CN114996116A - Anthropomorphic evaluation method for automatic driving system - Google Patents

Anthropomorphic evaluation method for automatic driving system Download PDF

Info

Publication number
CN114996116A
CN114996116A CN202210285074.9A CN202210285074A CN114996116A CN 114996116 A CN114996116 A CN 114996116A CN 202210285074 A CN202210285074 A CN 202210285074A CN 114996116 A CN114996116 A CN 114996116A
Authority
CN
China
Prior art keywords
driving
vehicle
data
sampling
automatic driving
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202210285074.9A
Other languages
Chinese (zh)
Inventor
朱冰
张培兴
赵健
范天昕
孙宇航
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jilin University
Original Assignee
Jilin University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Jilin University filed Critical Jilin University
Priority to CN202210285074.9A priority Critical patent/CN114996116A/en
Publication of CN114996116A publication Critical patent/CN114996116A/en
Priority to US17/943,167 priority patent/US20230304896A1/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/36—Preventing errors by testing or debugging software
    • G06F11/3668—Software testing
    • G06F11/3672—Test management
    • G06F11/3688—Test management for test execution, e.g. scheduling of test suites
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/36—Preventing errors by testing or debugging software
    • G06F11/3664—Environments for testing or debugging software
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/80—Technologies aiming to reduce greenhouse gasses emissions common to all road transportation technologies
    • Y02T10/84—Data processing systems or methods, management, administration

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Pure & Applied Mathematics (AREA)
  • Mathematical Optimization (AREA)
  • Quality & Reliability (AREA)
  • Computer Hardware Design (AREA)
  • Mathematical Analysis (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Operations Research (AREA)
  • Probability & Statistics with Applications (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Algebra (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Software Systems (AREA)
  • Traffic Control Systems (AREA)

Abstract

The invention belongs to the technical field of test and evaluation of an automatic driving automobile, and particularly relates to a anthropomorphic evaluation method for an automatic driving system. The method comprises the following steps: step one, acquiring running data of a tested automatic driving system; step two, collecting driving data of a driver; step three, calculating a driving track field; step four, calculating the probability distribution of the data of the real driving track field of the driver; step five, constructing a quasi-human index; and sixthly, sampling the data of the driving track field and obtaining a final evaluation result. The evaluation method comprises the steps of calculating a driving track field according to driving data of a detected automatic driving system, comparing and analyzing the track field and the driving track field of a real human driver, and calculating the driving similarity of the detected automatic driving system and the real human driver when the detected automatic driving system drives in the scene in a probability statistics mode.

Description

Anthropomorphic evaluation method for automatic driving system
Technical Field
The invention belongs to the technical field of test and evaluation of an automatic driving automobile, and particularly relates to a anthropomorphic evaluation method for an automatic driving system.
Background
With the continuous improvement of the automatic driving technology, more and more attention is paid to how to evaluate the anthropomorphic property of the automatic driving system in the driving process. The anthropomorphic nature of the driving process of the automatic driving system mainly affects two aspects: for the passengers, the psychological acceptance degree of the passengers can be influenced by the difference of the driving behaviors of the automatic driving system, so that the passengers cannot trust the automatic driving system; for other traffic participants, the low-intelligence autopilot system cannot correctly interact with the other traffic participants, so that the driving experience of other drivers is influenced, the traffic efficiency is influenced, and even traffic accidents are caused. Therefore, how to correctly evaluate the anthropomorphic property of the driving process of the automatic driving system has important significance.
Most of the existing methods are used for evaluating the similarity at a certain moment, the similarity evaluation of the whole driving process is lacked, the difference of the driving processes of different drivers is lacked to be considered, and only a single behavior is used as the human driving or ideal driving data.
Disclosure of Invention
The invention provides a personification evaluation method for an automatic driving system, which is characterized in that a driving track field is calculated according to driving data of a detected automatic driving system, the track field and the driving track field of a real human driver are compared and analyzed, and the similarity degree of the detected automatic driving system driving in the scene and the driving of the real human driver is calculated in a probability statistics mode, so that the problems that the existing method is mainly used for evaluating the similarity at a certain moment, the similarity evaluation of the whole driving process is lacked, and the difference of the driving processes of different drivers is lacked to be considered are solved.
The technical scheme of the invention is explained by combining the drawings as follows:
an anthropomorphic evaluation method for an automatic driving system comprises the following steps:
step one, acquiring running data of a tested automatic driving system;
step two, collecting driving data of a driver;
step three, calculating a driving track field;
step four, calculating the probability distribution of the data of the real driving track field of the driver;
step five, constructing a quasi-human index;
and sixthly, sampling the data of the driving track field to obtain a final evaluation result.
The specific method of the first step is as follows:
building a simulation test platform and building a corresponding test scene in a simulation environment; and embedding the tested automatic driving system into a simulation environment for testing, and acquiring the driving data of the tested automatic driving system in a simulation test scene, wherein the driving data comprises speed data, position data and operation data.
The specific method of the second step is as follows:
and acquiring running data including speed data, position data and operation data in a scene corresponding to the simulation test scene of the real human driver in the first step through natural driving data, real vehicle driving and a driving simulator.
The concrete method of the third step is as follows:
calculating a driving track field of the two groups of data in a test scene through formulas (1) to (2) according to the driving data of the tested algorithm obtained in the step one and the driving data of the real human driver acquired in the step two;
regarding the whole driving scene as a coordinate system, regarding the vehicle as an x axis along the advancing direction of a road, regarding the lane direction as a y axis, and regarding different positions on the lane as specific coordinates of the coordinate system;
S＝∑s (1)
Figure BDA0003557892420000021
in the formula, S is a driving track field, namely the sum of the influences of the whole driving process of the vehicle on the surrounding space; s is an instantaneous field, namely, the influence of the driving moment of the vehicle on the surrounding space and time; r is ij In different positions anda vector consisting of vehicle centers; v. of i Is the speed of the vehicle; theta i Is r ij And v i The included angle of (A); k is a radical of 1 And k 2 To correct the parameters;
regarding the vehicle as a mass point, and regarding the instantaneous field value of the vehicle in the advancing direction, which is 1m away from the center of mass of the vehicle, as the instantaneous field value of the whole vehicle in 1 m;
and calculating the instantaneous field value of the vehicle every time the vehicle advances by 0.5m until the vehicle stops or leaves the driving scene, and adding all the obtained instantaneous fields to obtain the driving track field value in the whole scene, wherein the formula (1) is shown.
The concrete method of the fourth step is as follows:
due to differences of the driving data of different drivers, the driving track fields in corresponding scenes obtained by calculation according to the driving data of different real drivers have differences, and values of the driving track fields of the real drivers at different positions in the whole scene are described by Gaussian distribution, namely, the values of the driving track fields at different positions are described by Gaussian distribution, as shown in (3);
Figure BDA0003557892420000031
in the formula, h is a specific numerical value of a driving track field at different positions in a coordinate system; μ and σ are the mean and standard deviation of the values of the travel trajectory field at the corresponding locations.
The concrete method of the step five is as follows:
the quasi-human index comprises four parts of contents, namely: an operation number correction factor; a second part: increasing a driving mileage correction factor; and a third part: similarity of travel trajectories; the fourth part: similarity of driving speeds at corresponding positions; according to the four indexes, the set anthropomorphic index is shown as a formula (4), and the maximum value of the anthropomorphic index is 1;
Figure BDA0003557892420000032
wherein, L is the running distance of the detected automatic driving system in a specific scene, when a vehicle controlled by the detected automatic driving system stops, L is the distance between a starting point and a terminating point along the road direction, and if the vehicle does not stop, L is the length along the road direction in the scene; l is mean The average value of the driving distance of the real human drivers in the corresponding scene is obtained, and the driving distance of each real human driver is obtained in the same mode as L; n is h The number of times of operation of the tested automatic driving system on the vehicle is determined, and the absolute value of the acceleration of the specified braking or acceleration is more than 0.5m/s 2 The back brake pedal or the accelerator pedal returns to the initial position to be one-time vehicle operation, and the back reverse direction returns to the initial position from the back to be one-time vehicle operation after the steering wheel angle of the vehicle is larger than 10 degrees; n is A Is the average operation times of the real driver, the acquisition mode of the operation times of a single real driver and n h The same; r is the position of road sampling, the length along the road between each sampling is defined to be 0.5m, the sampling position is a line segment, namely a line segment taking a lane boundary as a boundary in a coordinate system when x is equal to r, and when the distance between the last sampling position and the road end point is less than 0.5m, the road end point position of the line segment is directly sampled without considering the interval of 0.5 m; p is a radical of t_r The method comprises the steps that a Gaussian distribution is formed by using vehicle positions of a real driver during driving on a sampling line segment, the probability of a value of the vehicle position of a detected automatic driving system in the Gaussian distribution is described, and the probability distribution of different y values at the sampling position of the real driver during driving is described as shown in a formula (3); where h becomes the value of the vehicle position on the y-axis at the sampling position for a different real driver; μ and σ become the mean and standard deviation corresponding to the y values; p is a radical of t_2σ_r The probability that the vehicle position belongs to the position of the mean value of the y-axis position of the real driver plus two times of the standard deviation is obtained; p is a radical of v_j Calculating the probability that the running track field at different sampling positions on the sampling line segment is the specific numerical value when a real driver drives by using a formula (3); p is a radical of v_2σ_j The driving track field of different sampling positions on the sampling line segment when a real driver drives is the mean value of the sampling position plus two times of the standard deviation of the sampling positionThe corresponding probability at the location; n is s The number of sampling points on a sampling line segment is 9; n is m Is the number of all sample points; k is a radical of 3 、k 4 For correcting the coefficient, when the operation frequency of the detected automatic driving system is less than that of a real driver and the driving distance is longer than that of the real driver, k is used 3 、k 4 Correcting the maximum value of D to 1, correcting the results of other tested automatic driving systems which are already involved in test evaluation, and otherwise, correcting k 3 、k 4 Directly taking 1.
The concrete method of the sixth step is as follows:
sampling road data forwards at intervals of 0.5m in the road direction until the end of the road; when a sampling point is selected in a sampling line segment, 4 points are sampled by taking the vehicle position as the center and taking 0.5m as a step length upwards and downwards, 9 points are sampled in a whole slice, when the distance between the vehicle position and the road boundary on one sampling line segment is less than 2m, the length between the vehicle position and the road boundary is divided into four equal parts, and data are sampled at the four equal parts without considering the interval of 0.5 m; and (3) substituting the sampled data result into a formula (4) to calculate the similarity between the tested automatic driving system and the driving of a real driver so as to obtain the anthropomorphic evaluation result of the tested automatic driving system, and if the evaluation result is greater than 0.85, the tested algorithm is considered to have better anthropomorphic performance.
The beneficial effects of the invention are as follows:
according to the driving track field of the tested automatic driving system and the driving track field of the real human driver in the corresponding scene, the similarity of the driving track field and the driving track field is compared and analyzed to obtain the anthropomorphic property of the tested automatic driving system. The method provided by the invention can evaluate the anthropomorphic property of the whole driving process of the whole driving scene, fully considers the difference of the driving process of human drivers, and has strong universality.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained according to the drawings without inventive efforts.
FIG. 1 is an overall flow chart of the evaluation;
fig. 2 is a schematic view of a driving scenario.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, a method for evaluating personification of an automatic driving system includes the steps of:
step one, acquiring running data of a tested automatic driving system;
the specific method comprises the following steps:
building a simulation test platform and building a corresponding test scene in a simulation environment; and embedding the tested automatic driving system into a simulation environment for testing, and acquiring the driving data of the tested automatic driving system in a simulation test scene, wherein the driving data comprises speed data, position data and operation data.
Step two, collecting driving data of a driver;
the specific method comprises the following steps:
and acquiring driving data including speed data, position data and operation data in a scene corresponding to a simulation test scene of a real human driver in the first step through natural driving data, real vehicle driving and a driving simulator.
Step three, calculating a driving track field;
the specific method comprises the following steps:
calculating a driving track field of the two groups of data in a test scene through formulas (1) to (2) according to the driving data of the tested algorithm obtained in the step one and the driving data of the real human driver acquired in the step two;
regarding the whole driving scene as a coordinate system, regarding the vehicle as an x axis along the road advancing direction, regarding the lane direction as a y axis, and regarding different positions on the lane as specific coordinates of the coordinate system;
S＝∑s (1)
Figure BDA0003557892420000061
in the formula, S is a driving track field, namely the sum of the influences of the whole driving process of the vehicle on the surrounding space; s is an instantaneous field, namely, the influence of the driving moment of the vehicle on the surrounding space and time; r is ij Vectors composed of different positions and the center of the vehicle; v. of i Is the speed of the vehicle; theta.theta. i Is r ij And v i The included angle of (A); k is a radical of 1 And k 2 To correct the parameters;
regarding the vehicle as a mass point, and regarding the instantaneous field value of the vehicle in the advancing direction, which is 1m away from the center of mass of the vehicle, as the instantaneous field value of the whole vehicle in 1 m;
and calculating the instantaneous field value of the vehicle every time the vehicle advances by 0.5m until the vehicle stops or leaves the driving scene, and adding all the obtained instantaneous fields to obtain the driving track field value in the whole scene, wherein the formula (1) is shown.
Step four, calculating the probability distribution of the data of the real driving track field of the driver;
the specific method comprises the following steps:
due to differences of the running data of different drivers, running track fields in corresponding scenes obtained by calculation according to the running data of different real drivers have differences, and the values of the running track fields of the real drivers at different positions in the whole scene are described by using Gaussian distribution, namely the values of the running track fields at different positions are described by using Gaussian distribution, as shown in (3);
Figure BDA0003557892420000071
in the formula, h is a specific numerical value of a driving track field at different positions in a coordinate system; μ and σ are the mean and standard deviation of the values of the travel trajectory field at the corresponding locations.
Step five, constructing a quasi-human index;
the specific method comprises the following steps:
the quasi-human index comprises four parts of contents, namely: an operation number correction factor; a second part: increasing a driving mileage correction factor; and a third part: similarity of travel trajectories; the fourth part: similarity of driving speeds at corresponding positions; according to the four indexes, the set anthropomorphic index is shown as a formula (4), and the maximum value of the anthropomorphic index is 1;
Figure BDA0003557892420000072
wherein, L is the running distance of the measured automatic driving system in a specific scene, when a vehicle controlled by the measured automatic driving system stops, L is the distance between a starting point and an end point along the road direction, and if the vehicle does not stop, the L is the length along the road direction in the scene; l is mean The average value of the driving distance of the real human drivers in the corresponding scene is obtained, and the driving distance of each real human driver is obtained in the same mode as L; n is h The number of times of operation of the tested automatic driving system on the vehicle is determined, and the absolute value of the acceleration of the specified braking or acceleration is more than 0.5m/s 2 The back brake pedal or the accelerator pedal returns to the initial position to be a vehicle operation, and the back reverse direction returns to the initial position after the steering wheel angle of the vehicle is larger than 10 degrees to be a vehicle operation; n is A Is the average operation times of the real driver, the acquisition mode of the operation times of a single real driver and n h The same; r is the position of road sampling, the length along the road between each sampling is defined to be 0.5m, the sampling position is a line segment, namely a line segment taking the lane boundary as the boundary in the coordinate system when x is equal to r, when the distance between the penultimate sampling position and the road end point is less than 0.5m, the line segment road end point position is directly sampled, and the 0.5m road end point position is not required to be consideredSpacing; p is a radical of t_r The method comprises the steps that a Gaussian distribution is formed by using vehicle positions of a real driver during driving on a sampling line segment, the probability of the value of the vehicle position of a detected automatic driving system in the Gaussian distribution is described as shown in a formula (3) in the probability distribution of different y values at the sampling position during driving of the real driver; where h becomes the value of the vehicle position on the y-axis at the sampling position for a different real driver; μ and σ become mean and standard deviation corresponding to the y value; p is a radical of t_2σ_r The probability that the vehicle position belongs to the mean value of the real driver y-axis position plus twice the standard deviation is obtained; p is a radical of formula v_j Calculating the probability that the running track field at different sampling positions on the sampling line segment is the specific numerical value when a real driver drives by using a formula (3); p is a radical of v_2σ_j The probability of the driving track field at different sampling positions on the sampling line segment is the corresponding probability of the position of the mean value at the position of the sampling point plus the position of the standard deviation which is twice of the position of the sampling point when a real driver drives; n is s The number of sampling points on a sampling line segment is 9; n is m Is the number of all sampling points; k is a radical of formula 3 、k 4 To correct the coefficients, k is used when the number of measured autopilot system operations is less than the actual driver and the mileage is longer than the actual driver 3 、k 4 Correcting the maximum value of D to 1, correcting the results of other tested automatic driving systems which are already involved in test evaluation, and otherwise, correcting k 3 、k 4 Directly taking 1.
And sixthly, sampling the data of the driving track field to obtain a final evaluation result.
The specific method comprises the following steps:
sampling road data forwards at intervals of 0.5m in the road direction until the end of the road; when a sampling point is selected on a sampling line segment, taking the position of a vehicle as the center, taking 0.5m as a step length to sample 4 points upwards and downwards, sampling 9 points on the whole slice, when the distance between the position of the vehicle and the road boundary on the sampling line segment is less than 2m, quartering the length between the position of the vehicle and the road boundary, and sampling data on the quartering points without considering the interval of 0.5 m; and substituting the sampled data result into a formula (4) to calculate the similarity between the tested automatic driving system and the driving of a real driver so as to obtain the anthropomorphic evaluation result of the tested automatic driving system, and if the evaluation result is greater than 0.85, considering that the tested algorithm has better human simulation.
Examples
Referring to fig. 2, a preceding vehicle low speed scenario is selected as an example of test evaluation.
The front vehicle runs at a low speed of 15 m/s, the vehicle runs forwards at a constant speed of 25m/s, the distance from the front vehicle to the front vehicle is 150m initially, and the similarity between the measured automatic driving system and the real driver in the running process is judged. And acquiring vehicle running data of 10 real drivers in the scene and running tracks of two tested automatic driving systems in the scene. The first tested automatic driving system adopts the operation of changing the lane to the left after approaching the front vehicle in the driving process, the second tested automatic driving system adopts the operation of decelerating and following the front vehicle in the driving process, and 10 real drivers all adopt the operation of changing the lane to the left. And inputting the running data of 10 real drivers and two tested automatic driving systems into the formulas (1) to (4), wherein the calculated result of the anthropomorphic effect of the first tested automatic driving system is 0.96, and the result of the anthropomorphic effect of the second automatic driving system is 0.41, and is consistent with the subjective condition of vehicle running.
Although the preferred embodiments of the present invention have been described in detail with reference to the accompanying drawings, the scope of the present invention is not limited to the specific details of the above embodiments, and any person skilled in the art can substitute or change the technical solution of the present invention and its inventive concept within the technical scope of the present invention, and these simple modifications belong to the scope of the present invention.
It should be noted that the various technical features described in the above embodiments can be combined in any suitable manner without contradiction, and the invention is not described in any way for the possible combinations in order to avoid unnecessary repetition.
In addition, any combination of the various embodiments of the present invention is also possible, and the same should be considered as the disclosure of the present invention as long as it does not depart from the spirit of the present invention.

Claims (7)

1. An anthropomorphic evaluation method for an automatic driving system is characterized by comprising the following steps:
step one, acquiring running data of a tested automatic driving system;
step two, collecting driving data of a driver;
step three, calculating a driving track field;
step four, calculating the probability distribution of the data of the real driver driving track field;
step five, constructing a quasi-human index;
and sixthly, sampling the data of the driving track field to obtain a final evaluation result.
2. The anthropomorphic evaluation method for the automatic driving system according to claim 1 is characterized in that the specific method of the first step is as follows:
building a simulation test platform and building a corresponding test scene in a simulation environment; and embedding the tested automatic driving system into a simulation environment for testing, and acquiring the driving data of the tested automatic driving system in a simulation test scene, wherein the driving data comprises speed data, position data and operation data.
3. The anthropomorphic evaluation method for the automatic driving system according to claim 1 is characterized in that the specific method of the second step is as follows:
and acquiring driving data including speed data, position data and operation data in a scene corresponding to a simulation test scene of a real human driver in the first step through natural driving data, real vehicle driving and a driving simulator.
4. The anthropomorphic evaluation method for the automatic driving system according to claim 1 is characterized in that the concrete method of the third step is as follows:
calculating a driving track field of the two groups of data in a test scene through formulas (1) to (2) according to the driving data of the tested algorithm obtained in the step one and the driving data of the real human driver acquired in the step two;
regarding the whole driving scene as a coordinate system, regarding the vehicle as an x axis along the road advancing direction, regarding the lane direction as a y axis, and regarding different positions on the lane as specific coordinates of the coordinate system;
S＝∑s (1)
Figure FDA0003557892410000021
in the formula, S is a driving track field, namely the sum of the influences of the whole driving process of the vehicle on the surrounding space; s is an instantaneous field, namely, the influence of the driving moment of the vehicle on the surrounding space and time; r is ij Vectors composed of different positions and the center of the vehicle; v. of i Is the speed of the vehicle; theta i Is r ij And v i The included angle of (c); k is a radical of formula 1 And k 2 To correct the parameters;
regarding the vehicle as a mass point, and regarding the instantaneous field value of the vehicle in the advancing direction, which is 1m away from the center of mass of the vehicle, as the instantaneous field value of the whole vehicle in 1 m;
and (3) calculating the instantaneous field value of the vehicle every time the vehicle moves forward by 0.5m until the vehicle stops or moves out of the driving scene, and adding all the obtained instantaneous fields to obtain the driving track field value in the whole scene, as shown in a formula (1).
5. The anthropomorphic evaluation method for the automatic driving system according to claim 1 is characterized in that the concrete method of the step four is as follows:
due to differences of the running data of different drivers, running track fields in corresponding scenes obtained by calculation according to the running data of different real drivers have differences, and the values of the running track fields of the real drivers at different positions in the whole scene are described by Gaussian distribution, namely the values of the running track fields at different positions are described by Gaussian distribution, as shown in a formula (3);
Figure FDA0003557892410000022
in the formula, h is a specific numerical value of a driving track field at different positions in a coordinate system; μ and σ are the mean and standard deviation of the values of the travel trajectory field at the corresponding locations.
6. The anthropomorphic evaluation method for the automatic driving system according to claim 5 is characterized in that the concrete method of the fifth step is as follows:
the pseudo-human index comprises four parts of contents, namely: an operation number correction factor; a second part: increasing a driving mileage correction factor; and a third part: similarity of travel trajectories; the fourth part: similarity of driving speeds at corresponding positions; according to the four indexes, the established anthropomorphic index is shown as a formula (4), and the maximum value of the anthropomorphic index is 1;
Figure FDA0003557892410000031
wherein, L is the running distance of the measured automatic driving system in a specific scene, when a vehicle controlled by the measured automatic driving system stops, L is the distance between a starting point and an end point along the road direction, and if the vehicle does not stop, the L is the length along the road direction in the scene; l is mean The driving distance of each real human driver is obtained in the same mode as L; n is h The number of times of operation of the tested automatic driving system on the vehicle is determined, and the absolute value of the acceleration of the specified braking or acceleration is more than 0.5m/s 2 The back brake pedal or the accelerator pedal returns to the initial position to be one-time vehicle operation, and the back reverse direction returns to the initial position from the back to be one-time vehicle operation after the steering wheel angle of the vehicle is larger than 10 degrees; n is A Is a real driverAverage operation times of driving, obtaining mode of operation times of single real driver and n h The same; r is the position of road sampling, the length along the road between each sampling is defined to be 0.5m, the sampling position is a line segment, namely a line segment taking a lane boundary as a boundary in a coordinate system when x is equal to r, and when the distance between the last sampling position and the road end point is less than 0.5m, the road end point position of the line segment is directly sampled without considering the interval of 0.5 m; p is a radical of t_r The method comprises the steps that a Gaussian distribution is formed by using vehicle positions of a real driver during driving on a sampling line segment, the probability of the value of the vehicle position of a detected automatic driving system in the Gaussian distribution is described as shown in a formula (3) in the probability distribution of different y values at the sampling position during driving of the real driver; where h becomes the value of the vehicle position on the y-axis at the sampling position for a different real driver; μ and σ become mean and standard deviation corresponding to the y value; p is a radical of t_2σ_r The probability that the vehicle position belongs to the position of the mean value of the y-axis position of the real driver plus two times of the standard deviation is obtained; p is a radical of formula v_j Calculating the probability that the driving track field at different sampling positions on the sampling line segment is the specific numerical value when a real driver drives by using a formula (3); p is a radical of v_2σ_j The probability of the driving track field at different sampling positions on the sampling line segment is the corresponding probability of the position of the mean value at the position of the sampling point plus the position of the standard deviation which is twice of the position of the sampling point when a real driver drives; n is s The number of sampling points on a sampling line segment is 9; n is m Is the number of all sampling points; k is a radical of 3 、k 4 To correct the coefficients, k is used when the number of measured autopilot system operations is less than the actual driver and the mileage is longer than the actual driver 3 、k 4 Correcting the maximum value of D to 1, correcting the results of other tested automatic driving systems which are already involved in test evaluation, and otherwise, correcting k 3 、k 4 Directly taking 1.
7. The anthropomorphic evaluation method for the automatic driving system according to claim 6 is characterized in that the concrete method of the sixth step is as follows:
sampling road data forwards at intervals of 0.5m in the road direction until the end of the road is reached; when a sampling point is selected in a sampling line segment, 4 points are sampled by taking the vehicle position as the center and taking 0.5m as a step length upwards and downwards, 9 points are sampled in a whole slice, when the distance between the vehicle position and the road boundary on one sampling line segment is less than 2m, the length between the vehicle position and the road boundary is divided into four equal parts, and data are sampled at the four equal parts without considering the interval of 0.5 m; and substituting the sampled data result into a formula (4) to calculate the similarity between the tested automatic driving system and the driving of a real driver so as to obtain the anthropomorphic evaluation result of the tested automatic driving system, and if the evaluation result is greater than 0.85, considering that the tested algorithm has better human simulation.
CN202210285074.9A 2022-03-22 2022-03-22 Anthropomorphic evaluation method for automatic driving system Pending CN114996116A (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN202210285074.9A CN114996116A (en) 2022-03-22 2022-03-22 Anthropomorphic evaluation method for automatic driving system
US17/943,167 US20230304896A1 (en) 2022-03-22 2022-09-12 Method for evaluating performance of self-driving vehicle oriented to full parameter space of logical scenario

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210285074.9A CN114996116A (en) 2022-03-22 2022-03-22 Anthropomorphic evaluation method for automatic driving system

Publications (1)

Publication Number Publication Date
CN114996116A true CN114996116A (en) 2022-09-02

Family

ID=83023777

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210285074.9A Pending CN114996116A (en) 2022-03-22 2022-03-22 Anthropomorphic evaluation method for automatic driving system

Country Status (1)

Country Link
CN (1) CN114996116A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117217422A (en) * 2023-11-07 2023-12-12 国汽(北京)智能网联汽车研究院有限公司 Vehicle motion control capability assessment method, system, device and medium thereof

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117217422A (en) * 2023-11-07 2023-12-12 国汽(北京)智能网联汽车研究院有限公司 Vehicle motion control capability assessment method, system, device and medium thereof
CN117217422B (en) * 2023-11-07 2024-03-22 国汽(北京)智能网联汽车研究院有限公司 Vehicle motion control capability assessment method, system, device and medium thereof

Similar Documents

Publication Publication Date Title
CN112668153B (en) Method, device and equipment for generating automatic driving simulation scene
CN111009153B (en) Training method, device and equipment of trajectory prediction model
CN109657355A (en) A kind of emulation mode and system of road vehicle virtual scene
CN108820039A (en) A kind of automatic driving vehicle bend crosswise joint system and method
CN110276952B (en) Traffic information simulation acquisition method and device
CN114454878B (en) Method and device for determining vehicle speed control model training sample
CN113127466B (en) Vehicle track data preprocessing method and computer storage medium
CN106956680A (en) A kind of electric automobile driving behavior identifying and analyzing method
CN106960189A (en) A kind of driving intention decision method based on hidden Markov model
CN114996116A (en) Anthropomorphic evaluation method for automatic driving system
CN113312733B (en) Method, device, equipment and storage medium for calibrating simulation model parameters of signal-controlled roundabout
CN112116813B (en) State recognition model establishing method, road traffic state recognition method and road traffic state recognition device
CN111723672B (en) Method and device for acquiring video recognition driving track and storage medium
CN111460879B (en) Neural network operation method using grid generator and device using the same
CN112749508A (en) Road feel simulation method based on GMM and BP neural network
CN115204755B (en) Service area access rate measuring method and device, electronic equipment and readable storage medium
CN113642682B (en) Trajectory primitive extraction and analysis method and system under multi-vehicle interaction environment
CN111932829B (en) Fatigue driving prevention facility utility testing method and system
CN115366891A (en) Driving style recognition method, system and storage medium
CN115410366A (en) Intersection collision early warning test method and system, electronic device and storage medium
CN114701517A (en) Multi-target complex traffic scene automatic driving solution based on reinforcement learning
CN106997678B (en) Optimization design method for vehicle-mounted road directing information at upstream of expressway exit ramp
CN113092135A (en) Test method, device and equipment for automatically driving vehicle
CN106585634A (en) Method and device for analyzing driving behavior
CN111881121A (en) Automatic driving data filling method and device

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination