Abstract

According to CNESST data, workers compensated because of a work-related traffic accident represent approximately 2% of all those who receive compensation, but traffic fatalities represent between 25% and 30% of all accidental deaths at work. In fact, they are the leading cause of accidental death at work. According to a statistical study conducted on data from more than 8000 workers who received compensation from the CNESST following such an accident, between 2000 and 2008, over 83% were the drivers of the vehicle involved. While drivers’ individual characteristics and behaviours are important considerations when developing a prevention strategy, according to previous studies driver fatigue while on the job is the most significant factor among the range of risk factors identified.

Among the various accident prevention strategies, the use of in-vehicle driving safety aids to warn drivers of drowsiness is the most direct way to reduce this risk. Recent studies suggest that the variability of physiological parameters such as heart rate and breathing are promising indicators for use in drowsy driving detection systems.

The aim of this research project is to build a functional prototype based on the use of sensors to detect vital signs without direct skin contact, integrated into the vehicle seat cover and connected to a microprocessor. An algorithm will be developed to convert the signals detected into biometric data, while filtering out noises such as body movements. Post-processing and formatting of this data is envisaged to train a machine-learning algorithm to determine the stages of drowsiness while driving.

The study will consolidate knowledge on the development of biometric sensors, using smart textile technologies. It will also provide a better understanding of the biometric variables that identify the signs of drowsiness while driving. The research team plans to develop a suitable method to obtain useable biometric data and to implement a tool to detect drowsiness in workers who are driving. It will also seek to validate the processing of biometric signals as a predictive measure of drowsiness and deterioration of driving skills.