More efficient method developed to monitor drivers’ attention levels

Fifty-two people have died on Catalan roads so far this year, according to data from the Catalan Traffic Service. Although there have been fewer accidents in the first four months of the year than in the same period last year, accidents continue to have a terrible impact: more than 3,500 people were involved in 2,414 traffic accidents in 2024, 48 of them with fatalities.

In more than 85% of cases, the human factor is to blame, with distractions, fatigue, drowsiness, and consumption of alcohol and other substances being recurrent causes. Aware of this, the automotive industry is working on active safety systems that—among other things—monitor drivers’ alertness and warn them when they are not paying attention. Researchers from the NEPHOS research group have developed a prototype capable of doing this more efficiently than conventional systems.

The paper is published in the IEEE Sensors Journal.

The system they propose is relatively simple and therein lies its main strength. A small, low-cost radar device capable of emitting and detecting electromagnetic waves is mounted on the driver’s sun visor.

Two frequency selective surfaces (FSS) are placed on the headrest on either side of the driver’s head. Each surface modifies the radar waves in such a way that when they bounce back to the receiver, the receiver is able to distinguish which surface they are coming from: right or left. When the driver moves his head, the line of sight between the radar and one of the FSS surfaces is blocked to a greater or lesser extent.

In this way, the system is able to monitor the driver’s movements.

The research team calibrated the system in the laboratory and inside a car (moving and stationary) in order to relate head movements to the driver’s state of attention at the wheel. With the information gathered in the experimental tests, they developed an algorithm that identifies the driver’s state of attention based mainly on the frequency of head movements.

“When you drive, you normally look straight ahead, with small changes in the direction of your head to check the mirrors and the sides; the movements are different when you are distracted,” explains Farid Morabet, researcher at the Department of Electrical, Electronic and Automatic Engineering.

Thanks to its simplicity, this proposal has great advantages over conventional systems. While many use a camera that records the driver, the URV alternative relies on radar technology. This has more benefits than one might expect. As the technique involves no images, there are no issues concerning image rights or privacy. Also, the hardware can be optimized and simplified, significantly reducing manufacturing and maintenance costs and energy consumption.

“You don’t need a big computer; a small microprocessor is able to do the calculations,” explain the researchers.

Another distinguishing feature of the prototype is its so-called “acoustic immunity.” Systems that detect the positioning of objects by ultrasound—in the same way that bats do—are vulnerable to acoustic interference. This is not the case with radar, which has another advantage over cameras insofar as it does not depend on lighting conditions.

Despite the potential of this technology, Morabet is cautious about its application.

“It will take a lot of field testing and a long validation process before it can be implemented,” he says.

In addition to developing the prototype, the researchers have discovered other possible uses of FSS surfaces in motion detection, with very diverse potential implications. They are currently working on a system that can recognize more complex driver gestures, which can be used to control vehicle accessories such as the radio or heating.

Other ways this technology can be applied are in improving accessibility and autonomy for people with reduced mobility, and in medicine, in developing wearable communication systems that can read biomarkers, thus eliminating the need for invasive procedures