A team from Abu Dhabi University’s Electrical and Computer Engineering Department has invented a real-time fatigue detection system using artificial intelligence (AI). With many drivers working long hours, the system is designed detect drowsiness and alert the driver to prevent accidents and loss of life.
The low cost, easy to use system, which connects to the driver’s smart phone, consists of a Raspberry Pi 3 with an attached camera module and a deep learning fatigue detection system. It does not negatively affect the driving process at all and works by installing a small camera on the dashboard that triggers an alarm once persistent fatigue is detected.
Conventional fatigue detection systems based on computer vision are fast but less accurate. They are more sensitive to lighting conditions and imaging resolution and need to be calibrated for each driver. The ADU team’s system is designed based on deep learning approaches that boost detection performance and accuracy. It has been tested on a large number of drivers and works with different facial features. Although still under development, the design and testing stages have been completed and the team is now working on enhancing performance in terms of accuracy and speed of detection.
Dr. Mohammed Ghazal, Chair of the Electrical and Computer Engineering Department, Yasmin Abu Haeyeh, Teaching Assistant, Abdelrahman Abed, master’s student, and Sara Ghazal, a mobile applications developer, make up the team.