惯导系统辅助CSS的室内定位方法

摘要

A high-precision indoor positioning method based on inertial navigation system(INS)assisted Chirp Spread Spectrum(CSS)is proposed. Firstly,an inertial measurement unit(IMU)based on MPU9250 multi-axis sensor isdesigned,using the Digital Motion Processing(DMP)database to obtain accurate and stable heading angle by quaternion calculation. Becauseof the magne tometer with MPU9250,the cumulative error of the heading angleis avoided. Then,based on the analysis of the NLOS(Non Line of Sight)positioning error of CSS,a method of selecting star based on triangular criterion is proposed,which effectively reduces the influence of NLOS errors in indoor com-plex environment. Finally,the extended Kalman filter(EKF)is used to fuse the inertial navigation system and the CSS localization,inorder to output stable and accurate localization results. Through the analysis and comparison of the experiment in 52 m×20 m underground garage,the results show that the mean error of inertial navigation and CSS separate positioning are 0.3456 m and 0.3659 m,and the mean error of the combined navigation method in this paper is 0.1581 m,and the positioning performance is higher more than 50% compared to the previous two single positioning method. The proposed method reduces the costof positioning and the complexity of the system,and im-proves the accuracy of positioning.%提出一种惯性导航系统INS(Inertial Navigation System)辅助线性调频扩频CSS(Chirp Spread Spectrum)的高精度室内定位方法.首先设计了基于MPU9250多轴传感器的惯性测量单元IMU(Inertial Measurement Unit),利用数字运动处理(DMP)数据库经四元数解算可求得准确稳定的航向角,由于MPU9250自带的磁力计,规避了航向角的累积误差问题.随后在分析CSS非视距定位误差基础上,提出了基于三角形三边准则的选星方法,有效降低了室内复杂环境下非视距的影响;最后利用扩展卡尔曼滤波对惯导系统与CSS定位进行融合,以输出稳定、准确的定位结果.通过在50 m×20 m的地下车库实验分析比较,结果表明:惯导与CSS单独定位的平均误差分别为0.3456 m、0.3659 m,本文组合导航方法平均误差为0.158 l m,较前两种单一的定位方式性能提高了50%以上,该方法降低了定位的成本,系统实现复杂度降低,定位精度提高.