An Experimental Analysis on Enhancement of Electric Vehicle Safety using ADAS Technology and Forward Collision Avoidance with an Automatic Braking System

Abstract

The rapid expansion of the automobile sector in developing nations has intensified road safety concerns, particularly in congested urban environments where human error accounts for approximately 90% of all accidents. This paper presents and experimentally validates an integrated Advanced Driver Assistance System (ADAS) for electric vehicles comprising three complementary safety modules: a Forward Collision Avoidance System (FCAS) employing an HC-SR04 ultrasonic sensor interfaced with an Arduino Uno R3 to detect frontal obstacles within 50 cm and trigger automatic braking via an L298N motor driver; a Blind Spot Detection (BSD) system using three directional ultrasonic sensors with colour-coded LED alerts for lateral and rear blind zones; and a Driver Drowsiness Detection (DDD) system implemented on a Raspberry Pi 4 Model B using real-time computer vision, monitoring the Eye Aspect Ratio (EAR), Mouth Opening Ratio (MOR), and Nose Length Ratio (NLR) via the Dlib 68-point facial landmark model. Experimental results confirm reliable real-time operation of each module. The FCAS successfully engaged braking upon obstacle detection; the BSD system accurately identified directional blind-spot intrusions; and the DDD algorithm correctly classified four driver states: Active, Drowsy, Sleeping, and Head Bending with consistent accuracy across all test subjects. The proposed system is cost-effective (approximately USD 25 for hardware), modular, and scalable, providing a practical safety upgrade for commuter-segment vehicles currently underserved by commercial ADAS solutions.