曹红燕, 刘长明, 沈小林,等.基于GPS/INS的自适应无迹Kalman滤波算法[J].河南理工大学学报（自然科学版）,2021,40(6):149-156.

CAO H Y, LIU C M, SHEN X L, et al.Adaptive traceless Kalman filtering algorithm based on GPS/INS[J].Journal of Henan Polytechnic University(Natural Science) ,2021,40(6):149-156.

基于GPS/INS的自适应无迹Kalman滤波算法

曹红燕1, 刘长明1, 沈小林1, 牛兴龙1, 李大威1, 陈燕2

1.中北大学 电气与控制工程学院，山西 太原 030051;2.陆军装备部驻北京地区军事代表局某军代室，山西 太原 030051

摘要:载体的姿态参数是导航系统重要的影响因素，为提高姿态角测量精度，以INS和GPS紧组合导航系统为研究背景，针对无迹Kalman滤波算法对误差模型较敏感、新息噪声干扰数据需预处理、算法实效性低等缺点，提出改进的自适应无迹Kalman滤波算法。首先，用自适应窗口在线估计系统噪声和量测噪声的协方差值，得到与实际噪声更贴近的统计特性，减小数据预处理的干扰；其次，对状态预测方差阵引入次优渐消因子减少计算量，同时为了减少模型的精度损耗，对滤波过程引入统计量，确定模型不确定性检测阈值；最后，用扩展Kalman滤波、无迹 Kalman滤波和改进后的新滤波算法对无人机航向轨迹进行数据处理。结果分析可得，改进的滤波融合算法能将姿态的测量精度提高到0.1°，具有更强的收敛性，能较好地抑制漂移误差。

关键词:GPS;INS;Kalman滤波算法;扩展Kalman滤波算法;无迹Kalman滤波算法;姿态角

doi:10.16186/j.cnki.1673-9787.2020050108

基金项目:国家自然科学基金资助项目（61903343 ）；山西省自然科学基金资助项目（201901D111151）

收稿日期:2020/05/31

修回日期:2020/07/23

出版日期:2021/11/15

Adaptive traceless Kalman filtering algorithm based on GPS/INS

CAO Hongyan1, LIU Changming1, SHEN Xiaolin1, NIU Xinglong1, LI Dawei1, CHEN Yan2

1.School of Electrical and Control Eengineering,North University of China ,Taiyuan  030051 ,Shanxi,China;2.Military Representative Office of Military Equipment Department in Beijing,Taiyuan  030051 ,Shanxi,China

Abstract:Abstract ：The attitude information of the carrier is one of the most important information in the navigation parameters. In order to improve the measurement accuracy of the attitude angle ,taking INS and GPS tight integrated navigation system as the background frame ,in view of the traceless Kalman filtering algorithm is more sensitive to the error model ，the new information noise interference data preprocessing and the low effective ness of the algorithm ,an improved adaptive traceless Kalman filtering algorithm was proposed. The covariance difference between system noise and measurement noise was estimated online by adaptive window ,and the statistical characteristics closer to the actual noise were obtained to reduce the interference of data preprocessing； For the variance matrix of state prediction ,the suboptimal fading factor was introduced to reduce the calculation a-mount； In order to reduce the precision loss when the model was determined ,the statistics were introduced to the filtering process to determine the detection threshold of model uncertainty. Finally ,extended Kalman filtering ,traceless Kalman filtering and improved new filtering algorithm were used to process the data of UAV's heading trajectory. The results showed that the improved filtering fusion algorithm could improve the attitude measurement accuracy to 0. 1 °.

Key words:GPS;INS;Kalman filtering algorithm;extended Kalman filtering algorithm;trackless Kalman filtering algorithm;attitude angle

 基于GPS_INS的自适应无迹Kalman滤波算法_曹红燕 (2).pdf