Volume XLII-4/W18
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-4/W18, 905–912, 2019
https://doi.org/10.5194/isprs-archives-XLII-4-W18-905-2019
© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-4/W18, 905–912, 2019
https://doi.org/10.5194/isprs-archives-XLII-4-W18-905-2019
© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

  19 Oct 2019

19 Oct 2019

PEDESTRIAN DEAD RECKONING USING SMARTPHONES SENSORS: AN EFFICIENT INDOOR POSITIONING SYSTEM IN COMPLEX BUILDINGS OF SMART CITIES

E. Saadatzadeh1, A. Chehreghan2, and R. Ali Abbaspour1 E. Saadatzadeh et al.
  • 1School of Surveying and Geospatial Engineering, College of Eng., University of Tehran, Tehran, Iran
  • 2Faculty of Mining Engineering, Sahand University of Technology, Tabriz, Iran

Keywords: Indoor Positioning, Smartphone Mode Detection, Pedestrian Dead Reckoning, Smartphone Sensors, Filter

Abstract. This paper proposes an indoor positioning method using Pedestrian Dead Reckoning (PDR) based on the detection of the mode of the user’s smartphone. In the first step, to determine the mode of carrying the smartphone (Holding, Calling, Swinging) by suitably formed feature vectors based on sensor data, three classification algorithms (Decision Tree (DT), Support Vector Machine (SVM), and K-Nearest Neighbor (KNN)) are evaluated. From the classification algorithm perspective, the decision tree algorithm had the best performance in terms of processing time and classification. Secondly, to determine the user position, the step detection is performed by defining the upper threshold and time threshold for Acceleration norm values. The orientation component is obtained by combining accelerometer, magnetometer, and gyroscope data using Complementary Filtering and Principal Component Analysis based on Global Acceleration (PCA-GA) methods. The mean standard deviation along the direct path for the three modes of carrying (Holding, Calling, and Swinging) were obtained 6.22, 6.82, and 14.68 degrees, respectively. Localization experiments were performed on 3 modes of carrying a smartphone in a rectangular geometry path. The mean final error of positioning from ordinary walking for the three modes of holding (Calling, Holding, Swinging) were obtained 2.11, 2.34, and 4.5 m, respectively.