Deep Learning-Based Direction Estimation Scheme Using Smartphone Inertial Sensors 


Vol. 47,  No. 6, pp. 898-907, Jun.  2022
10.7840/kics.2022.47.6.898


PDF
  Abstract

Pedestrian dead reckoning (PDR) which is widely used in smartphone-based indoor localization, updates user"s location by calculating the movement variations using the inertial measurement unit (IMU) inside the smartphones. The PDR consists of three components: step detection, stride length estimation, and direction estimation. In the indoor environment, although the step detection and the stride length estimation obtain accurate results relatively, the direction estimation accuracy can be easily affected by the cumulative errors and the magnetic noises. In this paper, we propose a deep learning-based direction estimation scheme by taking into account the gyroscope drift problem and the external magnetic field effects. The proposed scheme collects the IMU data when the user freely walks with the smartphone as the training data and labels the data using the actual walking direction as the output data. The proposed deep learning model receives the IMU sensor values while the user walks, and predicts the direction of the current step. In order to evaluate the performance of the proposed scheme, we built an Android app to collect the IMU sensor data, and train the deep learning model by utilizing the Google’s TensorFlow machine learning library.

  Statistics
Cumulative Counts from November, 2022
Multiple requests among the same browser session are counted as one view. If you mouse over a chart, the values of data points will be shown.


  Cite this article

[IEEE Style]

C. Lin and Y. Shin, "Deep Learning-Based Direction Estimation Scheme Using Smartphone Inertial Sensors," The Journal of Korean Institute of Communications and Information Sciences, vol. 47, no. 6, pp. 898-907, 2022. DOI: 10.7840/kics.2022.47.6.898.

[ACM Style]

Chenxiang Lin and Yoan Shin. 2022. Deep Learning-Based Direction Estimation Scheme Using Smartphone Inertial Sensors. The Journal of Korean Institute of Communications and Information Sciences, 47, 6, (2022), 898-907. DOI: 10.7840/kics.2022.47.6.898.

[KICS Style]

Chenxiang Lin and Yoan Shin, "Deep Learning-Based Direction Estimation Scheme Using Smartphone Inertial Sensors," The Journal of Korean Institute of Communications and Information Sciences, vol. 47, no. 6, pp. 898-907, 6. 2022. (https://doi.org/10.7840/kics.2022.47.6.898)

Search
(IN TITLE, AUTHOR, ABSTRACT,KEYWORDS)

Advanced Search ▼

POPULAR KEYWORDS
(TOP 10 KEYWORDS)

Recent Publications
(LAST 3 YEARS)






pISSN : 1226 - 4717
eISSN : 2287 - 3880