2026-06-13T01:59:54Z https://meral.edu.mm/oai

oai:meral.edu.mm:recid/2433 2021-12-13T03:03:23Z 1582963390870:1582967549708 user-uy

MICROCONTROLLER-IMPLEMENTED ARTIFICIAL NEURAL NETWORK FOR ELECTROOCULOGRAPHY-BASED WEARABLE DROWSINESS DETECTION WITH ALERTSYSTEM Tabar, Keith Marlon R. Caluyo, Felicito S. Ibarra, Joseph Bryan G. Drowsiness has been one of the leading causes of work-related accidents. One reason that is pointed out by the Royal Society for the Prevention of Accidents (2001) is that, drowsiness tends to reduce the reaction time and attentiveness of a person resulting to poor performance on attention-based activities. This is due to the fact that the speed at which information is processed in the brain is also reduced by drowsiness (NCSDR/NHTSA 1998). Different methods have been explored to develop an effective drowsiness detection system (DDS) to give drivers warning of impending drowsiness. However, no study implemented a wearable and standalone DDS. Most of the existing DDS require the use of a computer application or a separate processor for signal processing and drowsiness detection. The present study aimed to design a wearable electrooculography (EOG)-based DDS that doesn't require a computer to operate; to implement an artificial neural network (ANN) into a microcontroller; to determine the best electrode placement setup on the visor cap for optimal extraction of EOG signals; to test the accuracy, precision and sensitivity of the system in real-time; and to evaluate the system in terms of comfort and unobtrusiveness. Figure 1 below shows the methodology from which the development of the system was based upon. 2015 http://hdl.handle.net/20.500.12678/0000002433 https://meral.edu.mm/records/2433