基于自适应ZUPT与航向角误差修正的INS约束算法
INS constraint algorithm based on adaptive ZUPT with heading angle error correction
刘宇 1贺光瑞 1陈燕苹 1邹梦强 1刘小玮1
作者信息
- 1. 重庆邮电大学 自主导航与微系统重庆市重点实验室,重庆 400065;重庆邮电大学 智能传感技术与微系统重庆市高校工程研究中心,重庆 400065
- 折叠
摘要
为提高行人惯性导航系统(INS)的定位精度,提出了一种基于自适应零速修正算法(ZUPT)与航向角误差修正的INS约束算法.首先,利用基于广义似然比的零速检测阈值与检验统计量的特征值建立关系式,实现对零速检测阈值的自适应调整.然后,对INS解算位置建立线性回归模型,结合主航向与解算航向的航向差,根据行人运动类型构建观测向量和观测矩阵,对INS误差进行修正.最后,经实验验证,当行人在复杂路线变速运动时,相较于固定阈值的零速检测算法和主航向修正算法,所提算法的解算轨迹更接近真实轨迹,平均闭环误差从 3.31%D减小至 1.45%D,有效提高了INS定位精度,具有较好的工程应用价值.
Abstract
To improve the positioning accuracy of pedestrian inertial navigation system(INS),an INS constraint algorithm based on adaptive zero-velocity update algorithm(ZUPT)and heading angle error correction is proposed.Firstly,the threshold of zero-velocity detection based on the generalized likelihood ratio and the eigenvalue of the test statistic are used to establish an equation,adaptive adjusting of the threshold of zero-velocity detection.Then,a linear regression model is established based on the solution position of INS.Combining the heading difference between the main heading and the solved heading,the observation vector and observation matrix are constructed according to the pedestrian movement type to correct the INS error.Ultimately,experimental validation confirms that when pedestrians move in a complex route with variable velocity,compared with the zero-velocity detection algorithm with fixed threshold and the main course correction algorithm,the proposed algorithm is closer to the real trajectory,and the average closed-loop error is reduced from 3.31%D to 1.45%D,which effectively improves the positioning accuracy of INS and has good engineering application value.
关键词
行人导航/零速修正/广义似然比/扩展卡尔曼滤波/航向角修正Key words
pedestrian navigation/zero-velocity update/generalized likelihood ratio/extended Kalman filter/heading angle correction引用本文复制引用
基金项目
国家自然科学基金(52175531)
国家自然科学基金(62305039)
重庆市自然科学基金(CSTB2023NSCQ-MSX0568)
重庆市自然科学基金(CSTB2022NSCQ-LZX0050)
重庆市自然科学基金(cstc2022ycjhbgzxm0190)
重庆市自然科学基金(CSTB2023NSCQ-LMX0028)
出版年
2024
NETL
NSTL
维普
万方数据