Designing an Intelligent System to Detect Stress Levels During Driving
Mohammad Karimi1, Zahra Khandaghi2, and Mahsa Shahipour1
1Department of Biomedical Engineering, Islamic Azad University, Iran
2Department of Medical Radiation Engineering, Shahid Beheshti University, Iran
Abstract: In addition to the devastating effects of anxiety and stress on the development and exacerbation of the cardiovascular disease, lack of stress control increases drivers' risk of accidents. This paper aims to identify the stress of drivers in various driving situations to warn the driver to control the tense conditions during driving. In order to detect stress while driving, we used the heart signals in the Physionet database. To analyze the conditions of the electrocardiogram (ECG) under various driving situations, linear and non-linear features were used. The characteristics of the RRIs are the only able to identify driver stress in different driving modes relative to rest periods, while the return mapping features, in addition to identifying driver stress while resting, have the ability to identify stress between different driving positions also brought. The results showed that driver's stress level during driving in city 1 and highway 1 with a P-value of 0.028 and also in city 3 and highway 2 with a P-value of 0.041 can be distinguished. The accuracy obtained from the proposed detection method is 98±2% for 100 iterations. The result indicated an efficiency of our proposed method and enhanced the reliability.
Keywords: Driver's stress level, Heart rate signal, return map, linear and non-linear features, statistical analysis.
Received March 20, 2020; accepted November 24, 2020