A two-stage intelligent model to extract features from PPG for drowsiness detection

This paper presented a two-stage intelligent model that combined the wavelet packet transform (WPT) and functional-link-based fuzzy neural network (FLFNN) to access drowsy level. Early detection of drowsiness can help prevent drivers from involving in car accidents. According to the report of U.S. National Highway Traffic Safety Administration, drivers falling asleep while driving were responsible for at least 100,000 automobile crashes annually that resulted in an annual average of 40,000 non-fatal injuries and 1,550 fatalities [3]. Furthermore, the National Sleep Foundation had reported that 60% of adult drivers drove with drowsy and 37% of them had eventually fallen asleep during driving [4]. The fact behind those reports indicated that there is a dire need to develop a sensor device that detects drowsy status at an early stage.