Driving fatigue is something that every driver has encountered, especially in the driving environment of long-distance and traffic jam. According to statistics, fatigue related traffic accidents account for about 20% of the total accidents and more than 40% of major traffic accidents in China. Fatigue driving is an international problem in the field of traffic safety management, which widely exists, which is closely related to the inadequate implementation of the functions required by the standards of fatigue driving monitoring equipment.
Of course, to solve this problem, car factories and aftermarket are trying.
At the fourth World Internet Conference, Robin Li said that intelligent driving is the area where Baidu focuses on artificial intelligence. In this area, Baidu will launch a system of anti fatigue driving with FAW next year, because “we hope that the truck drivers will not lose their lives before they lose their jobs.” Baidu fatigue driving monitoring system will be one of the landing methods of image recognition technology based on Baidu brain. It will judge through infrared face recognition. When the truck driver enters the fatigue state, the system will send corresponding warning information.
The detection methods of driver’s fatigue state can be roughly divided into detection methods based on driver’s physiological signal, physiological response characteristics, operation behavior and vehicle state information.
Based on driver’s physiological signal
The physiological indexes of drivers in fatigue state will deviate from the indexes of normal state, and the research on fatigue driving was first based on this aspect. At present, the more mature detection methods include the measurement of EEG and ECG of drivers.
The detection method based on the driver’s physiological signal has high accuracy in judging fatigue, but the physiological signal needs to be measured by contact and is highly dependent on individuals. Therefore, there are many limited factors in the actual operation, so it is mainly used in the experiment as the control parameter.
Based on the physiological response characteristics of drivers
The detection method infers the driver’s fatigue state by using the driver’s eye movement characteristics and head movement characteristics. The eye movement and blink information are considered to be important features reflecting fatigue.
At present, there are many algorithms on eye movement mechanism. PERCLOS, a widely used algorithm, takes the percentage of eyelid closure time in a period of time as the measurement index of physiological fatigue. However, the problem faced by this kind of algorithm is that the adaptability of the algorithm is difficult to optimize due to different races and different eye socket depths. Moreover, when facing the change of scene light, the accuracy of face recognition will also be affected.
Based on driver’s operation behavior
The driver’s fatigue state can be inferred by processing the driver’s steering wheel operation data. At present, there are many uncertainties in this way, which is limited by personal habits, driving speed, road environment and operation skills, so there are few in-depth research results in this aspect.
Based on vehicle trajectory
The driver’s fatigue state is inferred from the driving information such as vehicle trajectory change and lane line deviation.
It is reported that there are two technical thresholds for fatigue detection, that is, the complexity of the application scenario and the bottleneck of the algorithm itself.
Let’s take a look at the tips provided by major car companies and aftermarket to deal with fatigue driving.
Mercedes Benz has its own unique system in fatigue driving detection and early warning, which is called “attention assistance system”, and is configured on C, Cl, e, GLK, s, GL, SLK and SL models.
Firstly, the system collects the driver’s driving behavior within 15-20 minutes after the vehicle starts, and files it as a reference system.
Through a series of sensors equipped on the vehicle, each command of the driver is monitored. Up to 70 kinds of data, such as driving time, steering angle, vehicle speed, acceleration and driver behavior, are recorded. The system monitors these data and compares them with the data in the file. So as to make a judgment when the driver is tired driving and give corresponding reminders. It is worth noting that this system starts when the vehicle speed is 60km / h.
Some models of Lexus, a high-end car brand of Toyota, such as ls600h L, can be equipped with driver attention monitor to record the driver’s facial expression through the infrared LED display installed around the steering rod. If there are continuous blinking, eye closing and line of sight deviation, the alarm will be remembered.
Volkswagen determines whether it is in a fatigue state through the information of steering angle sensor or electronic power steering system, that is, the driver’s steering action. This system is called “fatigue identification system”. If the driver’s steering behavior exceeds a specific value and the vehicle speed exceeds 65km / h, the system will recognize that the driver has signs of fatigue driving. Passat alltrack is equipped with this system as standard.
The system developed by Hyundai to prevent fatigue driving is to monitor the driver’s heart rate and vehicle driving status. The new Hyundai Sonata is equipped with a fatigue driving reminder system.
The fatigue monitoring system equipped by BYD is called “fatigue driving early warning system”. It is a device based on the driver’s physiological image response and composed of ECU and camera. It uses the driver’s facial features, eye signals and head mobility to infer the driver’s fatigue state, give an alarm and take corresponding measures.
At present, most of the vehicle manufacturers’ systems to deal with fatigue driving are equipped with medium and high-end models.
New ways to play in the aftermarket
Compared with small passenger car drivers, commercial vehicles such as buses, intercity buses, airport buses and operating trucks, especially passenger and freight vehicles running long-distance transportation, crash is a high-risk group of fatigue. Therefore, there is a demand for the aftermarket. It is reported that the fatigue driving monitoring system has two technical difficulties to overcome. On the one hand, it is the bottleneck of the algorithm itself, on the other hand, it is the complexity of the use scenario.
If you want to develop your own fatigue monitor system products, brands or software; please don’t hesitate to contact [email protected] , we won’t let you down.