An Innovative Solution with Wearable and Aboard-the-Vehicle Sensors Integrated with Machine Learning Algorithms for Monitoring the Driver's Psycho-Physical Condition for Safety Purposes

Recent progress in automotive safety has led to the development of an intelligent multi-sensor system for monitoring driver stress and fatigue. Monitoring the driver's psycho-physical state helps reduce risks to drivers and pedestrians. A literature review has been conducted on existing systems for detecting bio-signals and their use in evaluating the driver's psycho-physical condition. This study proposes the architecture of a monitoring system to classify driver fatigue using wearable devices and in-vehicle sensors to detect physiological signals, vehicle dynamics data, and traffic and weather information accessed by the driver's smartphone. The system employs an ensemble learning strategy to enhance the accuracy of fatigue detection compared to classic video-based systems. This approach aims to improve road safety by enabling intervention before a loss of attention occurs.