Human Activity Recognition Using a Single Wrist IMU Sensor via Deep Learning Convolutional and Recurrent Neural Nets

E. Valarezo; P. Rivera; J. M. Park; G. Gi; T. Y. Kim; M. A. Al-Antari; M. Al-Masni; T.-S. Kim

doi:10.33150/JITDETS-1.1.1

E. Valarezo Department of Biomedical Engineering, College of Electronics and Information, Kyung Hee University, Yongin,South Korea
P. Rivera Department of Biomedical Engineering, College of Electronics and Information, Kyung Hee University, Yongin,South Korea
J. M. Park Department of Biomedical Engineering, College of Electronics and Information, Kyung Hee University, Yongin,South Korea
G. Gi Department of Biomedical Engineering, College of Electronics and Information, Kyung Hee University, Yongin,South Korea
T. Y. Kim Department of Biomedical Engineering, College of Electronics and Information, Kyung Hee University, Yongin,South Korea
M. A. Al-Antari Department of Biomedical Engineering, College of Electronics and Information, Kyung Hee University, Yongin,South Korea
M. Al-Masni Department of Biomedical Engineering, College of Electronics and Information, Kyung Hee University, Yongin,South Korea
T.-S. Kim Department of Biomedical Engineering, College of Electronics and Information, Kyung Hee University, Yongin,South Korea

DOI: https://doi.org/10.33150/JITDETS-1.1.1

Keywords: Human Activities, Inertial Measurement Units (IMUs), Convolutional Neural Nets (CNN),, Recurrent Neural Nets (RNN), HAR System

Abstract

In this paper, the authors aimed to propose novel deep learning-based HAR systems with a single wrist IMU sensor. This research used time-series activity data from only one IMU sensor at a wrist to build two deep learning algorithm-based HAR systems: one is based on Convolutional Neural Nets (CNN) and the other Recurrent Neural Nets (RNN). Our two HAR systems are evaluated by 5-fold cross-validation tests to compare the performance of both systems. Five primary daily activities, including standing, walking, running, walking downstairs, and walking upstairs, were recognized. Our results show that the CNN-based HAR system achieved an average accuracy of 95.43% and the RNN-based HAR system accuracy of 96.95%. This result presents the feasibility of HAR for some macro human activities with only a single wearable IMU device.

Human Activity Recognition Using a Single Wrist IMU Sensor via Deep Learning Convolutional and Recurrent Neural Nets

Abstract

Browse

Indexation