CN115545118B

CN115545118B - Vehicle driving evaluation and training method, device, equipment and medium of model thereof

Info

Publication number: CN115545118B
Application number: CN202211463206.9A
Authority: CN
Inventors: 林泰哲
Original assignee: Beijing Jidu Technology Co Ltd
Current assignee: Beijing Jidu Technology Co Ltd
Priority date: 2022-11-16
Filing date: 2022-11-16
Publication date: 2023-04-07
Anticipated expiration: 2042-11-16
Also published as: CN115545118A

Abstract

The embodiment of the application provides a vehicle driving evaluation and model training method, device, equipment and medium thereof. In an embodiment of the present application, the method may include: acquiring evaluation data of a driving intelligence function of a vehicle by a target group, performance parameters of the driving intelligence function, user pictures of the target group and expert evaluation results corresponding to the evaluation data; performing data annotation on the evaluation data based on the user images of the target group and the performance parameters of the intelligent driving function to obtain the evaluation data after the data annotation; and training a vehicle driving evaluation model by taking the evaluation data marked by the data as a training sample set and taking an expert evaluation result corresponding to the evaluation data as a sample label.

Description

Vehicle driving evaluation and training method, device, equipment and medium of model thereof

Technical Field

The application relates to the technical field of machine learning, in particular to a vehicle driving evaluation and model training method, device, equipment and medium.

Background

Along with the continuous promotion of vehicle intelligent degree, the vehicle can drive the function for the intelligence that the user provided also more and more. In order to enable the intelligent driving function provided by the vehicle to reach certain standards in the aspects of safety performance, comfort and the like, the intelligent driving function of the vehicle is generally required to be evaluated. The existing intelligent driving function evaluation method generally needs subjective evaluation on the use experience of the intelligent driving function by a part of people, and needs a test engineer to test the intelligent driving function of a vehicle in various use scenes so as to obtain objective evaluation data.

However, it is difficult for the subjective evaluation data and the objective evaluation data to correspond to each other in the smart driving function due to the difference between the evaluation angle and the evaluation dimension. In addition, the subjective evaluation population may also be different for different intelligent driving functions, which results in great difference in the obtained subjective evaluation data. Moreover, different people have different degrees of understanding about the intelligent driving function or experience about subjective evaluation of the intelligent driving function, so that the obtained subjective evaluation data are different in scoring scale. When the subjective evaluation data correspond to the objective evaluation data, it is difficult to trace back to the product development.

Therefore, how to provide an objective and accurate vehicle driving evaluation method for the intelligent driving function of various vehicles in the prior art to provide a favorable basis for the development and improvement of the intelligent driving function of the vehicle still needs to provide a further solution.

Disclosure of Invention

Aspects of the application provide a vehicle driving evaluation and a training method, device, equipment and medium of a vehicle driving evaluation model, so as to improve the accuracy of the intelligent driving function evaluation of a vehicle, thereby providing a favorable basis for the research and development and improvement of the intelligent driving function of the vehicle.

In a first aspect, an embodiment of the present application provides a training device for a vehicle driving evaluation model, including:

the system comprises a data acquisition module, a data processing module and a data processing module, wherein the data acquisition module is used for acquiring evaluation data of a driving intelligence function of a vehicle by a target group, performance parameters of the driving intelligence function, user pictures of the target group and expert evaluation results corresponding to the evaluation data; the target group at least comprises one of a user group, a vehicle engineer group and a vehicle expert group, and the performance parameters of the intelligent driving function are obtained by testing the intelligent driving function corresponding to the evaluation data;

the data labeling module is used for carrying out data labeling on the evaluation data based on the user images of the target group and the performance parameters of the intelligent driving function to obtain the evaluation data after the data labeling;

and the model training module is used for training the vehicle driving evaluation model by taking the evaluation data marked by the data as a training sample set and taking an expert evaluation result corresponding to the evaluation data as a sample label.

In a second aspect, an embodiment of the present application further provides a vehicle driving evaluation device, including:

the parameter acquisition module is used for acquiring performance parameters of a target intelligent driving function to be evaluated;

the function evaluation module is used for inputting the performance parameters of the target intelligent driving function into a vehicle driving evaluation model so as to output and obtain the grade of at least one user group on the performance parameters of the target intelligent driving function;

the parameter determination module is used for obtaining the optimal performance parameters of the target intelligent driving function based on the grade of the performance parameters of the target intelligent driving function by the at least one user group;

the vehicle driving evaluation model is obtained by training based on the training device of the vehicle driving evaluation model of the first aspect.

In a third aspect, an embodiment of the present application further provides a training method for a vehicle driving evaluation model, including:

acquiring evaluation data of a driving intelligence function of a vehicle by a target group, performance parameters of the driving intelligence function, user pictures of the target group and expert evaluation results corresponding to the evaluation data; the target group at least comprises one of a user group, a vehicle engineer group and a vehicle expert group, and the performance parameters of the intelligent driving function are obtained by testing the intelligent driving function corresponding to the evaluation data;

performing data annotation on the evaluation data based on the user images of the target group and the performance parameters of the intelligent driving function to obtain the evaluation data after the data annotation;

and training a vehicle driving evaluation model by taking the evaluation data marked by the data as a training sample set and taking an expert evaluation result corresponding to the evaluation data as a sample label.

In a fourth aspect, an embodiment of the present application further provides a vehicle driving evaluation method, including:

acquiring performance parameters of a target intelligent driving function to be evaluated;

inputting the performance parameters of the target intelligent driving function into a vehicle driving evaluation model so as to output and obtain the grade of at least one user group on the performance parameters of the target intelligent driving function;

acquiring the optimal performance parameters of the target intelligent driving function based on the grade of the performance parameters of the target intelligent driving function by the at least one user group;

the vehicle driving evaluation model is obtained by training based on the training device of the vehicle driving evaluation model in the first aspect.

In a fifth aspect, an embodiment of the present application further provides a computer device, including: a memory and a processor; wherein the memory is used for storing a computer program; the processor is coupled to the memory for executing the computer program for performing the steps of the method of the third or fourth aspect.

In a sixth aspect, the present application further provides a computer readable storage medium storing a computer program, which when executed by a processor, causes the processor to implement the steps of the method in the third or fourth aspect.

As can be seen from the technical solutions provided in the embodiments of the present application, the embodiments of the present application have at least one of the following technical effects:

according to one or more embodiments provided by the application, the evaluation data of the intelligent driving function of the vehicle by a target group, the performance parameters of the intelligent driving function, the user portrait of the target group and the expert evaluation result corresponding to the evaluation data can be obtained by the data acquisition module; then, the data annotation module carries out data annotation on the evaluation data based on the user image of the target group and the performance parameters of the intelligent driving function; and finally, training the vehicle driving evaluation model by using the evaluation data marked by the data as a training sample set and using an expert evaluation result corresponding to the evaluation data as a sample label through a model training module. The data for training the vehicle driving evaluation model is from one or more of the user group, the vehicle engineer group, the vehicle expert group and the like, the requirements of common users and professional users on the intelligent driving function of the vehicle are considered, the coverage is wide, and the performance parameters of the intelligent driving function for marking the evaluation data are obtained by testing the intelligent driving function corresponding to the evaluation data, so that the obtained vehicle driving evaluation model can accurately predict the preference value of the users of each group on the performance index of the intelligent driving function of the vehicle, meanwhile, the data marking in the training stage also realizes the correspondence between the evaluation data of the intelligent driving function of the user and the test data of the objective performance index, and provides a favorable basis for the improvement of the intelligent driving function of the vehicle from backtracking to the research and development stage.

Drawings

The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:

FIG. 1 is a schematic structural diagram of a training apparatus for a vehicle driving evaluation model according to an exemplary embodiment of the present application;

FIG. 2 is a schematic diagram of a first index dimension in a training apparatus for a vehicle driving evaluation model according to an exemplary embodiment of the present application;

FIG. 3 is a diagram illustrating a second index dimension in a training apparatus for a vehicle driving evaluation model according to an exemplary embodiment of the present application;

FIG. 4 is a schematic diagram of an algorithmic logic of a model in a training apparatus for a vehicle driving evaluation model according to an exemplary embodiment of the present application;

FIG. 5 is a diagram illustrating results of model output in a training apparatus for a vehicle driving evaluation model according to an exemplary embodiment of the present application;

fig. 6 is a schematic structural diagram of a vehicle driving evaluation device according to an exemplary embodiment of the present application;

FIG. 7 is a flowchart illustrating an implementation of a training method for a vehicle driving evaluation model according to an exemplary embodiment of the present application;

FIG. 8 is a schematic flow chart illustrating an implementation of a vehicle driving evaluation method according to an exemplary embodiment of the present disclosure;

fig. 9 is a schematic structural diagram of an electronic device according to an exemplary embodiment of the present application;

fig. 10 is a schematic structural diagram of an electronic device according to an exemplary embodiment of the present application.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application will be described in detail and completely with reference to the following specific embodiments of the present application and the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application, 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 application.

As described in the background art, it is generally difficult to correspond subjective evaluation data and objective evaluation data of an existing vehicle intelligent driving function to evaluation data due to different evaluation angles and dimensions. In addition, the subjective evaluation population may also be different for different intelligent driving functions, which results in great difference in the obtained subjective evaluation data. Moreover, different people have different degrees of understanding about the intelligent driving function or experience about subjective evaluation of the intelligent driving function, so that the obtained subjective evaluation data are different in scoring scale. When the subjective evaluation data correspond to the objective evaluation data, it is difficult to trace back to the product development.

In view of this, for the accuracy that the intelligence of improving the vehicle drives function evaluation to improve the research and development that drives the function for the vehicle intelligence and provide favorable basis, this application embodiment provides a solution, and the basic thinking is: the method comprises the steps that a data acquisition module acquires evaluation data of a target group on the intelligent driving function of a vehicle, performance parameters of the intelligent driving function, a user portrait of the target group and expert evaluation results corresponding to the evaluation data; then, the data annotation module carries out data annotation on the evaluation data based on the user image of the target group and the performance parameters of the intelligent driving function; and finally, training the vehicle driving evaluation model by using the evaluation data marked by the data as a training sample set and using an expert evaluation result corresponding to the evaluation data as a sample label through a model training module.

The technical solutions provided by the embodiments of the present application are described in detail below with reference to the accompanying drawings.

Fig. 1 is a schematic structural diagram of a training device for a vehicle driving evaluation model according to an embodiment of the present application. The apparatus of fig. 1 may comprise:

the data acquisition module 11 is used for acquiring evaluation data of the intelligent driving function of the vehicle by a target group, performance parameters of the intelligent driving function, a user portrait of the target group and expert evaluation results corresponding to the evaluation data; the intelligent driving function performance parameters are obtained by testing the intelligent driving function corresponding to the evaluation data;

the data annotation module 12 is used for performing data annotation on the evaluation data based on user images of a target group and performance parameters of the intelligent driving function to obtain the evaluation data after the data annotation;

and the model training module 13 is configured to train the vehicle driving evaluation model by using the evaluation data after the data labeling as a training sample set and using an expert evaluation result corresponding to the evaluation data as a sample label.

The data for training the vehicle driving evaluation model is from one or more of user groups, vehicle engineer groups, vehicle expert groups and the like, the requirements and the evaluation of common users and professional users on the intelligent driving function of the vehicle are fully considered, the coverage is wide, the performance parameters of the intelligent driving function for marking the evaluation data are obtained by testing the intelligent driving function corresponding to the evaluation data, and therefore the obtained vehicle driving evaluation model can accurately predict the preference value of the performance index of the intelligent driving function of the vehicle by the users of each group. Meanwhile, the data marking in the training stage also realizes the correspondence between the evaluation data of the intelligent driving function and the test data of objective performance indexes by the user, and provides a favorable basis for the improvement of the vehicle intelligent driving function backtracking to the research and development stage. In addition, the data labeling module is used for carrying out data labeling on the evaluation data based on the user image, namely character characteristics of users in all groups when the users evaluate the intelligent driving function of the vehicle are fully considered, so that the accuracy of the scoring of a certain performance parameter of each intelligent driving function of the vehicle by the users predicted by the trained vehicle driving evaluation model is further improved.

Optionally, since different objective environments may also have certain influence on the evaluation data of each group, for example, objective environment parameters such as weather, indoor environment, outdoor environment, and the like may all have influence on the evaluation data to different degrees, the embodiment of the present application further considers the environmental characteristics corresponding to the evaluation data on the basis of considering the characteristics of the person when each group evaluates the smart driving function. Specifically, the data obtaining module is further configured to obtain an environmental parameter corresponding to the evaluation data, and the data labeling module is configured to:

and carrying out data annotation on the evaluation data based on the user image of the target group, the performance parameters of the intelligent driving function and the environment parameters corresponding to the evaluation data to obtain the evaluation data after data annotation.

Optionally, when the target group includes a user group, the degree of understanding, the degree of demand, and the degree of preference of the user group for the intelligent driving function of the vehicle are generally reflected in the index dimensions of usability, trustiness, comfort, practicality, recommendation, humanization, and the like of the user for the intelligent driving function. Based on the evaluation result, the evaluation result of the intelligent driving function of the user group on the vehicle can be obtained from the index dimension. Specifically, when the target group includes a user group, the data obtaining module is specifically configured to, when obtaining evaluation data of the target group on the smart driving function of the vehicle:

acquiring a first evaluation result of a user group on the intelligent driving function according to a first grading gear on a first index dimension, wherein the first index dimension at least comprises one index dimension of usability, trust, comfort, practicability, recommendation and humanization;

the target group comprises at least one user group of a vehicle holding user group, a vehicle experience user group and a vehicle potential user group.

The usability is used for representing whether a user can smoothly use a certain intelligent driving function, how the user uses the intelligent driving function and which users are suitable for using the intelligent driving function. The trust is used for representing the reliability of the intelligent driving function. The comfort is used for representing the comfortable experience feeling brought to the user by the intelligent driving function. The practicality is used for the sign intelligence to drive the function whether can satisfy user's actual demand. The recommendation is used for representing the willingness index of the user to voluntarily recommend or share the intelligent driving function to other users. The humanization is used for representing whether the intelligent driving function can be designed according to the operation habits and living habits of the user, so that the function appeal of the user can be met, and the index of the psychological demand of the user can be met.

As an example, the granularity of the evaluation of the smart driving function by the user group is often larger than that of a vehicle engineer and a vehicle expert, so that the first scoring gear can be set to 2 to be divided into one gear, and the user experience corresponding to each gear is different. Table 1 shows a table of correspondence between the evaluation results of the user group on the smart driving function and the user experience. The first grading gear can be obtained through continuous adjustment of multiple rounds of grading execution in practical application, so that the finally determined grading gear can be high in discrimination.

Table 1 correspondence table between evaluation results of smart driving function by user group and user experience

In table 1, the user experience represented by the evaluation result of 1-2 points is "poor experience, very unsatisfactory", the user experience represented by the evaluation result of 3-4 points is "poor experience, more problems", the user experience represented by the evaluation result of 5-6 points is "general experience, well-regulated moment", the user experience represented by the evaluation result of 7-8 points is "good experience, can be circled, and the user experience represented by the evaluation result of 9-10 points is" good experience, very satisfactory ". Meanwhile, the node can be divided into 7 points, and the intelligent driving function or the driving sub-function with the evaluation result of 7 points or more meets the mass production standard.

In addition, for a consumer-level user group of a non-fixed group, for a scoring gear of 2-degree system, the scoring of each gear and the understanding cost of the corresponding user experience of the user group can be effectively reduced, and meanwhile, the operation cost in the specific evaluation execution process of the user group is reduced.

Fig. 2 is a schematic diagram of a first index dimension in a training device for a vehicle driving evaluation model according to an exemplary embodiment of the present application. In fig. 2, the usability and the trust of the intelligent driving function are taken as the basis of the requirement of the user group on the intelligent driving function, on this basis, the user group evaluates the comfort and the practicability of the intelligent driving function, and under the condition that the user group is satisfied with the comfort and the practicability of the intelligent driving function, the intention of recommending the intelligent driving function to other users is provided, so that the recommendation of the intelligent driving function is above the comfort and the practicability of the intelligent driving function. The humanization of the intelligent driving function generally plays a role in improving the intelligence, so the humanization of the intelligent driving function can be used as the last basis for the user to grade the intelligent driving function, or the humanization can be used as the most intuitive basis for the intelligent driving function.

The target groups may include a vehicle holding user group, a vehicle experiencing user group, and a vehicle potential user group. The vehicle owner group represents a vehicle family, but is not specific to vehicle owners holding electric vehicles, fuel oil vehicles or fuel-electric hybrid vehicles, namely all vehicle types under the name of the vehicle owners are not distinguished, and the group can present universal understanding on the use degree of the intelligent driving function. The vehicle experience user group represents a user group who has used, or experienced, the smart driving function, which is a group with a certain understanding of the smart driving function. The vehicle potential user group is a user group willing to buy a list for the intelligent driving function of the vehicle, and the group has strong preference for the prospective function and the trend new function.

Optionally, when the target group includes a group of vehicle engineers, the group of vehicle engineers may evaluate the driving intelligence function of the vehicle from the perspective of their expertise, and the evaluation perspective may include the dimensions of the indicators such as visibility, legibility, operability, steering, dynamics, braking, accuracy, drivability, noise, stability, ride quality, and visibility during the use of the driving intelligence function. Based on the evaluation result, the evaluation result of the intelligent driving function of the vehicle by the vehicle engineer group can be obtained from the index dimension. Specifically, when the target group includes a group of vehicle engineers, the data acquisition module is specifically configured to, when acquiring evaluation data of the target group on the smart driving function of the vehicle:

acquiring a second evaluation result of the intelligent driving function of a vehicle engineer group on a second index dimension according to a second grading gear;

the second index dimension includes at least one of visibility, legibility, operability, steerability, drivability, braking, accuracy, drivability, noise, stability, ride quality, and visibility.

Wherein, the visual field is used for representing whether the intelligent driving function can be seen on the premise that the user has no burden on the body in a comfortable driving posture. Legibility is used to characterize whether the smart driving function is easily seen in color and shape. The operability is used for representing the index of whether the intelligent driving function can be observed, repeated and directly operated. The steering performance is used for representing the related intelligent driving function, so that stable and accurate steering characteristics can be provided for the vehicle. The dynamic property is used for representing the related intelligent driving function, and the average driving speed which can be reached by the vehicle and is determined by the longitudinal external force applied to the vehicle when the vehicle is driven on a good road surface in a straight line can be provided for the vehicle. The braking performance is used for representing the capability of a related intelligent driving function of stopping the vehicle in a short time and maintaining the stability of the driving direction when the vehicle runs and maintaining a certain speed when the vehicle runs on a long slope. Accuracy is used to characterize the accuracy of the intelligent driving function. Drivability is used to characterize the power output and transfer characteristics of the vehicle provided by the associated smart driving function. The noise is used for representing the noise generated when the related intelligent driving function is triggered. Stability is used to characterize the constant ability of the Smart-drive function over time. Ride quality is used to characterize the ride comfort that the relevant smart driving function provides to the user. The visibility is used for representing the advancing and retreating performance, the expansion and contraction performance and the brightness of the color of the intelligent driving function.

As an example, since the vehicle engineers belong to the field of vehicle profession, the granularity of the evaluation of the intelligent driving function by the vehicle engineer group is often more surprised than that by the user group, so that the second scoring gear can be set to 1 and divided into one gear, and the user experience corresponding to each gear is different. Table 2 shows a table of correspondence between the evaluation results of the vehicle engineer group on the smart driving function and the user experience. The second grading gear can be obtained through continuous adjustment of multiple rounds of grading execution in practical application, so that the finally determined grading gear can be evaluated more finely.

Table 2 correspondence table between evaluation result of the driver intelligence function by the vehicle engineer and the user experience

Evaluation results

1 minute (1)

2 is divided into

3 points of

4 is divided into

5 points of

6+ point

7-minute

7 points of

7+ point

8-minute

8 is divided into

9 minutes

10 minutes

User experience

Not evaluated

Very poor

Difference (D)

Is poor

Can also be used for

In general terms

Go well

Good wine

Is good

Is preferably used

Good taste

Is excellent in

Perfection

In table 2, the user experience represented by the evaluation result of 1 point is "no evaluation", and the corresponding evaluation is described as "the function is just developed, and it can be directly judged not to be a subjective evaluation item or not to be evaluated"; the user experience represented by the evaluation result of 2 points is 'very poor', and the corresponding evaluation is described as 'the function cannot work normally, the function is unstable and has obvious major problems (including software and hardware)'; the user experience represented by the evaluation result of 3 points is "very poor", and the corresponding evaluation is described as "the function is very poor in performance (or experience), and the function is only used as a judgment basis for whether the function can be started or not; the user experience represented by the evaluation result of 4 points is "poor", and the corresponding evaluation is described as "the function performance (or experience) is poor, so that the user can try to use the evaluation, and the evaluation can be tried under the precondition of ensuring safety"; the user experience represented by an evaluation result of 5 points is "poor", and the corresponding evaluation is described as "this function performance (or experience) is poor, and evaluation can be attempted, but there are many obvious problems".

The user experience represented by the evaluation result of 6 points is 'available', and the corresponding evaluation is described as 'the functional performance (or) experience is available, the evaluation can be carried out, and a plurality of comprehensive problem points exist to continue to improve'; the user experience represented by the evaluation result of 6+ score is general, and the corresponding evaluation is described as that the functional performance (or) experience can be further evaluated, so that a comprehensive problem point exists, and the mass production can be realized after the improvement is needed; the user experience represented by the evaluation result of 7-point is 'good', and the corresponding evaluation is described as 'the function has good performance (or experience) and can be used as a mass production condition for judgment, but has partial problem points and can be modified in a short time for mass production'; the user experience represented by an evaluation result of 7 points is 'good', and the corresponding evaluation is described as 'the functional performance (or experience) is good, and the functional performance (or experience) can be used as a mass production condition for judgment, no obvious problem point exists, and part of problem points need to be updated in an iterative manner'; the user experience represented by 8-point evaluation result is 'better', the corresponding evaluation is described as 'the function is better in performance (or experience), partial advantages are provided compared with the competitive products, and partial detailed functions can be optimized subsequently'.

The user experience represented by 8 points of evaluation result is 'good', the corresponding evaluation is described as 'the function is well represented (or experienced), the competitive products have advantages compared with the competitive products, and partial detailed functions can be subsequently optimized'; the user experience represented by the evaluation result of 9 points is 'excellent', and the corresponding evaluation is described as 'the function performance (or experience) is very good and can be used as a market and industry wind vane without a detailed function optimization part'; the user experience represented by the evaluation result of 10 points is 'perfect', and the corresponding evaluation is described as 'the function is very perfect, so that the conventional product and competitive products (including prospective, conceptual and imaginary functions) cannot exceed the function and have no problem points'. Meanwhile, the evaluation result can be longitudinally matched with the first scoring gear of the user group, and the evaluation result is judged to be 7 points or more, so that the intelligent driving function or the driving sub-function meets the mass production standard.

In addition, for the consumer-level user group of the non-fixed group, for the score grade of 2 scores, the score of each grade and the understanding cost of the corresponding user experience of the user group can be effectively reduced, and meanwhile, the operation cost in the specific evaluation execution process of the user group is reduced.

Fig. 3 is a schematic diagram of a second index dimension in a training device for a vehicle driving evaluation model according to an exemplary embodiment of the present application. In fig. 3, the smart driving function can be evaluated from the dimensions of vision, hearing, upper limb, lower limb operation, and overall experience. Wherein, the visual evaluation can be carried out by three index dimensions of visual field, visual recognition and easy recognition. The auditory perception can be evaluated by the index dimension of noisiness. The upper limb operation can be evaluated through several index dimensions of operability, steering and accuracy. The lower limb operation can be evaluated through several index dimensions of dynamic performance, braking system and driving performance. The overall experience can be evaluated through two index dimensions, namely ride quality and stability.

Obviously, compared with the user group, the evaluation dimension of the intelligent driving function is more comprehensive and deeper by the vehicle engineer group. The product PRD corresponding to the intelligent driving function is developed and matched with a research and development model in a targeted mode, the experience level of a user and the positioning of a problem point module on control logic are determined, and the product force for producing and researching a closed loop and further continuously grinding the intelligent driving function is integrally formed. For example, a vehicle engineer evaluates the intelligent driving function "efficiency lane change" according to the second scoring gear in the second indexing dimension, taking the stability dimension of the vehicle in the lane change process as an example, and if the vehicle is in a stable, smooth and smooth lane change scene, the evaluation result may be good. And carrying out comprehensive scoring according to different performances on the second index dimension in the vehicle lane changing process, so that the evaluation result of the intelligent driving function of 'efficiency lane changing' can be obtained.

However, if the intelligent driving function 'efficiency lane change' has obvious longitudinal pitching feeling and obvious rolling performance in the evaluation process, the user experience that the longitudinal pitching feeling is positioned as the problem of acceleration and braking performance, and the user experience that the lateral acceleration of the rolling positioning is too large or the lateral instantaneous torque is too large causes poor overall lane change body feeling. And in the evaluation process of the intelligent driving function 'efficiency lane change', corresponding different performance parameters, such as longitudinal speed, acceleration and acceleration change, lane change time, transverse speed, transverse acceleration and other corresponding specific experiences can be positioned.

Optionally, when the target group includes a vehicle expert group, the vehicle expert group may evaluate the intelligent driving function of the vehicle from a more professional perspective, and the evaluation angle may be performed on the basis of the first evaluation result and the second evaluation result, and meanwhile, the index dimensions may correspond to the performance indexes corresponding to the intelligent driving function of the vehicle obtained through objective testing, so that the subjective evaluation result and the objective test result correspond to the dimension of the performance indexes of the intelligent driving function, thereby achieving accurate problem location. Specifically, when the target group includes a vehicle expert group, the data acquisition module is used for acquiring evaluation data of the target group on the intelligent driving function of the vehicle, and specifically:

acquiring a first evaluation result of the intelligent driving function by a user group on a first index dimension according to a first grading gear, a second evaluation result of the intelligent driving function by a vehicle engineer group on a second index dimension according to a second grading gear, and a performance parameter of the intelligent driving function;

acquiring a third evaluation result of the intelligent driving function of a vehicle expert group according to a third grading gear on a third index dimension based on the first evaluation result, the second evaluation result and the performance parameter of the intelligent driving function;

the third index dimension includes a first index dimension and a second index dimension; the third grading grade is smaller than the second grading grade, and the second grading grade is smaller than the first grading grade.

As an example, since the vehicle expert is more specialized and detailed in research on the vehicle, the granularity of evaluation of the smart driving function is more detailed than that of a user group and a vehicle engineer, and thus the third scoring gear may be set to 0.1 to be divided into one gear aiming at grinding the production force of the smart driving function in more detail.

In addition, in order to realize one-to-one correspondence between subjective evaluation results and objective test results, time stamping can be performed on different performance parameters of the intelligent driving function to enable the different performance parameters to correspond to the subjective evaluation results. Specifically, a third evaluation result of the intelligent driving function of the vehicle expert group according to a third grading gear on a third index dimension is obtained based on the first evaluation result, the second evaluation result and the performance parameters of the intelligent driving function, and the third evaluation result not only comprises an overall subjective evaluation result on the third index dimension, but also can be corresponding to the specific performance parameters of the intelligent driving function.

For example, in the process of evaluating the intelligent driving function of the vehicle, namely the vehicle following start-stop, the deceleration acceleration values of the vehicle can be counted, and it is assumed that the intelligent driving function of the vehicle, namely the vehicle following start-stop, is evaluated five thousand times, wherein the evaluation result corresponding to the deceleration acceleration value of-0.5 is 5 minutes, the evaluation result corresponding to the deceleration acceleration value of-0.2 is 7 minutes, the evaluation result corresponding to the deceleration acceleration value of-0.05 is 8 minutes, and the scoring gear of a single performance index can be consistent with the second scoring gear of a vehicle engineer. The expert group can obtain the times of the deceleration acceleration value of the vehicle between-0.2 and-0.05 according to statistics to obtain the overall score of the intelligent driving function 'start and stop with the vehicle', if the times is less than or equal to 500 times, the corresponding evaluation result can be defined as 5.8 points, if the times is more than 500 times and less than or equal to 1000 times, the corresponding evaluation result can be defined as 6.1 points, and if the times is more than 1000 times and less than 3000 times, the corresponding evaluation result can be defined as 7.3 points. When historical data is counted for multiple times, evaluation can be carried out in a scoring gear of 0.1 point.

Optionally, since the vehicle expert group may evaluate the driving intelligence function of the vehicle from a more professional perspective, and the evaluation angle may be performed based on the first evaluation result and the second evaluation result, the obtained third evaluation result of the driving intelligence function of the vehicle expert may also be used as a sample label corresponding to the evaluation data for training the vehicle driving evaluation model. Specifically, when the data acquisition module acquires an expert evaluation result corresponding to the evaluation data, the data acquisition module is specifically configured to:

and acquiring an expert evaluation result corresponding to the evaluation data based on a third evaluation result of the vehicle expert group on the intelligent driving function in a third index dimension according to a third grading gear.

As an example, the third evaluation result under the same smart driving function may be used as the expert evaluation result corresponding to the evaluation data. And the third evaluation result and the evaluation data are obtained aiming at the same intelligent driving function evaluation.

Optionally, when the model training module trains the vehicle driving evaluation model by using the evaluation data after data labeling as a training sample set and using an expert evaluation result corresponding to the evaluation data as a sample label, the model training module is specifically configured to:

taking the evaluation data after data annotation as the input of a vehicle driving evaluation model;

and adjusting parameters of the vehicle driving evaluation model to enable the model output result of the vehicle driving evaluation model to approach an expert evaluation result corresponding to the evaluation data, or enabling the regression evaluation coefficient of the vehicle driving evaluation model to reach a preset value.

It should be understood that the process of training the vehicle driving evaluation model is a process of continuously adjusting the model parameters according to the sample labels, so that the output result of the vehicle driving evaluation model of the adjusted model parameters infinitely approaches the expert evaluation result corresponding to the evaluation data, or the regression evaluation coefficient of the vehicle driving evaluation model reaches a preset value. As an example, the preset value may be set to 0.99.

Optionally, when the evaluation data after the data annotation is used as the input of the vehicle driving evaluation model, the model training module is specifically configured to:

removing the evaluation result in the evaluation data after the data annotation to obtain model input data;

the model input data is used as the input of the vehicle driving evaluation model.

It should be understood that after the evaluation data after the data annotation is obtained, since the evaluation data may be from a user group, a vehicle engineer group or an expert group, the evaluation result included in the evaluation data cannot be guaranteed to be completely objective and accurate. In the training stage, the expert evaluation result corresponding to the evaluation data representing the objective and accurate evaluation result is used as a sample label, so that the evaluation result in the evaluation data after data labeling can be eliminated.

Optionally, the model training module is specifically configured to, when adjusting parameters of the vehicle driving evaluation model:

determining a parameter discrete relation between the evaluation result of the intelligent driving function by the target group and the performance parameter of the intelligent driving function;

determining the weight of the performance parameters of the intelligent driving function based on the importance degree of the performance parameters of the intelligent driving function to the evaluation result of the intelligent driving function;

based on the weight of the performance parameters of the parameter discrete relation and the intelligent driving function, the optimal parameters of each blade node are determined from a plurality of parameters of each blade node decided by the algorithm of the vehicle driving evaluation model, so that the model output result obtained based on the optimal parameters of each blade node approaches the expert evaluation result corresponding to the evaluation data, or the regression evaluation coefficient of the vehicle driving evaluation model reaches a preset value.

The intelligent driving system comprises a target group, an intelligent driving function, a plurality of expert evaluation results and a plurality of intelligent driving function evaluation results, wherein the target group is used for evaluating the intelligent driving function and the intelligent driving function, and the intelligent driving function evaluation results are respectively matched with the expert evaluation results corresponding to the intelligent driving function.

Fig. 4 is a schematic diagram of an algorithm logic of a model in a training device for a vehicle driving evaluation model according to an exemplary embodiment of the present application. In fig. 4, the evaluation data of the target group on the intelligent driving function is used as the input data of the vehicle driving evaluation model, the vehicle driving evaluation model is trained for multiple times until an expert evaluation result corresponding to the infinite approximation evaluation data is obtained through output, then the expert evaluation result corresponds to the performance parameter of the intelligent driving function when the target group evaluates the intelligent driving function, the evaluation data can be re-labeled according to the result, and the model parameter is continuously optimized. Meanwhile, the evaluation results of different user portraits under the performance parameters of the same intelligent driving function can be clustered to obtain preference values of users with different user portraits on the performance parameters of the intelligent driving function.

Wherein, X = { X1, X2 \8230andXn } is n blade nodes in the model, and each blade node in the blade nodes comprises a plurality of parameter values. The parameter discrete relation between the evaluation result of the intelligent driving function and the performance parameters of the intelligent driving function by the target group can be determined through different parameters in the blade nodes, and meanwhile, the weight Y can be calculated by taking the performance parameters of the intelligent driving function as parameters. Taking X1 as an example, the parameters a-d, \8230inX 1 have n values, the parameter selection of the whole blade node X1 can be regarded as the number of table lines of n X n, each point is taken as a value combination of one parameter, each point represents the point adopted by the model for calculation and prediction, and finally, the parameter corresponding to the point with the best prediction effect in each blade node is selected as the optimal parameter of the model.

According to the training device for the vehicle driving evaluation model provided by one or more embodiments, the data acquisition module can acquire evaluation data of the intelligent driving function of the vehicle by a target group, performance parameters of the intelligent driving function, a user figure of the target group and expert evaluation results corresponding to the evaluation data; then, the data annotation module carries out data annotation on the evaluation data based on the user image of the target group and the performance parameters of the intelligent driving function; and finally, training the vehicle driving evaluation model by using the evaluation data marked by the data as a training sample set and using an expert evaluation result corresponding to the evaluation data as a sample label through a model training module. The data for training the vehicle driving evaluation model is from one or more of a user group, a vehicle engineer group, a vehicle expert group and the like, the requirements of common users and professional users on the intelligent driving function of the vehicle are fully considered, the coverage is wide, and the performance parameters of the intelligent driving function for marking the evaluation data are obtained by testing the intelligent driving function corresponding to the evaluation data, so that the obtained vehicle driving evaluation model can accurately predict the preference value of the users of each group on the performance index of the intelligent driving function of the vehicle, meanwhile, the data marking in the training stage also realizes the correspondence between the evaluation data of the intelligent driving function of the users and the test data of the objective performance index, and a favorable basis is provided for the improvement of the intelligent driving function of the vehicle from backtracking to the research and development stage.

Fig. 5 is a schematic structural diagram of a vehicle driving evaluation device according to still another exemplary embodiment of the present application. As shown in fig. 5, the vehicle driving evaluation device includes:

the parameter obtaining module 51 is configured to obtain a performance parameter of a target intelligent driving function to be evaluated;

the function evaluation module 52 is configured to input the performance parameter of the target intelligent driving function into a vehicle driving evaluation model, so as to output a score of the performance parameter of the target intelligent driving function obtained by at least one user group;

the parameter determination module 53 is configured to obtain an optimal performance parameter of the target intelligent driving function based on the score of the at least one user group on the performance parameter of the target intelligent driving function;

the vehicle driving evaluation model is obtained by training through a training device based on the vehicle driving evaluation model.

Fig. 6 is a schematic diagram illustrating a result of a model output in a training apparatus for a vehicle driving evaluation model according to an exemplary embodiment of the present application. In fig. 6, for a specific performance parameter of the target smart driving function, preference values of different groups for the performance parameter of the target smart driving function can be obtained through prediction by a vehicle driving evaluation model. Wherein, 1 in fig. 6 may be a vehicle holding group, 2 may be a vehicle experiencing group, and 3 may be a vehicle potential group, and the shaded portion in fig. 6 is an optimal solution of preference values of the three user groups for the performance parameter of the target smart driving function.

According to the vehicle fault diagnosis device provided by one or more embodiments, data for training the vehicle driving evaluation model are from one or more of a user group, a vehicle engineer group, a vehicle expert group and the like, requirements of common users and professional users on the intelligent driving function of the vehicle are fully considered, the coverage is wide, and performance parameters of the intelligent driving function for marking the evaluation data are obtained by testing the intelligent driving function corresponding to the evaluation data, so that the obtained vehicle driving evaluation model can accurately predict preference values of users of each group on the performance indexes of the intelligent driving function of the vehicle, meanwhile, the data marking in the training stage also realizes the correspondence between the evaluation data of the intelligent driving function of the users and the test data of objective performance indexes, and a favorable basis is provided for the improvement of the intelligent driving function of the vehicle in the stage of backtracking to research and development.

The specific implementation of the vehicle driving evaluation device shown in fig. 6 has been described in detail in the embodiment of the training device of the vehicle driving evaluation model, and will not be elaborated here.

Fig. 7 is a schematic flow chart of an implementation process of a training method for a vehicle driving evaluation model according to an exemplary embodiment of the present application. In fig. 7, the method may include:

and 710, obtaining evaluation data of the intelligent driving function of the vehicle by a target group, performance parameters of the intelligent driving function, a user portrait of the target group and expert evaluation results corresponding to the evaluation data.

The target group at least comprises a user group, a vehicle engineer group and a vehicle expert group, and the performance parameters of the intelligent driving function are obtained by testing the intelligent driving function corresponding to the evaluation data.

Optionally, when the target group includes a user group, obtaining evaluation data of the target group on the intelligent driving function of the vehicle includes:

Optionally, when the target group includes a group of vehicle engineers, obtaining evaluation data of the target group on the smart driving function of the vehicle includes:

acquiring a second evaluation result of the intelligent driving function of the vehicle engineer group on a second index dimension according to a second grading gear;

Optionally, when the target group includes a vehicle expert group, obtaining evaluation data of the target group on the intelligent driving function of the vehicle includes:

acquiring a first evaluation result of the intelligent driving function by a user group on a first index dimension according to a first grading gear, a second evaluation result of the intelligent driving function by a vehicle engineer group on a second index dimension according to a second grading gear, and performance parameters of the intelligent driving function;

acquiring a third evaluation result of the intelligent driving function of a vehicle expert group according to a third grading gear on a third index dimension on the basis of the first evaluation result, the second evaluation result and the performance parameters of the intelligent driving function;

Optionally, acquiring an expert evaluation result corresponding to the evaluation data includes:

And 720, performing data annotation on the evaluation data based on the user images of the target group and the performance parameters of the intelligent driving function to obtain the evaluation data after the data annotation.

Optionally, the data acquired in step 710 further includes an environmental parameter corresponding to the evaluation data, and then the data annotation is performed on the evaluation data based on the user image of the target group and the performance parameter of the smart driving function, so as to obtain the evaluation data after the data annotation, including:

And step 730, training the vehicle driving evaluation model by taking the evaluation data after data labeling as a training sample set and an expert evaluation result corresponding to the evaluation data as a sample label.

Optionally, training a vehicle driving evaluation model by using the evaluation data after data labeling as a training sample set and using an expert evaluation result corresponding to the evaluation data as a sample label, including:

Optionally, the inputting the evaluation data after the data tagging as the vehicle driving evaluation model includes:

Optionally, adjusting parameters of the vehicle driving evaluation model comprises:

determining a parameter discrete relation between an evaluation result of the target group on the intelligent driving function and a user portrait of the target group;

The specific implementation of the training method for the vehicle driving evaluation model shown in fig. 7 has been described in detail in the embodiment of the training device for the vehicle driving evaluation model, and will not be described in detail here.

Fig. 8 is a schematic implementation flow chart of a vehicle driving evaluation method according to an exemplary embodiment of the present application. In fig. 8, the method may include:

step 810, acquiring performance parameters of a target intelligent driving function to be evaluated;

step 820, inputting the performance parameters of the target intelligent driving function into a vehicle driving evaluation model so as to output and obtain the score of at least one user group on the performance parameters of the target intelligent driving function;

step 830, obtaining the optimal performance parameters of the target intelligent driving function based on the scoring of the performance parameters of the target intelligent driving function by at least one user group;

the vehicle driving evaluation model is obtained by training a training device based on the vehicle driving evaluation model.

The detailed implementation of the vehicle driving evaluation method shown in fig. 8 has been described in detail in the embodiment of the vehicle driving evaluation device, and will not be described in detail here.

It should be noted that in some of the flows described in the above embodiments and the drawings, a plurality of operations are included in a specific order, but it should be clearly understood that these operations may be executed out of the order presented herein or in parallel, and the sequence numbers of the operations, such as 710, 720 and 810, 820, etc., are merely used for distinguishing different operations, and the sequence numbers themselves do not represent any execution order. Additionally, the flows may include more or fewer operations, and the operations may be performed sequentially or in parallel. It should be noted that, the descriptions of "first", "second", etc. in this document are used for distinguishing different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit the types of "first" and "second".

Fig. 9 is a schematic structural diagram of an electronic device according to an exemplary embodiment of the present application. Referring to fig. 9, the electronic device includes: a memory 91 and a processor 92.

The memory 91 is used to store computer programs and may be configured to store various other data to support operations on the computing platform. Examples of such data include instructions for any application or method operating on the computing platform, contact data, phonebook data, messages, pictures, videos, and so forth.

The memory 91 may be implemented by any type or combination of volatile or non-volatile memory devices, such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disks.

A processor 92, coupled to the memory 91, for executing the computer program in the memory 91 for: acquiring evaluation data of a driving intelligence function of a vehicle by a target group, performance parameters of the driving intelligence function, user pictures of the target group and expert evaluation results corresponding to the evaluation data; the target group at least comprises one or more of a user group, a vehicle engineer group and a vehicle expert group, and the performance parameters of the intelligent driving function are obtained by testing the intelligent driving function corresponding to the evaluation data;

performing data annotation on the evaluation data based on the user portraits of the target group and the performance parameters of the intelligent driving function to obtain the evaluation data after the data annotation;

Further optionally, when the target group includes a user group, the processor 92 is specifically configured to, when obtaining the evaluation data of the target group on the smart driving function of the vehicle:

acquiring a first evaluation result of the intelligent driving function of the user group on a first index dimension according to a first grading gear, wherein the first index dimension at least comprises one index dimension of usability, trust, comfort, practicability, recommendation and humanization;

the target group comprises at least one of a vehicle holding user group, a vehicle experience user group and a vehicle potential user group.

Further optionally, when the target group includes a group of vehicle engineers, the processor 92 is specifically configured to, when obtaining the evaluation data of the target group on the smart driving function of the vehicle:

acquiring a second evaluation result of the intelligent driving function by the vehicle engineer group on a second scale dimension according to a second grading gear;

the second index dimension includes at least one index dimension of visibility, legibility, operability, steering, dynamics, braking, accuracy, drivability, noise, stability, ride quality, and visibility.

Further optionally, when the target group includes a vehicle expert group, the processor 92 is specifically configured to, when obtaining the evaluation data of the target group on the driving intelligence function of the vehicle:

acquiring a first evaluation result of the intelligent driving function by the user group on a first index dimension according to a first grading gear, a second evaluation result of the intelligent driving function by the vehicle engineer group on a second index dimension according to a second grading gear, and performance parameters of the intelligent driving function;

obtaining a third evaluation result of the intelligent driving function by the vehicle expert group according to a third grading gear on a third index dimension based on the first evaluation result, the second evaluation result and the performance parameter of the intelligent driving function;

the third index dimension comprises the first index dimension and the second index dimension; the third scoring gear is smaller than the second scoring gear, and the second scoring gear is smaller than the first scoring gear.

Further optionally, when obtaining the expert evaluation result corresponding to the evaluation data, the processor 92 is specifically configured to:

and acquiring an expert evaluation result corresponding to the evaluation data based on a third evaluation result of the vehicle expert group on the intelligent driving function according to the third grading gear on the third index dimension.

Further optionally, when the evaluation data labeled with the data is used as a training sample set and an expert evaluation result corresponding to the evaluation data is used as a sample label to train the vehicle driving evaluation model, the processor 92 is specifically configured to:

taking the evaluation data marked by the data as the input of the vehicle driving evaluation model;

and adjusting parameters of the vehicle driving evaluation model to enable a model output result of the vehicle driving evaluation model to approach an expert evaluation result corresponding to the evaluation data, or enabling a regression evaluation coefficient of the vehicle driving evaluation model to reach a preset value.

Further optionally, when the evaluation data labeled with the data is used as the input of the vehicle driving evaluation model, the processor 92 is specifically configured to:

and taking the model input data as the input of the vehicle driving evaluation model.

Further optionally, when adjusting the parameters of the vehicle driving evaluation model, the processor 92 is specifically configured to:

determining a parameter discrete relation between the evaluation result of the target group on the intelligent driving function and a user portrait of the target group;

based on the parameter discrete relation and the weight of the performance parameter of the intelligent driving function, determining the optimal parameter of each blade node from a plurality of parameters of each blade node decided by the algorithm of the vehicle driving evaluation model, so that the model output result obtained based on the optimal parameter of each blade node approaches the expert evaluation result corresponding to the evaluation data, or the regression evaluation coefficient of the vehicle driving evaluation model reaches a preset value.

Further, as shown in fig. 9, the electronic device further includes: communication components 93, display 94, power components 95, audio components 96, and the like. Only some of the components are schematically shown in fig. 9, and the electronic device is not meant to include only the components shown in fig. 9. In addition, the components within the dashed line frame in fig. 9 are optional components, not necessary components, and may be determined according to the product form of the electronic device. The terminal device of this embodiment may be implemented as a desktop computer, a notebook computer, a smart phone, an IOT device, or other terminal devices that can be deployed in a vehicle, or may be a conventional server, a cloud server, or a server array, or other server-side devices. If the terminal device of this embodiment is implemented as a terminal device such as a desktop computer, a notebook computer, a smart phone, etc., the terminal device may include components within a dashed box in fig. 9; if the terminal device of this embodiment is implemented as a server device such as a conventional server, a cloud server, or a server array, the terminal device may not include components in the dashed line box in fig. 9.

Accordingly, the present application further provides a computer-readable storage medium storing a computer program, where the computer program is capable of implementing the steps that can be executed by the electronic device in the foregoing method embodiments when executed.

The communication component is configured to facilitate wired or wireless communication between the device in which the communication component is located and other devices. The device where the communication component is located can access a wireless network based on a communication standard, such as a WiFi, a 2G, 3G, 4G/LTE, 5G and other mobile communication networks, or a combination thereof. In an exemplary embodiment, the communication component receives a broadcast signal or broadcast related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, the communication component further includes a Near Field Communication (NFC) module to facilitate short-range communications. For example, the NFC module may be implemented based on Radio Frequency Identification (RFID) technology, infrared data association (IrDA) technology, ultra Wideband (UWB) technology, bluetooth (BT) technology, and other technologies.

The display includes a screen, which may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive an input signal from a user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure associated with the touch or slide operation.

The power supply assembly provides power for various components of the device in which the power supply assembly is located. The power components may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the device in which the power component is located.

The audio component may be configured to output and/or input an audio signal. For example, the audio component includes a Microphone (MIC) configured to receive an external audio signal when the device in which the audio component is located is in an operational mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signal may further be stored in a memory or transmitted via a communication component. In some embodiments, the audio assembly further comprises a speaker for outputting audio signals.

Fig. 10 is a schematic structural diagram of an electronic device according to an exemplary embodiment of the present application. Referring to fig. 10, the electronic device includes: a memory 101 and a processor 102.

The memory 101 is used to store computer programs and may be configured to store various other data to support operations on the computing platform. Examples of such data include instructions for any application or method operating on the computing platform, contact data, phonebook data, messages, pictures, videos, and so forth.

The memory 101 may be implemented by any type or combination of volatile or non-volatile memory devices, such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disks.

A processor 102, coupled to the memory 101, for executing the computer program in the memory 101 to: acquiring performance parameters of a target intelligent driving function to be evaluated;

Further, as shown in fig. 10, the electronic device further includes: communication component 103, display 104, power component 105, audio component 106, and other components. Only some of the components are schematically shown in fig. 10, and the electronic device is not meant to include only the components shown in fig. 10. In addition, the components within the dashed line frame in fig. 10 are optional components, not necessary components, and may be determined according to the product form of the electronic device. The terminal device of this embodiment may be implemented as a desktop computer, a notebook computer, a smart phone, an IOT device, or other terminal devices that can be deployed in a vehicle, or may be a conventional server, a cloud server, or a server array, or other server-side devices. If the terminal device of this embodiment is implemented as a desktop computer, a notebook computer, a smart phone, or other terminal devices, the terminal device may include components within a dashed line frame in fig. 10; if the terminal device of this embodiment is implemented as a server device such as a conventional server, a cloud server, or a server array, the components in the dashed box in fig. 10 may not be included.

As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include forms of volatile memory in a computer readable medium, random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.

Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.

It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrases "comprising one of 8230; \8230;" 8230; "does not exclude the presence of additional like elements in a process, method, article, or apparatus that comprises that element.

The above description is only an example of the present application and is not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A training device for a vehicle driving evaluation model, comprising:

the system comprises a data acquisition module, a data processing module and a data processing module, wherein the data acquisition module is used for acquiring evaluation data of a driving intelligence function of a vehicle by a target group, performance parameters of the driving intelligence function, user pictures of the target group and expert evaluation results corresponding to the evaluation data; the target group at least comprises a user group, a vehicle engineer group and a vehicle expert group, and the performance parameters of the intelligent driving function are obtained by testing the intelligent driving function corresponding to the evaluation data;

the data marking module is used for carrying out data marking on the evaluation data based on the user images of the target group and the performance parameters of the intelligent driving function to obtain the evaluation data after the data marking, so that the evaluation data of the intelligent driving function corresponds to the test data of objective performance indexes;

2. The apparatus of claim 1, wherein when the target group comprises a user group, the data acquisition module, when acquiring the evaluation data of the target group for the smart driving function of the vehicle, is specifically configured to:

the user group comprises at least one of a vehicle holding user group, a vehicle experience user group and a vehicle potential user group.

3. The apparatus of claim 1, wherein when the target group comprises a group of vehicle engineers, the data acquisition module, when acquiring the evaluation data of the target group for the smart driving function of the vehicle, is specifically configured to:

4. The apparatus of claim 1, wherein when the target group comprises a vehicle expert group, the data acquisition module, when acquiring the evaluation data of the target group for the intelligent driving function of the vehicle, is specifically configured to:

acquiring a first evaluation result of the intelligent driving function by the user group on a first index dimension according to a first scoring gear, a second evaluation result of the intelligent driving function by the vehicle engineer group on a second index dimension according to a second scoring gear, and a performance parameter of the intelligent driving function;

5. The apparatus according to claim 4, wherein the data obtaining module, when obtaining the expert evaluation result corresponding to the evaluation data, is specifically configured to:

6. The apparatus according to claim 1, wherein the model training module, when training the vehicle driving evaluation model by using the evaluation data labeled with the data as a training sample set and using an expert evaluation result corresponding to the evaluation data as a sample label, is specifically configured to:

7. The apparatus of claim 6, wherein the model training module, when taking the evaluation data labeled with the data as the input of the vehicle driving evaluation model, is specifically configured to:

8. The apparatus of claim 6 or 7, wherein the model training module, when adjusting the parameters of the vehicle driving evaluation model, is specifically configured to:

and respectively determining the optimal parameters of each blade node from a plurality of parameters of each blade node decided by the algorithm of the vehicle driving evaluation model based on the parameter discrete relation and the weight of the performance parameters of the intelligent driving function, so that the model output result obtained based on the optimal parameters of each blade node approaches the expert evaluation result corresponding to the evaluation data, or the regression evaluation coefficient of the vehicle driving evaluation model reaches a preset value.

9. The apparatus of claim 1, wherein the data obtaining module is further configured to obtain an environmental parameter corresponding to the evaluation data, and the data labeling module is configured to:

and carrying out data annotation on the evaluation data based on the user portrayal of the target group, the performance parameters of the intelligent driving function and the environment parameters corresponding to the evaluation data to obtain the evaluation data after the data annotation.

10. A vehicle driving evaluation device characterized by comprising:

the function evaluation module is used for inputting the performance parameters of the target intelligent driving function into a vehicle driving evaluation model so as to output and obtain the score of at least one user group on the performance parameters of the target intelligent driving function;

wherein the vehicle driving evaluation model is trained by a training device based on the vehicle driving evaluation model of any one of claims 1 to 9.

11. A training method for a vehicle driving evaluation model is characterized by comprising the following steps:

acquiring evaluation data of a driving intelligence function of a vehicle by a target group, performance parameters of the driving intelligence function, user pictures of the target group and expert evaluation results corresponding to the evaluation data; the target group at least comprises a user group, a vehicle engineer group and a vehicle expert group, and the performance parameters of the intelligent driving function are obtained by testing the intelligent driving function corresponding to the evaluation data;

performing data annotation on the evaluation data based on the user images of the target group and the performance parameters of the intelligent driving function to obtain the evaluation data after the data annotation, so that the evaluation data of the intelligent driving function corresponds to the test data of objective performance indexes;

12. A vehicle driving evaluation method characterized by comprising:

13. A computer device, comprising: a memory and a processor; wherein the memory is used for storing a computer program; the processor is coupled to the memory for executing the computer program for performing the steps of the method of any of claims 11 or 12.

14. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, causes the processor to carry out the steps of the method of one of the claims 11 or 12.