Comprehensive Measurement and Evidential Evaluation of Driver Drowsiness and Alertness Warning Systems.

The paper aims to improve the accurate quantification of driver drowsiness and to provide comprehensive, evidence-based validation of a vision-based driver drowsiness and attention warning system. Advanced quantification of driver drowsiness is designed to enhance distinction of true positive events from false positive and false negative events. The methodology included assessing inputs such as driver visibility, dynamic driving tasks, driving patterns, driving course time and vehicle speed along with Percentage of Eye Closure (PERCLOS). The system is programmed to actively learn and adapt personalised eye aspect ratio (EAR) threshold value and to process EAR frames exclusively against the learnt threshold value. This method refined the data frames for evaluation, facilitating the processing of necessary frames and mitigating delays in Human Machine Interface (HMI) warnings. Comprehensive validation is systematically conducted within a controlled test track environment to ensure precise execution of protocols, maintaining inputs closely aligned with real-time scenarios. The test methodology comprised the execution of pre-defined protocols and a technology-neutral procedure. Pre-defined protocols are scenarios created using the aforementioned inputs. The Cartesian coordinates of the system camera and driver eye point relative to the seating reference point (SgRP) are identified using a coordinate measurement machine (CMM) to measure the driver's position within the camera's field of view. The protocols are executed with precision using a global navigation satellite system (GNSS), visual sensor, audio sensor, eye tracking device, and data logger. Subsequently, the system is tested with number of drivers trained on the Karolinska Sleepiness Scale (KSS) to conduct technology-neutral method for statistical analysis. The paper details the calculation and analysis of various iterations of the tested data, concludes with improved results, and explores future prospects for quantifying driver drowsiness. The paper also discusses observations and challenges related to the functionality of conventional systems and protocols currently deployed in the market.