A Novel System for Real Time Drowsiness Warning and Engine Ignition Authorization using Face Recognition

With the advancement of technology, there is huge progress in automotive industry. This paper proposes two automotive solutions to enhance safety and security features in vehicles. The first proposed solution is Real Time Driver Drowsiness Warning System implemented on Raspberry Pi 4B platform, with Raspbian operating system. This system is developed by applying transfer learning on CNN based pretrained model MobileNetV2. OpenCV libraries are used to detect drowsiness. The second proposed solution is authorized engine ignition using face recognition. Face recognition is achieved by using Caffe based detector and FaceNet algorithm. The recognized face of the user will act as key to enable engine ignition. Thus, this system will be able to reduce private vehicle theft. Driver drowsiness warning system is trained using MRL Eye dataset and achieved 98.75% accuracy.

I. Introduction

Artificial Intelligence (AI) and Machine Learning act as key enablers for Future transport solutions. Many automotive industries continue research on AI and Deep Learning (DL) to harness the power of data for automation, prediction, and optimization in a wide range of application. This paper proposes two automotive solutions, where the first proposed solution is driver drowsiness detection and alarming the driver to prevent tragic accidents thereby securing the life of fellow passengers and other road users. The second proposed solution is an auto ignition system which enables keyless engine ignition using face recognition technology, thereby the security system of the vehicle can be improved, and the theft rate of automobiles can be reduced.