基于惯性导航的步行者零速检测算法
|
摘要
为解决惯性导航长时间误差漂移、步行者轨迹推算精度低的缺点,提出了一种基于人体脚部运动特征的零速检测算法。通过步态分析和极大似然估计将行人运动分为静止和运动两个阶段,只在运动阶段进行惯性导航确定步长,为减小误差随时间的漂移,静止阶段进行传感器误差修正,将步长信息连续推算,即可实现惯性导航定位。基于提出的零速检测算法模型,搭建以惯性测量单元为核心的实验平台,进行了3组实验。实验结果表明:基于惯性导航的零速检测算法检测零速区间的误差为0,单步步长误差为2. 7%,结合零速修正连续推算误差为1. 6%,验证了该方法的有效性。
Aiming at problems that inertial navigation error is drifting with time and the precision of pedestrian trajectory estimation is low,a zero velocity detection algorithm based on human foot movement characteristics is proposed. Through gait analysis and the maximum likelihood estimation,pedestrian motion is divided into two stages: static and movement. The step size is determined by inertial navigation only in the movement stage. In order to reduce the drift of the error with time,the sensor error is corrected in the static stage,and the step size information is deduced continuously,so that the*inertial navigation and positioning can be realized. Based on the proposed zero velocity detection algorithm model,construct experimental platform which uses inertial measurement element as core,and three groups of experiments are carried out. Experimental result shows that the detecting error of zero velocity detection algorithm based on inertial navigation is zero,step length error of single step is 2. 7 %,combined with zero velocity update( ZUPT),continous dead reckoning error is 1. 6 %,which verify the method is effective.
Aiming at problems that inertial navigation error is drifting with time and the precision of pedestrian trajectory estimation is low,a zero velocity detection algorithm based on human foot movement characteristics is proposed. Through gait analysis and the maximum likelihood estimation,pedestrian motion is divided into two stages: static and movement. The step size is determined by inertial navigation only in the movement stage. In order to reduce the drift of the error with time,the sensor error is corrected in the static stage,and the step size information is deduced continuously,so that the*inertial navigation and positioning can be realized. Based on the proposed zero velocity detection algorithm model,construct experimental platform which uses inertial measurement element as core,and three groups of experiments are carried out. Experimental result shows that the detecting error of zero velocity detection algorithm based on inertial navigation is zero,step length error of single step is 2. 7 %,combined with zero velocity update( ZUPT),continous dead reckoning error is 1. 6 %,which verify the method is effective.
引文
[1]孙璧文,樊启高,武亚恒,等.基于PDR/UWB紧耦合的足绑式行人导航技术[J].传感器与微系统,2017,36(3):43-47,50.
[2]王亚娜,蔡成林,李思民,等.基于行人航迹推算的室内定位算法研究[J].电子技术应用,2017,43(4):86-89,93.
[3]王云涛,钱伟行,徐昊,等.自适应零速修正辅助的微惯性定位系统研究[J].南京师范大学学报:工程技术版,2017,17(4):14-19.
[4]张新喜,张嵘,郭美凤,等.足绑式行人导航偏航角误差自观测算法(英文)[J].中国惯性技术学报,2015,23(4):457-466.
[5]吴斌,刘丽明,薛婷,等.基于惯性传感器的上肢位置跟踪[J].测控技术,2018,37(6):64-67,71.
[6]张苗,许建新,黄欣,等.基于人体运动学辅助的可穿戴式行人导航系统[J].导航与控制,2018,17(4):1-6,13.
[7]陈国良,杨洲.基于加速度量测幅值零速检测的计步算法研究[J].武汉大学学报:信息科学版,2017,42(6):726-730,788.
[8]费程羽,苏中,李擎.行人惯性导航零速检测算法[J].传感器与微系统,2016,35(3):147-150,153.
[9]陈国通,王小娜,张晓旭,等.基于惯性导航的室内定位误差修正算法[J].河北工业科技,2018,35(3):185-190.
[10]樊启高,孙艳,孙璧文,等.基于GLRT零速检测算法的行人室内定位系统[J].传感技术学报,2017,30(11):1706-1711.
[11]陈克川,李保国,刘思庆,等.足绑式PNS自适应滑动平均零速区间检测方法[J].导航与控制,2018,17(2):52,95-100.
[2]王亚娜,蔡成林,李思民,等.基于行人航迹推算的室内定位算法研究[J].电子技术应用,2017,43(4):86-89,93.
[3]王云涛,钱伟行,徐昊,等.自适应零速修正辅助的微惯性定位系统研究[J].南京师范大学学报:工程技术版,2017,17(4):14-19.
[4]张新喜,张嵘,郭美凤,等.足绑式行人导航偏航角误差自观测算法(英文)[J].中国惯性技术学报,2015,23(4):457-466.
[5]吴斌,刘丽明,薛婷,等.基于惯性传感器的上肢位置跟踪[J].测控技术,2018,37(6):64-67,71.
[6]张苗,许建新,黄欣,等.基于人体运动学辅助的可穿戴式行人导航系统[J].导航与控制,2018,17(4):1-6,13.
[7]陈国良,杨洲.基于加速度量测幅值零速检测的计步算法研究[J].武汉大学学报:信息科学版,2017,42(6):726-730,788.
[8]费程羽,苏中,李擎.行人惯性导航零速检测算法[J].传感器与微系统,2016,35(3):147-150,153.
[9]陈国通,王小娜,张晓旭,等.基于惯性导航的室内定位误差修正算法[J].河北工业科技,2018,35(3):185-190.
[10]樊启高,孙艳,孙璧文,等.基于GLRT零速检测算法的行人室内定位系统[J].传感技术学报,2017,30(11):1706-1711.
[11]陈克川,李保国,刘思庆,等.足绑式PNS自适应滑动平均零速区间检测方法[J].导航与控制,2018,17(2):52,95-100.