Sistema inteligente de prevenção de fadiga em condutores de veículos na mineração

Abstract

This work presents an intelligent system for preventing driver fatigue, validated in the operational environment of Vale S.A., which integrates Computer Vision, Vehicle Telematics, and Fuzzy Logic. In mineral exploration activities, prolonged working hours and adverse conditions—such as variable lighting and vibration— increase fatigue risk and reduce driver vigilance. The system combines, in real time, facial indicators such as Eye Aspect Ratio (EAR) and head tilt angle with vehicle operational variables, including speed variation and continuous driving time, to estimate the driver’s fatigue state. Fuzzy inference consolidates multimodal evidence and enables the anticipation of critical conditions, issuing preventive alerts before extreme levels of drowsiness are reached. In tests using video samples and vehicle telematics data collected at Vale, the method achieved an 86% approval rate, demonstrating superior effectiveness compared to approaches based solely on eye closure. The results reinforce the potential of Computational Intelligence for risk mitigation and the strengthening of Occupational Health and Safety in the mining sector.