IoT-based drowsiness detection and prevention system using face landmark detection and eye aspect ratio calculation for road freight transport

ABSTRACT: Drowsiness-related road incidents are common in road freight truck driving due to the nature of the occupation. With extended hours of operation and scheduling issues, there is a high risk of experiencing fatigue and drowsiness during the driver’s shift. This study addressed the problem by implementing Drowsiness Detection, Sleep Prevention, Website Application and Database, and Data Processing System to directly combat drowsiness and communicate the incident to the company. Specifically, this study utilized facial landmark detection and eye aspect ratio calculation to detect drowsiness while Arduino logic programming was used to control the output devices. The concept of WebSocket was used to enable a full-duplex communication channel between the two devices and HTTP POST to send information to the web server. To properly test the accuracy of the whole system, several testing procedures were done in three different setups wherein standby testing was done in environment setup and situational and accuracy testing was done in controlled setup. Most of the actual results matched the expected results in the test scenarios, with nighttime having lighting issues but was disregarded because of the presence of headliner and ambient lights. A test driver survey was conducted, and several suggestions were made such as better build design and placement. In conclusion, the drowsiness detection, and prevention system were successfully accomplished and proved to be effective. On this basis, the problem must be properly addressed to deploy certain wide-scale measures and ultimately prevent road accidents.

Filipiniana