差方程基础上,搭建SINS/GPS/地磁的降维联邦Kalman滤波器模型。一方面,结合载体匀速、加速、转弯、爬坡四种运动状态,利用奇异值分解对系统状态量进行可观测性分析;另一方面,开展该降维对偶四元数导航框架下的惯性/GPS、惯性/地磁子滤波器分级设计,以获得最佳状态维数与导航解算效率。仿真与实验结果表明,降维后的组合导航系统计算量仅为全状态的3.34%,且参数估算结果的"解的集中性"更好。Based on the research background of land vehicles, a multi-source navigation system information fusion method based on reduced-dimensional dual quaternion is proposed. Taking strapdown inertial navigation system(SINS) as the public reference system, the federated Kalman filter with dimensionality reduction for SINS/GPS/geomagnetism is constructed based on the derivation of thrust velocity, gravitational velocity and position dual quaternion error equations. On the one hand, singular value decomposition(SVD) method is used to implement the observability of the system state analysis in combination with the four motion states of the vehicles: uniform velocity, acceleration, turning and climbing. On the other hand, the hierarchical designs of INS/GPS and INS/geomagnetic subfilters in the dual quaternion navigation frame are conducted under the constructed navigation frameworks, in order to obtain the optimal system state dimensionality and navigation solution efficiency. Simulation and experimental results show that the computation amount of the integrated navigation system after dimension reduction is only 3.34% of that with full state model, which leads to a better "concentration of solutions" for the parameter estimation results.

关 键 词：降维对偶四元数 SINS/GPS/地磁导航系统 可观测性分析 联邦Kalman滤波 奇异值分解 

分 类 号：U666.1[交通运输工程—船舶及航道工程]