Publication: Smartwatch-based sitting detection with human activity recognition for office workers syndrome
Issued Date
2018-06-08
Resource Type
Other identifier(s)
2-s2.0-85049951304
Rights
Mahidol University
Rights Holder(s)
SCOPUS
Bibliographic Citation
1st International ECTI Northern Section Conference on Electrical, Electronics, Computer and Telecommunications Engineering, ECTI-NCON 2018. (2018), 160-164
Suggested Citation
Sakorn Mekruksavanich, Narit Hnoohom, Anuchit Jitpattanakul Smartwatch-based sitting detection with human activity recognition for office workers syndrome. 1st International ECTI Northern Section Conference on Electrical, Electronics, Computer and Telecommunications Engineering, ECTI-NCON 2018. (2018), 160-164. doi:10.1109/ECTI-NCON.2018.8378302 Retrieved from: https://repository.li.mahidol.ac.th/handle/20.500.14594/45631
Research Projects
Organizational Units
Authors
Journal Issue
Thesis
Title
Smartwatch-based sitting detection with human activity recognition for office workers syndrome
Other Contributor(s)
Abstract
© 2018 IEEE. The main cause of Office Workers Syndrome (OWS) is habitual unhealthy behavior. One such example is the habit of siting for long periods at a computer. It has recently become simple to study this kind of human activity by using smartwatch sensors, and as a result, HAR (human activity recognition) has become a field attracting considerable interest. Smartwatch accelerometers have been especially useful in HAR, while gyroscopes are also helpful, especially when combined with the accelerometer. This study therefore suggests the use of HAR to detect sitting in order to identify the threat of OWS through the use of data from the accelerometer and gyroscope in a smartwatch. Two ensemble learning approaches have therefore been examined in order to establish their capability in recognizing periods of sitting. The results reveal that the accelerometer and gyroscope are complementary in their application and using them in combination thus improves the accuracy of recognition. The ensemble learning-based techniques were able to achieve an accuracy level of 93.57% for activity recognition when detecting sitting.
