ACS develops non-intrusive detection methodology of drowsy driving based on eye tracking data

3 MinutesBy NZ Trucking magazineJune 25, 2019

According to the US National Sleep Foundation, driving while sleep deprived has similar effects on your body as drinking alcohol. For instance, staying awake for 18 hours straight can make you drive as if you have a blood alcohol concentration of .05. In addition to this, the National Highway Traffic Safety Administration states that 72,000 police-reported motor vehicle collisions on US roads in 2015 were related to drowsiness of drivers, causing 41,000 injuries and 800 deaths.

As the need for the development of early warning systems for monitoring drowsiness in drivers is critical for preventing motor vehicle collisions and human fatalities, ACS (Alcohol Countermeasure Systems), has recently published an article in SAGE Publishing and Transportation Research Record based on the use of eye tracking data as a non-intrusive measure of driver behaviour for detection of drowsiness. The ultimate goal of his research is to lead to the development of technologies for real-time assessment of the state of vigilance, providing early warning of fatigue and drowsiness in drivers.

“Driving is demanding; from planning and navigating a route, interacting with traffic, following road safety practices, such as maintaining appropriate speed levels to basic vehicle operation, such as changing gears, braking and even turning on windshield wipers,” says Dr Zandi, ACS research scientist. “Being drowsy and sleep deprived only adds more interference to an already complex task, presenting dangerous risks.”

The ACS research study included two independent experiments that involved 53 volunteers: a short morning control driving session that lasted for 10 minutes, and a longer mid-afternoon monotonous driving session that lasted for 30 minutes. Both simulated sessions included driving on a low-traffic straight highway. In each experiment, infra-red-based eye tracking systems and cameras monitored the various eye movements of participants and advanced machine learning methodologies were employed to analyse the data and estimate the state of vigilance.

The results of the experiments verify a high correspondence between extracted eye tracking features and behavioural and physiological measures of vigilance. Results also confirmed that drowsy driving can be reliably detected with high accuracy, sensitivity, and specificity using eye tracking data and an appropriate classification framework.

With drowsy driving currently listed as one of the leading causes of motor vehicle accidents in North America, ACS research is necessary and will ultimately lead to the development of innovate, user-friendly products and services to promote road safety and keep people safe.

Research article: