吉林大学学报(工学版) ›› 2023, Vol. 53 ›› Issue (8): 2364-2370.doi: 10.13229/j.cnki.jdxbgxb.20220568

• 计算机科学与技术 • 上一篇    

基于TrAdaBoost算法为内核的行人航迹推算技术

王玫(),宋志远   

  1. 桂林理工大学 信息科学与工程学院,广西 桂林 541004
  • 收稿日期:2022-05-12 出版日期:2023-08-01 发布日期:2023-08-21
  • 作者简介:王玫(1963-),女,教授,博士.研究方向:音视频信息感知与处理.E-mail:wangmei231313@163.com
  • 基金资助:
    国家自然科学基金项目(62071135);广西科技重大专项项目(创新驱动发展专项)(桂科AB17292058);广西科技计划项目(桂科AD18281044);广西重点研发计划项目(桂科AB17292058)

Pedestrian dead reckoning technology based on TrAdaBoost algorithm

Mei WANG(),Zhi-yuan SONG   

  1. School of Information Science and Engineering,Guilin University of Technology,Guilin 541004,China
  • Received:2022-05-12 Online:2023-08-01 Published:2023-08-21

摘要:

行人航迹推算(PDR)由于误差发散速度快,导致PDR定位可信时间短、累计误差大,因此,提出了基于TrAdaBoost算法为内核的行人航迹推算技术。该方法通过全球定位系统(GPS)离线收集行人室外运动信息,利用迁移学习中的TrAdaBoost算法筛选出最合适的行人运动特征,将其迁移到室内PDR定位中,纠正行人步长,实现行人航迹推算。实验结果表明:行人步长大致分布在65~75 cm内,而且经过纠正后的定位轨迹与真实定位轨迹吻合度高,不会因为惯性传感器的积累误差而出现严重偏离,平均水平误差与纯PDR定位算法相比大幅度下降,位置精度在2 m以内的概率达到了80%。因此,该方法降低了PDR发散的速度,延长了PDR定位的可信时间,提高了短距离内的定位结果的可信度。

关键词: 迁移学习, PDR定位, TrAdaBoost算法, 步长纠正

Abstract:

Due to the fast divergence of error in pedestrian track estimation, the PDR positioning trust time is short and the accumulated error is large. A pedestrian track estimation technology based on TrAdaBoost algorithm is proposed. In this method, outdoor pedestrian movement information is collected offline by GPS, and the TrAdaBoost algorithm in transfer learning is used to screen out the most suitable pedestrian movement features, which are transferred to indoor PDR location to correct pedestrian steps and realize pedestrian track calculation. The experimental results show that the pedestrians step roughly distribution within the scope of the 65—75 cm, and after correcting the positioning of the trajectory match degree is high, and real location will not be because of the accumulation of inertial sensor error and appear serious deviation, the average error compared with pure PDR localization algorithm has fallen dramatically, location accuracy is within 2 m of probability of 80%. Therefore, this method reduces the divergence speed of PDR, prolonging the credibility time of PDR location, and improving the credibility of the location results within a short distance.

Key words: migration learning, PDR localization, TrAdaBoost algorithm, step size correction

中图分类号: 

  • P228.1

图1

PDR定位原理图"

表1

GPS与IMU比较"

项目优点缺点
GPS定位精度高,数据相对准确响应时间慢
IMU计算速度快精度较低,运动时不稳定

图2

室外数据提升PDR精度"

图3

算法流程"

图4

室内外步长"

图5

迁移后的步长"

图6

步长纠正"

图7

纠正后的PDR流程图"

图8

PDR原始数据"

图9

PDR步长区间"

表2

平均水平误差统计"

定位方式平均水平误差/m
纯PDR3.4
步长区间为0.4~0.8 m2.1
步长区间为0.6~0.7 m1.2

图10

累计误差分布图"

1 高伟, 侯聪毅, 许万旸, 等. 室内导航定位技术研究进展与展望[J]. 导航定位学报, 2019, 7(1): 10-17.
Gao Wei, Hou Cong-yi, Xu Wan-yang, et al. Research progress and prospect of indoor navigation and positioning technology[J]. Journal of Navigation and Positioning, 2019, 7(1): 10-17.
2 刘富, 权美静, 王柯. 仿蝎子振源定位机理的位置指纹室内定位方法[J] .吉林大学学报: 工学版, 2019, 49(6):2076-2082.
Liu Fu, Quan Mei-jing, Wang Ke. A position fingerprint indoor localization method based on the mechanism of simulating scorpion vibration source localization[J] Journal of Jilin University(Engineering and Technology Edition), 2019, 49(6): 2076-2082.
3 李楠, 陈家斌, 袁燕. 基于WiFi/PDR的室内行人组合定位算法[J]. 中国惯性技术学报, 2017, 25(4): 483-487.
Li Nan, Chen Jia-bin, Yuan Yan. Indoor pedestrianintegrated localization strategy based on WiFi/PDR[J]. Journal of Chinese Inertial Technology, 2017, 25(4): 483-487.
4 金彦亮, 张晓帅, 齐崎, 等. 基于WiFi辅助的自适应步长的室内定位算法[J]. 电子测量技术, 2017, 40(12): 165-170.
Jin Yan-liang, Zhang Xiao-shuai, Qi Qi, et al. Indoor positioning algorithm based on adaptive step size and WiFi assisted[J]. Electronic Measurement Technology, 2017, 40(12): 165-170.
5 刘庆, 关维国, 李顺康, 等. 基于扩展Kalman滤波的室内WiFi-PDR融合定位算法[J]. 计算机工程, 2019, 45(4): 66-71, 77.
Liu Qing, Guan Wei-guo, Li Shun-kang, et al. Indoor WiFi-PDR fusion location algorithm based on extended Kalman filter[J]. Computer Engineering, 2019, 45(4): 66-71, 77.
6 郭倩倩, 崔丽珍, 杨勇, 等. 基于LSTM个性化步长估计的井下人员精准定位PDR算法[J]. 工矿自动化, 2022, 48(1): 33-39.
Guo Qian-qian, Cui Li-zhen, Yang Yong, et al. PDR algorithm for precise positioning of underground personnel based on LSTM personalized step size estimation[J]. Industry and Mine Automation, 2022, 48(1): 33-39.
7 宋震, 李俊良, 刘贵强. 基于深度学习和限幅模糊的变转速液压动力源恒流量预测方法[J]. 吉林大学学报: 工学版, 2021, 51(3): 1106-1110.
Song Zhen, Li Jun-liang, Liu Gui-qiang. A constant flow prediction method for variable speed hydraulic power source based on deep learning and limiting fuzzy control[J]. Journal of Jilin University(Engineering and Technology Edition), 2021, 51(3): 1106-1110.
[1] 王连明,吴鑫. 基于姿态估计的物体3D运动参数测量方法[J]. 吉林大学学报(工学版), 2023, 53(7): 2099-2108.
[2] 刘近贞,高国辉,熊慧. 用于脑组织分割的多尺度注意网络[J]. 吉林大学学报(工学版), 2023, 53(2): 576-583.
[3] 王生生,姜林延,杨永波. 基于最优传输特征选择的医学图像分割迁移学习[J]. 吉林大学学报(工学版), 2022, 52(7): 1626-1638.
[4] 王学智,李清亮,李文辉. 融合迁移学习的土壤湿度预测时空模型[J]. 吉林大学学报(工学版), 2022, 52(3): 675-683.
[5] 徐艳蕾,何润,翟钰婷,赵宾,李陈孝. 基于轻量卷积网络的田间自然环境杂草识别方法[J]. 吉林大学学报(工学版), 2021, 51(6): 2304-2312.
[6] 张根保,李浩,冉琰,李裘进. 一种用于轴承故障诊断的迁移学习模型[J]. 吉林大学学报(工学版), 2020, 50(5): 1617-1626.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!