Smart Alarm Driver Assistance as an Early Warning of Drowsiness Drivers Based on Raspberry pi 4 Model B

Abstract

A decrease in the level of awareness due to drowsiness while driving a motorized vehicle can increase the risk of traffic accidents. Smart Alarm Driver Assistance is a computing system development based on one board Raspberry Pi 4 model B. Digital image processing is intended to automatically recognize signs of drowsiness and provide warnings as early as possible. It is hoped that this research will reduce the number of road accidents caused by drowsy drivers. Raspberry pi is integrated with Raspberry pi camera noir v2 8MP for real-time eye monitoring, audio speaker and I2C OLED 128x64 as a warning output. Shape Predictor_68_Facial landmark model is used to detect Eye Aspect Ratio (EAR) as a parameter of driver drowsiness. The results of the test obtained an average response time of 0.82 seconds for the audio speaker to turn on and the average time the 128x64 OLED I2C lights up 0.90 seconds. This research can prove that the Raspberry Pi 4 model B can be implemented for drowsiness detection and warning. In the future, the protoype can be developed in terms of software and proper tool design.