Drowsiness Detection System: Integrating YOLOv5 Object Detection with Arduino Hardware for Real-Time Monitoring

Abstract

Drowsy driving remains a significant cause of accidents worldwide, prompting the need for effective real-time monitoring systems to detect and prevent driver fatigue. In this paper, we propose a novel approach for drowsiness detection leveraging state-of-the-art deep learning techniques and compact hardware implementation. Our system integrates the YOLOv5 object detection model with Arduino hardware, offering a portable and efficient solution for on-road application. The YOLOv5 model is employed for its superior speed and accuracy in detecting facial landmarks and identifying signs of drowsiness in real-time video streams. By focusing on the key features indicative of drowsiness, such as eye closure, head nodding, and yawning, our system can effectively discern driver fatigue levels with high precision. Furthermore, the utilization of Arduino hardware enables seamless integration of the detection system into vehicles, providing a cost-effective and accessible solution for widespread deployment. Leveraging the computational capabilities of Arduino, we optimize the inference process of YOLOv5 to ensure real-time performance on resource-constrained platforms. We present experimental results demonstrating the efficacy and efficiency of our proposed drowsiness detection system. Through rigorous testing in simulated driving conditions and real-world scenarios, we validate the system's ability to accurately identify drowsiness cues while maintaining low latency. Overall, our research contributes to advancing the field of driver safety technology by offering a practical and scalable solution for drowsiness detection. The integration of YOLOv5 with Arduino hardware showcases the potential for deploying sophisticated deep learning models in real-world applications, paving the way for enhanced road safety and accident prevention measures.