ISSN: 1673-825X    Imprint: Chongqing University of Posts and Telecommunications Journal
Volume listings
  • Current volume
  •  
  • Search by keywords:
  • Search by 
  • Keywords
  • To
  • Journal archive
Related Links
基于智能手机传感器的林区行人定位算法
Smartphone sensor-based algorithm for forest-pedestrian location
DOI:10.3979/j.issn.1673-825X.2019.04.005
Received:August 06, 2018  Revised:June 20, 2019
中文关键词:林区行人定位  智能手机  传感器  PDR算法  差分气压测高
英文关键词:forest pedestrian location  smartphone  sensors  PDR algorithm  differential barometric altimetry
基金项目:中央高校基本科研业务费专项基金(TD2014-02);地质矿产调查评价项目(1212011120436)
Author NameAffiliationE-mail
WANG Hanqing School of Information Science & Technology, Beijing Forestry University, Beijing 100083, P. R. China eric_wanghanqing@outlook.com 
WU Gang School of Information Science & Technology, Beijing Forestry University, Beijing 100083, P. R. China wugang@bjfu.edu.cn 
CHEN Yuelu School of Information Science & Technology, Beijing Forestry University, Beijing 100083, P. R. China cyl_226@163.com 
CHEN Feixiang School of Information Science & Technology, Beijing Forestry University, Beijing 100083, P. R. China fxchen@126.com 
Hits: 37
Download times: 13
中文摘要:
      针对林区卫星信号缺失、跟踪定位困难的问题,提出了基于智能手机传感器的林区行人定位算法(forest-pedestrian location,FPL)。算法在行人航位推算算法(pedestrian dead reckoning,PDR)基础上进行改进:采用扩展卡尔曼滤波(extended Kalman filter, EKF)与卡尔曼滤波(Kalman filter, KF)融合算法对磁力计、加速度计及陀螺仪输出进行多次融合,以提高方位角测量精度;随后,使用Savitzky-Golay(S-G)滤波处理方位角测量值,以提高PDR算法中方位角的估计精度;引入K邻近(K-nearest neighbor, KNN)算法估计步长,将拟合显式步长函数问题转化为“懒惰学习”问题;使用差分气压测高法求解行人高程信息,从而获取行人在林区内的3维定位信息。实验结果表明,该算法可以提高方位角及步长的估计精度,同时可以增加精准的高程定位信息,整体误差控制在5%以内,可以满足林区无信号条件下的定位需求。
英文摘要:
      Aiming at the problems of the lack of satellite signals and the difficulty in locating in the forest area, a smartphone sensor-based algorithm for forest pedestrian location (FPL) was proposed in this paper. The algorithm was improved based on the pedestrian dead reckoning (PDR) algorithm. First, the extended Kalman filter (EKF) and Kalman filter (KF) algorithms was used to fuse the outputs of magnetometer, accelerometer and gyroscope multiple times to improve the azimuth measurement accuracy. Next, Savitzky-Golay (S-G) filtering is used to process azimuth measurements to improve the accuracy of azimuth estimation in PDR algorithm. Then the K-nearest neighbor (KNN) algorithm was introduced to estimate the step length, then the problem of fitting the explicit step function was divided into a “lazy learning” problem. Finally, the differential barometric altimetry method was used to calculate the pedestrian elevation, so as to obtain the three dimensional location information. The result showed that the algorithm could improve the accuracy of azimuth and step length, and could increase the accurate elevation information. The overall error was controlled within 5%, which could meet the requirement of locating in the forest area without signal information.
HTML    View PDF   View/Add Comment  Download reader