Driver field driving ability evaluation method based on big data

Abstract

The invention discloses a driver field driving ability evaluation method based on big data. The method comprises the following steps: carrying out multi-dimensional acquisition on driving behavior data of gold-brand coaches and trainees; removing abnormal values from the collected data, analyzing the driving behavior characteristic distribution of the coach and the trainee, and determining the coach data value range per meter so as to form a standard process curve; scoring according to a difference value between the driving behavior curve of the student and the standard process curve to obtain a behavior index score of the student; performing weighted calculation on the behavior index score to obtain a sub-project score; and averaging the scores of the sub-projects covered in the project to obtain the score of the project. According to the method, process-oriented evaluation indexes are adopted, and the defects that result-oriented evaluation indexes are prone to ignoring behavior details in the driver process and lack of real-time feedback are overcome.

Description

Technical Field

[0001] The present invention relates to the technical field of motor vehicle driver training test evaluation, and in particular to a method for evaluating a driver's on-site driving ability based on big data. Background Art

[0002] Most of the existing methods for evaluating drivers' on-site driving skills are based on sensors (radar, lidar, and cameras, etc.) installed on the test vehicle and high-precision positioning technology (such as GPS and inertial navigation systems) to perceive the information of the test vehicle. Sensors are used to perceive the distance, speed, and direction of surrounding objects, and high-precision positioning technology is used to determine the position and posture of the vehicle. Based on the information data obtained by the on-board sensors and high-precision positioning technology, the positional relationship between the vehicle position and the lane edge and the vehicle status information are inferred, and the operating behaviors such as the turn signal condition, the line crossing condition, the driving route, and the parking condition are judged. The result-oriented evaluation indicators such as "not using or using the turn signal incorrectly", "the vehicle body touches the storage position edge", "not driving on the prescribed route", and "not turning off the engine before turning off the engine" are used. Static judgments are made according to the established thresholds, and points are deducted for wrong or irregular behaviors, and finally the total score of the subject two items is obtained.

[0003] The above evaluation method is a result-oriented evaluation method, which mainly focuses on the final driving results, such as whether the vehicle crosses the line or violates the rules, etc. It ignores the driver's dynamic performance and decision-making ability during driving, and cannot make a quantitative evaluation of the dynamic and continuous driving process, and lacks a comprehensive evaluation of the driver's driving skills. Summary of the invention

[0004] In view of this, the purpose of the present invention is to provide a driver's on-site driving ability assessment method based on big data, which adopts process-oriented evaluation indicators to make up for the shortcomings of result-oriented evaluation indicators that easily ignore the driver's behavioral details during the process and lack real-time feedback.

[0005] The present invention solves the technical problem by adopting the following technical solution:

[0006] A method for evaluating a driver's on-site driving ability based on big data comprises the following steps:

[0007] S1, training behavior data

[0008] Collect driving behavior data of gold medal coaches and students from multiple dimensions;

[0009] S2, standard process curve

[0010] The collected data is filtered out of abnormal values, and the driving behavior time series data is converted into distance coordinate data with a granularity of 1 meter. The driving behavior data of the coaches and students are statistically analyzed respectively, and the driving behavior characteristic distribution of the coaches and students is analyzed to determine the value range of the coach data per meter, thereby forming a standard process curve;

[0011] S3, behavioral indicator score

[0012] Scoring is performed based on the difference between the student's driving behavior curve and the standard process curve to obtain the student's behavior index score, and the student's driving performance is evaluated and fed back based on the score;

[0013] S4, sub-item scores

[0014] The behavioral indicator scores were weighted to obtain sub-item scores;

[0015] S5, item score

[0016] The project score is obtained by averaging the scores of the sub-items covered in the project.

[0017] Furthermore, the driving behavior data includes running data and operation data, wherein the running data includes GPS information, engine speed, actual vehicle speed, body angle and vehicle gear information; the operation data includes steering wheel angle, brake depth, clutch depth and throttle depth.

[0018] Furthermore, the method of weighting the behavioral indicator scores to obtain the sub-item scores includes:

[0019] If the sub-item covers multiple driving behavior indicators, the sub-item score is the cumulative value of the product of the scores of each behavior indicator and the weight, where the weight is determined based on the standard deviation method; if the sub-item has only a single behavior indicator, the sub-item score is the score of the behavior indicator it contains.

[0020] Furthermore, in the on-site driving skills test training, it is divided into five items: reversing into a garage, parallel parking, right-angle turning, curve driving and ramp starting. The five items contain a total of 12 sub-items, and the 12 sub-items involve a total of 11 behavioral indicators.

[0021] Furthermore, the 11 behavioral indicators include lane alignment, steering timing, parking timing, parking alignment, vehicle speed control, turn signal status, trajectory control, parking position, starting stability, number of passive braking times, and number of bad driving behaviors.

[0022] The present invention discloses a method for evaluating a driver's on-site driving ability based on big data, which has the following features:

[0023] Beneficial effects:

[0024] 1) Establishing the optimal driver benchmark assessment method based on real vehicle big data

[0025] The on-site driving behavior evaluation method of the present invention is based on a large amount of driving behavior data from real vehicle training, collects multi-dimensional data from the driving operation process of coaches and students, and draws a standard process curve based on the driving behavior process data of gold medal coaches. The standard curve is established based on a large amount of real vehicle data and is closer to the actual traffic environment, so the evaluation method based on this is more accurate.

[0026] 2) Realize the process evaluation of the driver's driving skills

[0027] The present invention constructs a process indicator system and develops a process-oriented driving behavior evaluation method, abandoning the original result-oriented driving behavior evaluation method, and scoring the driver's driving operation behavior in meters. This evaluation method evaluates the driver's driving skill operation level more comprehensively and accurately, and reduces the impact of accidental factors on the evaluation of the driver's driving skill operation level. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 is a flow chart of the method of the present invention;

[0029] Figure 2 It is the trajectory diagram of the coach and the students of the present invention;

[0030] Figure 3 This is a standard curve diagram of the steering wheel angle of the present invention. DETAILED DESCRIPTION

[0031] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0032] refer to Figure 1 The present invention discloses a method for evaluating a driver's on-site driving ability based on big data, comprising the following steps:

[0033] S1, training behavior data

[0034] The driving behavior data of gold medal coaches and students are collected from multiple dimensions; the driving behavior data includes running data and operation data, among which the running data includes GPS information, engine speed, actual vehicle speed, body angle and vehicle gear information; the operation data includes steering wheel angle, brake depth, clutch depth and throttle depth.

[0035] S2, standard process curve

[0036] The collected data is filtered out of abnormal values, and the driving behavior time series data is converted into distance coordinate data with a granularity of 1 meter. The driving behavior data of the coaches and students are statistically analyzed respectively, and the driving behavior characteristic distribution of the coaches and students is analyzed to determine the value range of the coach data per meter, thereby forming a standard process curve;

[0037] S3, behavioral indicator score

[0038] Scoring is performed based on the difference between the student's driving behavior curve and the standard process curve to obtain the student's behavior index score, and the student's driving performance is evaluated and fed back based on the score;

[0039] S4, sub-item scores

[0040] The sub-item score is obtained by weighted calculation of the behavioral indicator scores; if the sub-item covers multiple driving behavior indicators, the sub-item score is the cumulative value of the product of the scores of each behavioral indicator and the weight, where the weight is determined according to the standard deviation method; if the sub-item has only a single behavioral indicator, the sub-item score is the score of the behavioral indicator it contains.

[0041] S5, item score

[0042] The project score is obtained by averaging the scores of the sub-items covered in the project.

[0043] The present invention collects driving behavior data of multiple coaches and trainees, converts the driving behavior data into distance coordinate data, extracts and analyzes the driving behavior characteristics of the coaches and trainees respectively, and develops an evaluation method based on a large amount of real vehicle data.

[0044] To further optimize the technical solution, the on-site driving skills test training is divided into five items: reverse parking, side parking, right-angle turning, curve driving and ramp starting. The five items contain a total of 12 sub-items, and the 12 sub-items involve a total of 11 behavioral indicators. The 11 behavioral indicators include lane alignment, wheel turning timing, parking timing, parking space alignment, vehicle speed control, turn signal status, trajectory control, parking position, starting stability, passive braking times and bad driving behavior times. Each evaluation indicator is scored according to the driving behavior operation, such as steering wheel angle, point to storage distance, brake pedal depth, driving direction and storage angle, clutch pedal opening, turn signal status, throttle depth and error score and other operational behavior indicators. The specific evaluation indicators covered by each sub-item are shown in Appendix 1-5. Among them, Table 1 shows the evaluation index system for reverse parking, Table 2 shows the evaluation index system for side parking, Table 3 shows the evaluation index system for right-angle turning, Table 4 shows the evaluation index system for curve driving, and Table 5 shows the evaluation index system for ramp starting.

[0045]

[0046]

[0047] Table 1

[0048]

[0049] Table 2

[0050]

[0051] Table 3

[0052]

[0053] Table 4

[0054]

[0055]

[0056] Table 5

[0057] The present invention constructs a set of on-site driving skill evaluation index system, and develops a driver on-site driving behavior evaluation method based on big data based on the constructed index system. The process-oriented evaluation index makes up for the shortcomings of result-oriented evaluation indexes that easily ignore the details of the driver's behavior during the process and lack of real-time feedback. At the same time, the evaluation method is developed based on a large amount of driving behavior data of coaches and trainees collected from real vehicles. It is driven by big data and forms a standard person curve based on the driving behavior data of gold medal coaches, thereby realizing the full process evaluation of trainees' driving behavior. The data collected from the real vehicle comes directly from the real driving environment, so that the evaluation method developed based on real vehicle data is more accurate and close to reality.

[0058] Application Examples

[0059] Take the lane alignment evaluation index in the reverse parking-entry project as an example. The data set is divided into meters, and the data values ​​of 19m to 110m are excluded. The data set range is 0-19m. The distance from the vehicle to the parking edge and the steering wheel angle are the main evaluation indicators in the entry project process, as follows:

[0060] 1) Scoring the distance from the vehicle to the storage location

[0061] Determine the entire storage location boundary range based on the four points of the storage location coordinates, and calculate the distance between each track point of the trainee and the storage location boundary. Group them in units of 1 meter, and extract the data in each group that is 3 meters away from the storage location boundary. Figure 2 As shown, through Figure 2The coach’s trajectory range is used to determine the interval range of each score, and scores are given based on the distance range of the calculated distance mean.

[0062] √The average distance per meter is between 4.03-4.20, excellent, 100

[0063] √Except for excellent, the range is between 3.80-4.30, good, 70

[0064] √Except for excellent, the range is 3.50-4.90, medium, 50

[0065] √Except for excellent, the range is in other ranges, poor, 0

[0066] 2) Steering wheel angle score

[0067] The process evaluation is achieved by judging based on the similarity with the standard person curve. The higher the similarity between the student's driving behavior process curve and the standard person curve, the higher the score. The threshold range of the standard person curve for the same evaluation indicator in different projects is slightly different. The scoring criteria for the steering wheel angle of the reverse parking-entry project are as follows.

[0068] The steering wheel angle score is based on the difference between the average steering wheel angle per meter and the instructor's standard curve. The steering wheel angle standard curve is shown in the figure below. Figure 3 shown.

[0069] √The turning angle is within the standard curve range, excellent, 100

[0070] √The angle is 10 degrees different from the standard curve, good, 70

[0071] √The angle is 20 degrees different from the standard curve, medium, 50

[0072] √The angle differs from the standard curve by more than 20 degrees, difference, 0

[0073] After calculating the score for each meter, the average is the steering wheel angle score.

[0074] 3) Lane alignment indicator score

[0075] The index score is obtained by the product and sum of the driving operation behavior score and its weight.

[0076] Lane alignment score = 0.542*vehicle-to-storage-location distance score + 0.458*steering wheel angle score.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for evaluating a driver's on-site driving ability based on big data, characterized in that: The steps include: S1, training behavior data Collect driving behavior data of gold medal coaches and students from multiple dimensions; S2, standard process curve The collected data is filtered out of abnormal values, and the driving behavior time series data is converted into distance coordinate data with a granularity of 1 meter. The driving behavior data of the coaches and students are statistically analyzed respectively, and the driving behavior characteristic distribution of the coaches and students is analyzed to determine the value range of the coach data per meter, thereby forming a standard process curve; S3, behavioral indicator score Scoring is performed based on the difference between the student's driving behavior curve and the standard process curve to obtain the student's behavior index score, and the student's driving performance is evaluated and fed back based on the score; S4, sub-item scores The behavioral indicator scores were weighted to obtain sub-item scores; S5, item score The project score is obtained by averaging the scores of the sub-items covered in the project.

2. The method for evaluating a driver's on-site driving ability based on big data according to claim 1 is characterized in that: Driving behavior data includes running data and operation data. The running data includes GPS information, engine speed, actual vehicle speed, body angle and vehicle gear information; the operation data includes steering wheel angle, brake depth, clutch depth and throttle depth.

3. The method for evaluating a driver's on-site driving ability based on big data according to claim 2 is characterized in that: The method of weighting the behavioral indicator scores to obtain the sub-item scores includes: If the sub-item covers multiple driving behavior indicators, the sub-item score is the cumulative value of the product of the scores of each behavior indicator and the weight, where the weight is determined based on the standard deviation method; if the sub-item has only a single behavior indicator, the sub-item score is the score of the behavior indicator it contains.

4. The method for evaluating a driver's on-site driving ability based on big data according to claim 3 is characterized in that: In the on-site driving skills test training, there are 5 items: reversing into a garage, parallel parking, right-angle turning, curve driving and ramp starting. The 5 items contain a total of 12 sub-items, and the 12 sub-items involve a total of 11 behavioral indicators.

5. The method for evaluating a driver's on-site driving ability based on big data according to claim 4 is characterized in that: The 11 behavioral indicators include lane alignment, steering timing, parking timing, parking space alignment, vehicle speed control, turn signal status, trajectory control, parking position, starting stability, number of passive braking times and number of bad driving behaviors.

Images