Intelligent Driver State Monitoring and Warning System for Long-Haul Freight
Abstract
In long-haul freight transportation, driver fatigue and drowsiness are among the primary causes of road traffic accidents. To address this safety hazard, this focuses on long-haul truck driving scenarios and designs and implements a vision-based real-time monitoring and early warning system for long-haul truck driver drowsiness. The system collects video through a camera and utilizes deep learning algorithms to perform drowsiness recognition and risk assessment of the long-haul truck driver’s state. Based on the risk assessment results, it automatically warnings to the driver, thereby reducing the risk of accidents. This experiment not only completes a full closed-loop of data collection, state recognition, risk assessment, and warning feedback in laboratory, but also provides a systematic performance evaluation in both simulated and real-vehicle environments, verifying its engineering feasibility. The innovation of this study lies in the weighted fusion of multimodal temporal from the eyes and mouth. Through time-series analysis of multi-feature fusion, a comprehensive assessment of the long-haul truck driver’s drowsiness state is achieved, providing an technical solution for preventing long-haul truck driving accidents.