Abstract—Several physical characteristics of workers can be extracted from physical timeline data to understand acceleration and angular velocity. Although various approaches have been implemented globally for indoor and outdoor agricultural (agri-) working sites, there is room for improvement. In this study, we aim to adapt these approaches particularly for real agri-directors, leaders and managers to improve the quality of tasks and their security levels. Thus, we apply a deep learning-based method and qualitatively demonstrate the classification of physical timeline datasets. To create our dataset, our subjects were six experienced agri-manual workers and six completely inexperienced men. The targeted task was cultivating the semi-crunching position using a simple, Japanese-style hoe. We captured the subjects’ acceleration and angular velocity data from an integrated multi-sensor module mounted on a wood lilt 15 cm from the gripping position of the dominant hand. We used Python code and recent distributed libraries for computation. For data classification, we successively executed a Recurrent Neural Network (RNN), which we evaluated using wavelet analyses such as the Fast Fourier Transform (FFT). These methods of analyzing digital data could be of practical use for providing key suggestions to improve daily tasks.

Index Terms—Deep learning-based classification, characteristic extraction, physical timeline data, acceleration, angular velocity.

Shinji Kawakura is with The University of Tokyo/The Research Center for Advanced Science and Technology (RCAST), Bunkyo-ku, Tokyo, 153-8904, Japan (email: s.kawakura@gmail.com, kawakura@motolabo.net).
Ryosuke Shibasaki is with The University of Tokyo/Center for Spatial Information Science, Kashiwa-shi, Chiba, 277-8568, Japan.

Cite: Shinji Kawakura, Ryosuke Shibasaki, "Deep Learning Based System to Extract Agricultural Workers’ Physical Timeline Data for Acceleration and Angular Velocity," International Journal of Bioscience, Biochemistry and Bioinformatics vol. 10, no. 2, pp. 84-93, 2020.

Copyright © 2020 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).