Automotive technology to prevent accidents caused by drowsiness

Drowsiness during driving is a worldwide hazard, motivating a number of campaigns to encourage drivers to take regular breaks and plan journeys responsibly. Automating identification of the onset of drowsiness can help to eliminate tragic human errors of misjudging when it is no longer safe to drive. A Japanese company, Aisin Seiki, has developed an eyelid monitor and identified symptoms in eyelid behaviour that indicate drowsiness in motorists.

A Japanese company, Aisin Seiki, has developed an eyelid monitor and identified symptoms in eyelid behaviour that indicate drowsiness in motorists. The automated technology - Intelligent Transport System (ITS) - is designed to eliminate the fallibility of human judgement in gauging when drowsiness is becoming a driving hazard.

Physiological factors that may indicate drowsiness include eyelid movement, pupil fluctuation, heart rate and skin potential activity. Monitoring eyelid movements is a promising approach to detecting drowsiness before it becomes a hazard because monitoring it is generally less restrictive for the driver. Reliable use of eyelid movement symptoms to identify drowsiness required investigative analysis of eyelid behaviour recorded with special monitoring equipment to clarify the relationship.

Aisin’s eyelid monitor tracks a range of variables such as blink speed and frequency and average eyelid position. Trials of the new technology on motorists driving along a test track have confirmed the reliability of using these measured parameters to rate the drowsiness level of the driver.

Results from driving simulators include effects that are not always present in real-world circumstances, such as sickness and a greater tendency towards drowsiness due to the lower levels of stress in a simulator. For these reasons, Aisin Seiki used real test runs for motor equipment evaluations. Nine test drivers took part in the evaluation, each accompanied by an observer and an operator. They drove at a constant speed of 60 km/h for 80 minutes or until the observer judged that they were no longer fit to drive.

The monitoring equipment retrieves eyelid movement data from video frame images of the driver’s face recorded at 30 frames per second while driving. Eight quantities were recorded: average blinking intervals, average eyelid closure duration, maximum eyelid closure duration, blink frequency, average eyelid closure velocity, average eyelid opening velocity, average eyelid position, and integrated eyelid closure duration. These values were then used to identify five levels of drowsiness:

1. Not drowsy;
2. A little drowsy;
3. Drowsy;
4. Considerably drowsy; and
5. Very drowsy.

Significant difference statistical analysis confirmed the reliability of the drowsiness monitoring system. A significant difference of 0.1% was found for the recorded quantities.

Aisin Seiki has also identified the link between eye movement and lapses in motorist concentration during driving simulations. It used specialised eye monitoring equipment to record the movements of the left and the right eye during driving. Slow fluctuation of the eyes and rapid fluctuation of the vergence angle, which measures the focal point of the line of sight of the right and the left eye, were both found to precede incidents when the car deviated from its lane or the driver’s reaction times were increased.

The company has identified symptoms in eye behaviour that can reliably indicate driver drowsiness before an accident occurs. Due to the standard deviation overlap between the values recorded for eyelid movement quantities at adjacent drowsiness levels, more than one quantity may need to be monitored for reliable results. The company is developing an algorithm with a secured estimate accuracy that combines multiple measured eyelid movement quantities.

Drowsiness is a serious driving hazard worldwide and reportedly ranks as the number one cause of fatal car accidents in Japan. According to the company, the drowsiness monitoring systems will soon be available in Toyota’s cars.