Don’tFall Asleep at the Wheel!
October 31, 2013
Sławomir Gruszczyński talks to the Polish Science Voice’s Karolina Olszewska.
Using funds from the National Center for Research and Development (NCBiR), Sławomir Gruszczyński, Ph.D., head of the Department of Electronics at the AGH University of Science and Technology in Cracow, has developed a system using radar technology to prevent drivers from falling asleep behind the wheel. His driver fatigue monitoring system is composed of a motion detector with multiple sensors, infrared cameras, a digital signal processing unit, and a decision-making unit capable of generating alarm signals.
While working on the project, Gruszczyński had to develop special algorithms to monitor the position and movements of the driver’s head and to analyze micro-Doppler signals to monitor the driver’s heart rate, breathing and pupil movements. In order to increase the reliability of the design and to provide better pattern recognition, the inventor used a radar unit that scans the space around the driver.
The research project ended in May. Information about it has been published in scientific journals and presented at international conferences on radar technology and signal analysis. Gruszczyński has built a prototype and now plans to approach automakers, hoping to persuade them to apply the system in mass-produced vehicles.
What inspired you to undertake this kind of research?
Statistics on road accidents caused by drivers falling asleep; they show a huge number of fatalities and serious injuries. A driver who nods off to sleep cannot respond to danger in any way. Research carried out by experts from the Department of Clinical Neuroscience at the Karolinska Institute in Stockholm, Sweden, on a representative group of several dozen people has found that the level of driver fatigue can be measured by analyzing the way in which they blink their eyes. After viewing these findings, I decided to design a device to detect and monitor the way the driver blinks his or her eyes.
I came up with the idea and managed a project carried out at the Department of Electronics of the AGH University of Science and Technology in Cracow. The project aimed to develop a driver fatigue monitoring system and involved a total of seven researchers, including Bogusław Cyganek and Krzysztof Wincza.
Various countries in Europe realized long ago that the costs of road accidents are so huge that even a several-percent reduction in these unfortunate events results in tangible savings. That’s why the European Commission promotes all solutions, especially automatic and autonomous ones, that support the driver and enable them to avoid potentially dangerous situations. This results from the fact that 57 percent of accidents are directly caused by human error, and in 90 percent human error plays a significant role.
How can you tell that the driver is close to falling asleep?
The device is equipped with two types of sensors. The first sensor uses an infrared camera mounted in the car thanks to which the whole system is independent of changing lighting conditions. These are a major problem in optical systems operating in the visible light range. In the image from the camera, the sensor singles out the driver’s eyes and detects how the driver blinks his or her eyes and analyzes this.
Such image analysis is a very effective tool. But it does not work properly when the driver is wearing sunglasses. Consequently, it is necessary to use a second sensor that works independently of the first one. This is a microwave radar device that uses the so-called Doppler effect. Microwave radiation easily penetrates through sunglasses and, reflecting off the driver’s eyes, provides information on eye movements. It is worth noting that this does not pose any threat to the health of the driver and passengers. The power of the radar used is very small.
Based on information collected on the blinking process, the level of the driver’s fatigue is measured. If it is high, the system will notify the driver and get him or her to stop the vehicle.
Who can install such a system in a car or bus?
Specific solutions depend on the design of the vehicle cabin. Both the infrared camera and the radar device must be directed towards the driver’s face. Depending on how much space there is, the system can be mounted on the dashboard, for example behind the wheel or above the rear-view mirror on the upper edge of the windshield.
The unit for processing the data from the system, in turn, can be hidden in or integrated into the car’s onboard computer. The device is a set of three modules and can be adapted for use in a specific vehicle model depending on the amount of space available, for example.
The monitoring system, due to the need to integrate the alarm with the onboard computer, for example, and because of the design of the radar and camera antennas, needs to be installed at the stage of producing the vehicle. Of course, the monitoring system could be produced as a separate device, to be attached by the user. However, selecting the right position for the sensors at an early stage of building the vehicle will increase the effectiveness of this device.
When it detects a dangerous level of driver fatigue, it generates a warning signal. The driver can get a warning through a voice message, a blinking light or another signal. The manufacturer of the car chooses the type of alarm.
What is innovative about your project?
Such devices mainly use optical systems operating in the visible range and are exposed to highly variable conditions in terms of car lighting and to reflections appearing when the driver is wearing glasses. What is especially innovative about our project is the use of the microwave radar sensor, thanks to which the detector is not affected by these factors. In addition, the detector works even when the driver puts on sunglasses. Another new thing is the use of an infrared camera with a special system for illuminating the driver’s face (in infrared), which improves the quality of the analyzed images and increases the reliability of the device.
Our idea of combining the microwave radar and the infrared vision system is an innovative solution in such systems. This will help increase the reliability of driver fatigue monitoring systems and thus help reduce the number of car accidents.
In the case of such systems, one of the key factors is the price of the product. As a result, research into compact and low-cost driver fatigue detectors is particularly being promoted. This is largely why I received around zl.900,000 from the National Center for Research and Development for my research work.
When will the prototype be put to commercial use?
The project ended in May. The prototype works as we expected. But in order to use it in vehicles, it is necessary to adapt its parameters to the specific solutions of car makers.
Tests have revealed some problems. One of these is a high level of interference in the radar sensor. We must improve the blink detection algorithms in the presence of high-amplitude interfering signals.
Work is still in progress to upgrade the radar part as well. Several concepts are being developed, which will lead to a situation in which the radiation beam will be focused on the driver’s face, instead of being dissipated to the surroundings.
The prototypes of the radar and optical sensors have been tested in both laboratory and real-life conditions. We are currently working to improve the effectiveness of the device before we offer it to potential customers.
Talks with automotive industry professionals are planned after this stage of research work is completed. We are now preparing the technology, which will soon be offered to potential producers. We are thinking mainly of our foreign partner, with whom we have been working for quite a while, and who could help us put our work to commercial use. Even if the system is not offered in whole, then using the modules in industry on a mass scale could prove a commercial hit.
It is difficult to determine how much such a system will cost because this depends on the specific solution. Still, the price should not exceed zl.1,500 or so.