This website requires JavaScript.

Embedded System Performance Analysis for Implementing a Portable Drowsiness Detection System for Drivers

Minjeong KimJimin Koo
Sep 2022
Drowsiness on the road is a widespread problem with fatal consequences; thus,a multitude of solutions implementing machine learning techniques have beenproposed by researchers. Among existing methods, Ghoddoosian et al.'sdrowsiness detection method utilizes temporal blinking patterns to detect earlysigns of drowsiness. Although the method reported promising results,Ghoddoosian et al.'s algorithm was developed and tested only on a powerfuldesktop computer, which is not practical to apply in a moving vehicle setting.In this paper, we propose an embedded system that can process Ghoddoosian'sdrowsiness detection algorithm on a small minicomputer and interact with theuser by phone; combined, the devices are powerful enough to run a web serverand our drowsiness detection server. We used the AioRTC protocol on GitHub toconduct real-time transmission of video frames from the client to the serverand evaluated the communication speed and processing times of the program onvarious platforms. Based on our results, we found that a Mini PC was mostsuitable for our proposed system. Furthermore, we proposed an algorithm thatconsiders the importance of sensitivity over specificity, specificallyregarding drowsiness detection algorithms. Our algorithm optimizes thethreshold to adjust the false positive and false negative rates of thedrowsiness detection models. We anticipate our proposed platform can help manyresearchers to advance their research on drowsiness detection solutions inembedded system settings.