Stress Level Detection Using Physiological Sensors

Gunaydin O., ARSLAN R. B.

20th IEEE International Conference on Bioinformatics and Bioengineering, BIBE 2020, Ohio, United States Of America, 26 - 28 October 2020, pp.509-512 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Volume:
  • Doi Number: 10.1109/bibe50027.2020.00088
  • City: Ohio
  • Country: United States Of America
  • Page Numbers: pp.509-512
  • Keywords: EDA, ECG, HR, SFS, RDF


© 2020 IEEE.According to the World Health Report published in 2018, in every 24 seconds, someone dies on the road. One of the causes that lead to traffic accidents is drivers' mental workload and stress. In this paper, ways of detecting drivers' stress are discussed, previous studies are examined, and an experimental setup for detecting stress is built. For experiments, a racing game is used. One subject played five different levels of a racing game while her face and play screen were recorded, together with EDA (Electrodermal Activity) and ECG (Electrocardiogram) signals acquired through sensors attached to her body. Recorded games are used for identifying hidden stressors that may cause stress on the subject. Statistical features were extracted and the Random Decision Forest (RDF) algorithm is used for classification. RDF yields an accuracy rate of 70.74% when all five level records are used and between 70% - 80% for individual records.