CN108492527B - Fatigue driving monitoring method based on overtaking behavior characteristics - Google Patents

Abstract

The invention discloses a fatigue driving monitoring method based on overtaking behavior characteristics, comprising the following steps: (1) by monitoring the turn signal, steering wheel angle and vehicle speed, to determine whether the driver is overtaking; (2) obtaining the vehicle through sensors and other equipment Fatigue characteristics in the process of overtaking, among which fatigue characteristics include: vehicle speed parameters, lane position parameters, and overtaking duration parameters; (3) The fatigue characteristics are used as the fatigue discrimination index set, and the SVM classification has been trained based on the fatigue driving experimental data that has been carried out in advance. The device is used to classify and determine the driver's fatigue state. Through the above method, the present invention can analyze the characteristic changes of the driver's overtaking behavior in the fatigue state, further improve the comprehensive fatigue monitoring accuracy of the driver, reduce the phenomenon of missed alarms and frequent false alarms caused by low algorithm accuracy, and prevent the occurrence of fatigue accidents .

Description

A fatigue driving monitoring method based on overtaking behavior characteristics

technical field

The invention relates to a vehicle driving assistance technology, in particular to a fatigue driving monitoring method based on overtaking behavior characteristics.

Background technique

Among the road traffic accidents in my country, fatigue driving is one of the main reasons for the highest number of fatalities in a single accident. When the driver is in a fatigued state, the reaction time increases, the action slows down, the cognition, judgment, operation error and the "micro-sleep" state appear, all of which are the direct causes of traffic accidents. According to a conservative estimate by the National Highway Traffic Safety Administration, there are at least 100,000 recorded drowsy-driving traffic accidents in the United States each year, resulting in 1,550 deaths, 71,000 injuries, and an economic loss of $12.5 billion. At the same time, the research results of the American Automobile Association show that in every 6 traffic accidents, one is caused by drowsiness. Fatigue driving also accounts for about 25% of traffic accidents involving casualties on German autobahns. The analysis of the causes of road traffic accidents in my country in 2011 shows that although fatigue driving is not the cause of the largest number of fatalities, the number of fatalities in a single accident ranks second among all major causes, only after illegal loading, with an average of every two accidents. A traffic accident due to fatigue causes the death of 1 traffic participant. At the same time, from 2004 to 2011, the results of in-depth analysis of the causes of serious traffic accidents with more than 10 fatalities in my country show that there is an average of 1 accident caused by fatigue driving every year. The analysis further shows that fatigue driving is one of the main reasons for malignant driving accidents. .

At present, the monitoring of fatigue driving is mainly divided into the monitoring of the driver and the monitoring of the vehicle. The monitoring of drivers mainly includes eye feature recognition, facial feature recognition, driver operation behavior monitoring, and physiological response monitoring. In addition, there are also studies on the mechanism and characterization of fatigue by using contact sensors to detect human peripheral physiological signals. This method has high cost, complex structure, poor scalability, and is easily affected by light and individual driver factors. Most of the research on vehicle monitoring is based on vehicle state information, such as driving time, driving speed, driving route, steering wheel angle, relative road offset, etc. This method is affected by various external factors such as vehicle type, road conditions and weather, so the accuracy of fatigue detection is not high, and its anti-interference and adaptability are poor.

To sum up, due to the late start of research on fatigue driving in my country, and most of the research is carried out in the environment of simulated experiments, there is still a certain gap with the driving in the real vehicle environment, so many problems still need to be solved. In the process of studying fatigue state, the method of fatigue state calibration is relatively simple, and it is difficult to make further breakthroughs. Most of the detection methods have problems such as high result error, long monitoring time, poor real-time performance, low sensitivity, poor reliability, and high cost.

SUMMARY OF THE INVENTION

The technical problem to be solved by the present invention is to provide a fatigue driving monitoring method based on overtaking behavior characteristics, aiming at the defects in the prior art.

The technical solution adopted by the present invention to solve the technical problem is: a fatigue driving monitoring method based on overtaking behavior characteristics, comprising the following steps:

A fatigue driving monitoring method based on overtaking behavior characteristics, comprising the following steps:

1) By monitoring the turn signal, steering wheel angle and vehicle acceleration, according to the monitoring results to determine whether the driver is overtaking;

The details are as follows: according to the monitoring results, confirm whether the left turn signal is turned on, whether the steering wheel angle is continuously greater than 6°, and whether the vehicle is in an accelerating state; when the three conditions are met at the same time, it is determined that the overtaking behavior is being performed;

2) Collect overtaking data by sensors, and the overtaking data includes: vehicle speed record data starting from the moment of overtaking behavior, lane position information from the moment of overtaking behavior, and the start time and overtaking end time of overtaking behavior; the The method for judging the end of the overtaking behavior is as follows: when the right turn signal of the vehicle is turned off, it is determined that the overtaking behavior ends.

3) Obtain the fatigue characteristics during the overtaking process of the vehicle according to the collected overtaking data, wherein the fatigue characteristics include: vehicle speed parameters, lane position parameters, and overtaking duration parameters;

The vehicle speed parameters are obtained by the following methods:

Obtain the maximum value of the overtaking speed V _max : that is, the maximum value of the vehicle speed in an overtaking sample;

Obtain the minimum value of the overtaking speed V _min : that is, the minimum value of the vehicle speed in an overtaking sample;

Obtain the overtaking speed range V _range : that is, the difference between the maximum and minimum vehicle speeds in an overtaking sample: V _range =V _max -V _min ;

Obtain the average speed of overtaking V _mean : that is, the average speed of vehicles in an overtaking sample;

The lane position parameters are obtained by the following methods:

Obtain the lane departure standard deviation SDLP, which describes the current driver's operating state;

Among them, d _avg is the mean value of lane positions in the sampling period; d _i is the lane position value in the sampling period; n is the number of lane position samples in the analysis sampling period; the sampling period is the time from the beginning to the end of an overtaking behavior.

Wherein, the overtaking duration parameter is obtained by the following method:

Obtain the overtaking behavior start time T _start : that is, the overtaking start time in an overtaking sample;

Get the overtaking behavior end time T _end : that is, the overtaking end time in an overtaking sample;

Obtain the overtaking behavior duration T: that is, the difference between the overtaking end time and the start time in an overtaking sample: T=T _end -T _star ;

4) The fatigue feature is used as the fatigue discrimination index set, and the classification is performed based on the SVM classifier that has been trained with the data of the fatigue driving experiment performed in advance, and the driver's fatigue state is determined.

According to the vehicle speed parameters, lane position parameters, and overtaking duration parameters extracted from the fatigue data in step 3) as the fatigue discrimination index set, the classification is performed based on the SVM classifier that has been trained with the fatigue driving experimental data performed in advance, and the driver's fatigue is determined. state;

The driver's fatigue state is divided into three levels, namely, awake, fatigued, and very fatigued, and the output of the classifier is one of the three fatigue levels.

The beneficial effects of the present invention are as follows: the present invention provides a fatigue driving monitoring method based on the characteristics of overtaking behavior. The method can aim at frequent overtaking behaviors on expressways with serious accident consequences. The characteristic changes of the algorithm can further improve the comprehensive fatigue monitoring accuracy of drivers, reduce the phenomenon of missed alarms and frequent false alarms caused by low algorithm accuracy, and prevent the occurrence of fatigue accidents.

Description of drawings

The present invention will be further described below in conjunction with the accompanying drawings and embodiments, in which:

Fig. 1 is the method flow chart of the embodiment of the present invention;

FIG. 2 is a flowchart of a method for judging an overtaking behavior according to an embodiment of the present invention.

Detailed ways

In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, but not to limit the present invention.

As shown in Figure 1, a fatigue driving monitoring method based on overtaking behavior characteristics includes:

Step 1: Monitor the driver's behavior and determine whether to overtake.

Comprehensive judgment is made by three indicators: turn signal, steering wheel angle and vehicle speed, including monitoring whether the left turn signal is turned on, monitoring whether the steering wheel angle is continuously greater than 6°, and monitoring whether the vehicle is accelerating. When two or more of the three conditions are satisfied, it is determined that the overtaking behavior is performed, and the specific process is shown in Figure 2.

Step 2: Use sensors and other equipment to collect overtaking data.

The speed data during the overtaking process is obtained from the CAN output of the vehicle; the lane position information is collected by the Mobileye device, including the left and right lane line spacing. The center of the vehicle is used as the vehicle position. When the lane line spacing is 1.8m, it means that the vehicle is in the middle of the lane. The default value is also 1.8m when the image recognition algorithm cannot detect the lane line. ; The beginning of the overtaking behavior is the overtaking start time, and the overtaking end and returning to the original lane are the overtaking end time.

Step 3: Fatigue feature extraction. The first step is to perform data processing. The first step is to perform synchronization processing to delete the unsynchronized data in the initial stage to ensure that the collection time of each indicator is synchronized; the next step is to standardize the data. Due to the differences in the collection frequency of each sensor, the experimental indicator data Process by sampling frequency or time series. And because the sampling frequency of the sensor is high, the amount of data per second is large. In order to facilitate the analysis, one second is taken as the minimum timing unit, and the average value of the data per second is regarded as the data value of this second. For each fatigue data sample, when the minimum vehicle speed during overtaking is less than 80km/h, the sample is considered invalid. This is because traffic accidents caused by fatigue driving are common in free-flow conditions on expressways or urban expressways, and in crowded Traffic accidents due to driver fatigue rarely occur on urban roads. Therefore, in order to reduce computational complexity and improve the accuracy of evaluation results, samples with a minimum value less than 80 km/h are ignored in the embodiment of the present invention.

Then, the vehicle speed parameters are extracted to obtain the maximum value of the overtaking speed V _max : that is, the maximum value of the vehicle speed in an overtaking sample; the maximum value of the overtaking speed V _min : that is, the minimum value of the vehicle speed in an overtaking sample; to obtain the overtaking speed range V _range : that is The difference between the maximum vehicle speed and the minimum vehicle speed in an overtaking sample: V _range =V _max -V _min ; obtain the average overtaking speed V _mean : that is, the average value of vehicle speeds in an overtaking sample.

Extract the lane position parameters, and calculate the standard deviation of lane departure (SDLP) from the lane line spacing:

where d _avg is the mean value of the lane positions in the sampling period, which is 1.8m in the example of the present invention; d _i is the lane position value in the sampling period, and the left lane line spacing value is used in the calculation in the example of the present invention; n is the lane in the analysis sampling period Number of location samples.

Extract the overtaking duration parameter, and obtain the overtaking behavior start time T _start : that is, the time when overtaking starts in an overtaking sample; obtain the overtaking behavior end time T _end : that is, the time when overtaking ends in an overtaking sample; the time used to obtain the overtaking behavior T: that is, a The difference between the overtaking end time and the start time in the overtaking sample: T=T _end -T _star . The method for judging the end of the overtaking behavior is as follows: when the right turn signal of the vehicle is turned off, it is determined that the overtaking behavior ends.

Step 4: Determine the fatigue state.

According to the vehicle speed parameters, lane position parameters, and overtaking time parameters extracted from the fatigue data in step 3 as the fatigue discrimination index set, the classification is performed based on the SVM classifier that has been trained with the fatigue driving experimental data performed in advance, and the driver's fatigue state is determined. .

The classifier used in the present invention is the SVM classifier, and the driver's fatigue state is divided into three levels, namely, awake, fatigued, and very fatigued. The SVM classifier has been trained using the data of the fatigue driving experiment performed in advance, and the fatigue characteristics of a valid fatigue data sample are input. In the present invention, these fatigue characteristics include the indicators related to the overtaking behavior described in step 3. The output is one of three fatigue levels.

It should be understood that, for those skilled in the art, improvements or changes can be made according to the above description, and all these improvements and changes should fall within the protection scope of the appended claims of the present invention.

Claims (3)

1. A fatigue driving monitoring method based on overtaking behavior characteristics is characterized by comprising the following steps:

1) judging whether the driver overtakes or not;

2) acquiring overtaking data through a sensor, the overtaking data comprising: vehicle speed record data of the start of the moment of the overtaking behavior, lane position information of the start of the moment of the overtaking behavior, the start moment of the overtaking behavior and the end moment of the overtaking behavior;

3) obtaining fatigue characteristics of the overtaking process of the vehicle according to the acquired overtaking data, wherein the fatigue characteristics comprise: the system comprises a vehicle speed parameter, a lane position parameter and an overtaking duration parameter;

the vehicle speed parameter is obtained by the following method:

obtaining the maximum value V of overtaking speed_max: the maximum value of the vehicle speed in one overtaking sample;

obtaining the minimum value V of overtaking speed_min: namely the minimum value of the vehicle speed in one overtaking sample;

acquiring the passing speed range V_range: i.e. the difference between the maximum and minimum vehicle speed values in an overtaking sample: v_range＝V_max-V_min；

Obtaining the average overtaking speed V_mean: i.e. speed in one overtaking sampleAverage value of (d);

the lane position parameter is obtained by the following method:

acquiring a lane departure standard deviation SDLP, wherein the lane departure standard deviation describes the current operation state of a driver;

wherein d is_avgThe average value of the lane positions in the sampling period is obtained; d_iThe position value of the lane in the sampling period is obtained; n is the number of lane position samples in the analysis sampling period; the sampling period is the time period from the beginning to the end of the one-time overtaking behavior;

the overtaking duration parameter is obtained by the following method:

obtaining the starting time T of overtaking behavior_start: namely the overtaking starting time in an overtaking sample;

obtaining the overtaking behavior ending time T_end: namely the overtaking ending time in an overtaking sample;

acquiring the overtaking behavior duration T: i.e. the difference between the overtaking ending time and the overtaking starting time in one overtaking sample: t ═ T_end-T_star；

4) According to the vehicle speed parameter, the lane position parameter and the overtaking duration parameter extracted from the overtaking data in the step 3), as a fatigue judgment index set, classifying the overtaking duration parameter based on an SVM classifier trained by using fatigue driving experimental data which is performed in advance, and judging the fatigue state of a driver;

the fatigue state of the driver is divided into three levels, namely waking, fatigue and very fatigue, and the output of the classifier is one of the three fatigue levels.

2. The fatigue driving monitoring method based on the overtaking behavior feature as claimed in claim 1, wherein the step 1) is to determine whether the driver overtakes specifically as follows: whether the driver overtakes or not is judged according to the monitoring result by monitoring the steering lamp, the steering wheel corner and the vehicle acceleration; whether a left steering lamp is turned on or not is confirmed according to the monitoring result, whether the steering wheel angle is continuously larger than 6 degrees or not is judged, and whether the vehicle is in an acceleration state or not is judged; and when the three conditions are simultaneously met, judging that the overtaking action is performed.

3. The fatigue driving monitoring method based on overtaking behavior characteristics as claimed in claim 1, wherein the method for judging the overtaking behavior ending in step 2) is as follows: and judging that the overtaking behavior is ended when the right turn light of the vehicle is turned off.

Images