Driver Drowsiness Detection System for “in-use vehicles”

Abstract

This paper presents the development of a low-cost, automotive-grade OBD-based driver drowsiness detection system designed for in-use passenger vehicles lacking such a system as a built-in feature. The system leverages real-time vehicle data accessed via the On-Board Diagnostics (OBD) interface to monitor driving behavior and detect signs of driver fatigue and inattentiveness. By learning the vehicle’s normal operational patterns, the system dynamically sets upper and lower thresholds for key parameters such as speed, acceleration, engine load, RPM, and braking behavior. The algorithm is designed to minimize false positives. A key feature that makes the system more reliable is its time-dependent sensitivity, which adjusts detection thresholds based on day and night driving conditions. Apart from that, the algorithm identifies critical events such as sudden braking, abrupt acceleration, and unfocused driving maneuvers, which may be considered indicative of drowsiness or distraction. The plug-and-play nature of the device ensures easy installation and compatibility across a wide range of vehicles, making it a scalable solution for enhancing the road safety of in-use vehicles. This paper details the system architecture, detection methodology, and validation results, highlighting its potential as a practical and accessible safety enhancement. Keywords: Drowsiness detection, Automotive Safety, On-Board Diagnostics, in-use vehicle