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

Drowsiness on the road is a widespread problem with fatal consequences; thus, a multitude of solutions implementing machine learning techniques have been proposed by researchers. Among existing methods, Ghoddoosian et al.'s drowsiness detection method utilizes temporal blink patterns to detect early signs of drowsiness. Although the method reported promising results, Ghoddoosian et al.'s algorithm was developed and tested only on a powerful desktop computer, which is not practical to apply in a moving vehicle setting. We propose an embedded system that can process Ghoddoosian et al.'s drowsiness detection algorithm on a MiniPC and interact with the user by phone; combined, the vehicles are powerful enough to run a web server and our drowsiness detection server. In this paper, we explain how we implemented Ghoddoosian et al.'s drowsiness detection method in an embedded system using the AioRTC Protocol and propose the methods to compare the potential of various hardware setups and ultimately choose the most practical device for our embedded system. We evaluated the communication speed and processing times of the program on various platforms. Based on our results, we found that the Mini PC was most suitable for our proposed system. Furthermore, we proposed an algorithm that prioritizes the sensitivity of the drowsiness detection over the specificity. Based on the false positive and false negative rates, the algorithm optimizes the threshold of alerting the driver. We anticipate that our proposed platform can help many researchers advance their research on drowsiness detection solutions in embedded system settings.