CN116956044A

CN116956044A - Automatic driving vehicle and performance evaluation method and system thereof

Info

Publication number: CN116956044A
Application number: CN202310961138.7A
Authority: CN
Inventors: 周宏图; 肖坤; 洪泽鑫
Original assignee: Beijing Dashi Particle Technology Co ltd
Current assignee: Beijing Dashi Particle Technology Co ltd
Priority date: 2023-08-01
Filing date: 2023-08-01
Publication date: 2023-10-27

Abstract

The application discloses an automatic driving vehicle and a performance evaluation method and an evaluation system thereof, and relates to the technical field of vehicle evaluation, wherein the method comprises the steps of utilizing preset acquisition equipment to acquire vehicle parameter data in an automatic driving process; preprocessing the vehicle parameter data; establishing a vehicle evaluation model based on the preprocessed data; extracting feature data in the preprocessed data according to an evaluation dimension, wherein the extracted feature data can characterize the evaluation dimension, and the vehicle evaluation model comprises at least one evaluation dimension; training the vehicle assessment model based on the feature data; testing the trained vehicle evaluation model on a test set to realize model tuning; and inputting vehicle data corresponding to any evaluation dimension in the vehicle evaluation model, wherein the driving behavior is evaluated based on a plurality of evaluation dimensions, so that the advantages and disadvantages of the automatic driving behavior can be comprehensively and objectively reflected, and a more comprehensive evaluation result is provided for a user.

Description

Automatic driving vehicle and performance evaluation method and system thereof

Technical Field

The application relates to the technical field of vehicle evaluation, in particular to an automatic driving vehicle, a performance evaluation method and an evaluation system thereof.

Background

With the development of automobile technologies, more and more automobiles have intelligent functions, such as ADAS (Advanced Driver Assistance System) intelligent driving assistance systems, internet of vehicles systems and the like. However, the prior art lacks pertinence and accuracy, cannot effectively evaluate the safety and reliability of the automatic driving automobile, and is difficult to output effective optimization suggestions, so that how to scientifically and comprehensively evaluate the intelligent function of the automobile becomes a hot problem of the current automobile technology development.

Disclosure of Invention

In order to solve the problem of inaccurate scoring of a vehicle evaluation system in the prior art, in a first aspect, an embodiment of the present application provides a performance evaluation method for an automatic driving vehicle, including:

acquiring vehicle data, and acquiring vehicle parameter data in an automatic driving process by using preset acquisition equipment;

preprocessing data, namely preprocessing the vehicle parameter data;

modeling, establishing a vehicle assessment model based on preprocessed data, wherein the preprocessed data is proportionally used as a training set and a testing set of the vehicle assessment model;

extracting data, extracting feature data in the preprocessed data according to an evaluation dimension, wherein the extracted feature data can represent the evaluation dimension, and the vehicle evaluation model comprises at least one evaluation dimension;

training a model, based on the characteristic data and human driving data, to train the vehicle assessment model;

model tuning, namely testing the trained vehicle evaluation model on a test set to realize model tuning; and

and (3) vehicle evaluation, namely inputting vehicle data corresponding to any evaluation dimension into the vehicle evaluation model according to the set evaluation dimension, and representing the advantages and disadvantages of the evaluation dimension by an evaluation result output by the model.

In one embodiment of the present application, in the step of acquiring vehicle data, the vehicle parameter data includes one or more of a vehicle speed, an acceleration, a braking force, a head angle, a GPS position, and a lane line position.

In one embodiment of the application, the acquisition device for acquiring data comprises one or more of millimeter wave radar, lidar, sensor, camera and GPS.

In one embodiment of the present application, the data preprocessing step includes data cleansing, data smoothing, and data conversion.

In one embodiment of the application, the assessment dimension includes driving behavior, vehicle safety, vehicle performance, and environmental adaptation capability.

In an embodiment of the present application, in the step of extracting data, the preprocessed data is extracted based on a statistical method, a time-frequency analysis method, or a machine learning algorithm.

In one embodiment of the application, the algorithm that trains the model employs linear regression, logistic regression, support vector machines, decision trees, or random forests.

In one embodiment of the present application, the step of evaluating the data includes the step of obtaining the evaluation result by the test set data according to an evaluation algorithm;

wherein the evaluation algorithm includes cross-validation, ROC validation, precision, recall, and F1-score.

In one embodiment of the application, the vehicle assessment model is built based on a neural network algorithm, a support vector algorithm, or a random forest algorithm.

In one embodiment of the present application, the step of tuning the model further includes: and optimizing the vehicle evaluation model according to the test set data and the user feedback data in the vehicle evaluation model.

In a second aspect, the present application provides a performance evaluation system of an autonomous vehicle, comprising:

the computer data acquisition module comprises acquisition equipment arranged on a vehicle to be evaluated, wherein the acquisition equipment is used for acquiring vehicle state information, and the vehicle state information is divided into a plurality of groups according to different evaluation dimensions;

the algorithm module can evaluate the vehicle to be evaluated according to the input state information during running and generate evaluation information of corresponding evaluation dimensions;

and the user side generates and presents an evaluation report according to the evaluation information.

In one embodiment of the present application, the acquisition device includes one or more of a radar, a sensor, a camera, and a GPS, and the acquisition device is capable of acquiring one or more of a distance between the vehicle to be evaluated and a lane line or an environment or other vehicles, a vehicle offset, a vehicle traveling direction, and a vehicle head angle.

In one embodiment of the application, the computer data acquisition module and/or the algorithm module is/are configured with a vehicle assessment model for identifying, measuring and assessing objects in the surroundings of a vehicle to be assessed;

the computer data acquisition module and/or the algorithm module can input pre-acquired vehicle data to train the vehicle assessment model.

In one embodiment of the present application, the computer data acquisition module and/or the algorithm module is connected with:

one or more of GPU units, distributed computer ports, cloud computing ports.

In one embodiment of the present application, the system further includes a storage module, configured to store the evaluation information, and configured to output the evaluation information to the client in response to a request from the client;

the storage module can store assessment information including assessment time information and assessment scenario information, the assessment scenario information including one or more of lane line width information, geographic location information, altitude information, grade information.

In a third aspect, the present application provides an autonomous vehicle equipped with an intelligent driving system and with any one of the above-described autonomous vehicle performance evaluation systems, the performance evaluation system performing an evaluation using any one of the above-described evaluation methods.

The beneficial effects of the application are as follows:

1. the application evaluates the driving behavior based on a plurality of evaluation dimensions, including but not limited to safety, fuel economy and comfort, can comprehensively and objectively reflect the advantages and disadvantages of the automatic driving behavior, and provides more comprehensive vehicle driving evaluation results for users;

2. the accuracy and the stability of the automatic driving system are improved through training the model and optimizing the model, the performance of the automatic driving vehicle can be more accurately estimated, and the safety and the stability are effectively improved;

3. during the evaluation, all human driving data are recorded and marked in view of the collection of human driving data. These data will provide an important reference for us to help the vehicle assessment model more accurately assess the performance of the autonomous vehicle;

4. and analyzing various performance indexes of the automatic driving automobile through the constructed automatic driving performance evaluation model, and outputting targeted optimization suggestions so as to improve the performance and safety of the automatic driving automobile.

Drawings

The accompanying drawings are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification. The drawings illustrate embodiments of the application and, together with the description, serve to explain the principles of the application.

Fig. 1 shows a flowchart provided by an embodiment of the present application.

Detailed Description

The present application now will be described more fully hereinafter with reference to the accompanying drawings, in which embodiments are shown. The application may, however, be embodied in many different forms and should not be construed as being limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the application to those skilled in the art. Like numbers refer to like elements throughout. Also, in the drawings, the thickness, ratio, and size of the parts are exaggerated for clarity of illustration.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting. As used herein, unless the context clearly indicates otherwise, "a," "an," "the," and "at least one" are not meant to limit the amount, but are intended to include both the singular and the plural. For example, unless the context clearly indicates otherwise, the meaning of "a component" is the same as "at least one component". The "at least one" should not be construed as limited to the number "one". "or" means "and/or". The term "and/or" includes any and all combinations of one or more of the associated listed items.

Unless otherwise defined, all terms used herein, including technical and scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art. Terms as defined in commonly used dictionaries should be interpreted as having the same meaning as the relevant art context and are not interpreted in an idealized or overly formal sense unless expressly so defined herein.

The meaning of "comprising" or "including" indicates a property, quantity, step, operation, component, element, or combination thereof, but does not preclude other properties, quantities, steps, operations, components, elements, or combinations thereof.

Embodiments are described herein with reference to cross-sectional illustrations that are idealized embodiments. Thus, variations from the shapes of the illustrations as a result, for example, of manufacturing techniques and/or tolerances, are to be expected. Thus, embodiments described herein should not be construed as limited to the particular shapes of regions as illustrated herein but are to include deviations in shapes that result, for example, from manufacturing. For example, a region shown or described as being flat may typically have rough and/or nonlinear features. Also, the acute angles shown may be rounded. Thus, the regions illustrated in the figures are schematic in nature and their shapes are not intended to illustrate the precise shape of a region and are not intended to limit the scope of the claims.

Various automatic driving automobile performance evaluation tools and algorithms exist in the market at present, but the algorithms often adopt simple statistical models or traditional machine learning methods, and the performance of the automatic driving automobile is difficult to evaluate accurately for complex scenes and driving conditions of the automatic driving automobile.

In view of the problems of low test efficiency, high cost, incomplete evaluation result and the like existing in the prior art, which depend on manual test and actual drive test, the applicant provides a method and a device for evaluating the automatic driving function of a vehicle based on a learning model. This approach is based on the following overall logical structure:

step 1: identifying and tracking vehicle information, surrounding object information and environment information by using equipment;

step 2: analyzing and evaluating behaviors of the vehicle in the automatic driving process by using a model and a deep learning algorithm, including but not limited to evaluation dimensions of driving behaviors, vehicle safety and the like;

in the evaluation process, the difference between the behavior of the autonomous vehicle and the human driving behavior is compared by using a series of metrics. These metrics include: "STH" for comparing whether or not the speed of autopilot is 5.0 or more slower than that of human drive at the same location; "PLTH" is used to compare whether the projection of the autopilot position on the human driving trajectory is less than 5.0 at the same point in time.

The present disclosure does not limit the evaluation criteria of the above metric, and the above "5.0" is exemplary, and the scoring values of the metrics such as speed, trajectory projection, and the like.

Furthermore, the model is also able to check for differences between the human driving trajectory, the old version of the re-simulated trajectory and the new version of the re-simulated trajectory, using a statistical method in which each simulated segment is counted only once, and grouping consecutive events.

Step 3: and generating an evaluation report according to the analysis result, and feeding back to the client.

Based on the technical ideas, a corresponding device can be designed, the basic configuration of the device comprises a computer acquisition module, a deep learning module and a feedback module, wherein the computer acquisition module can identify and collect vehicle information, the deep learning module is used for analyzing and evaluating behaviors in the running process of the vehicle, and the feedback module is used for generating an evaluation report according to an analysis result and feeding the evaluation report back to a user.

The above-described device can also be desirably provided on a vehicle having an autopilot function.

The present application is described below with reference to the drawings in the embodiments of the present application, and the inventor proposes a performance evaluation method and system for an automatic driving vehicle, and an automatic driving vehicle, so as to accurately and pertinently evaluate running scores of an automatic driving vehicle.

Hereinafter, exemplary embodiments according to the present application will be described with reference to the accompanying drawings.

In a first aspect, referring to fig. 1, the present application proposes a performance evaluation method of an autonomous vehicle, comprising the steps of:

acquiring vehicle data, wherein corresponding data are acquired through acquisition equipment preset in and/or outside the vehicle in the automatic driving process of the vehicle so as to acquire parameter data in the automatic driving process of the vehicle;

and preprocessing the data, namely preprocessing the vehicle parameter data, removing redundant data, and normalizing the data. The data collection stage is used for collecting a large amount of natural driving behavior data, and the data are processed, including cleaning, standardization and feature extraction, so that a foundation is provided for subsequent model training and evaluation;

modeling, namely establishing a vehicle evaluation model based on the preprocessed data, and proportionally taking the preprocessed data as training set data and test set data of the vehicle evaluation model;

and extracting data, namely extracting feature data corresponding to the evaluation dimension in the training set data according to the evaluation dimension required to be evaluated by the vehicle, wherein the set of the extracted feature data can represent the advantages and disadvantages of the evaluation dimension. Wherein, due to the difference of the selected evaluation dimensions, the extracted characteristic data may be different;

training a model, namely training a vehicle assessment model based on the extracted characteristic data and the human driving data; the human driving data can be collected in the non-automatic driving process, and can also be non-automatic driving data provided by other users;

model tuning, namely inputting data corresponding to the selected evaluation dimension in the test set data into a vehicle evaluation model for testing so as to realize model tuning; and

and (3) vehicle evaluation, namely inputting vehicle data collected in real time into a vehicle evaluation model with optimized vehicle according to the selected evaluation dimension, wherein an output result of the model represents the quality of the evaluation dimension corresponding to the vehicle.

In one embodiment, the vehicle data acquired by the collected data may include one or more of vehicle acceleration, braking force, head angle, GPS position, and lane line position. The vehicle data is not limited to the above data, but may include data capable of characterizing vehicle parameter information such as a braking distance, a steering angle, a vehicle offset distance, and the like.

In one embodiment, the acquisition device includes one or more of millimeter wave radar, lidar, sensors, cameras, and GPS. For example, a vehicle detects a distance between the vehicle and an adjacent vehicle using a sensor, acquires position information of the vehicle by GPS, and acquires a lane line position by a camera.

In one embodiment, the advantages of autopilot data over human drive data can be evaluated by comparing the mean and variance of the human drive model and the autopilot model. The training effect of the model can be optimized by taking a plurality of standard deviations (for example, 4) from the human driving model, and counting the times of rapid acceleration and rapid deceleration in the current journey and the accumulated journey, the mileage and the time.

The data preprocessing step specifically may include data cleaning, data smoothing and data conversion, where the data cleaning can clean redundant information collected in the collecting device, and the data smoothing can eliminate noise or irregularity of the data, so that the data is easier to analyze and understand, and the data conversion can convert the data from one form to another form, so that normalization of the collected data is performed, and thereby reliability and consistency of the data can be achieved through the data preprocessing step.

Specifically, the vehicle data may be uploaded to a data center to perform a preprocessing operation.

In the step of extracting data according to the evaluation dimension, in one embodiment, a decision description architecture of the autonomous vehicle is used. In this decision architecture, the behavior of an autonomous vehicle is divided into a plurality of discrete states, each state corresponding to a particular driving mode, such as driving, parking, lane changing, etc. Each state has one or more continuous kinetic models describing the behavior of the vehicle in that state. Transitions between states are controlled by a discrete set of hopping rules that define under what conditions a vehicle may transition from one state to another.

The advantage of this architecture is that it can clearly describe the complex behaviour of an autonomous vehicle, well assessing the various driving modes of an autonomous vehicle in complex traffic environments.

In one embodiment, the autopilot model is configured with reference to an acceleration profile model of human driving behavior data. Corresponding acceleration distribution models are built or updated according to different vehicle types, and meanwhile, an overall model without distinguishing the vehicle types is built. These models may be used to evaluate stability and comfort of the autopilot algorithm.

In one embodiment, the assessment dimension may include one or more of driving behavior, vehicle safety, vehicle performance, and environmental adaptation capability.

In another embodiment, the evaluation dimension may further include battery attenuation, safety, comfort, etc. of the vehicle, whereby the set of evaluation dimensions can comprehensively and objectively reflect the advantages or disadvantages of driving behavior.

For the evaluation dimension, specifically, for example, a lane keeping task is selected, and the feature data corresponding to the task may include data of a distance between the vehicle and the lane line, a vehicle speed, an angle at which the vehicle deviates from the lane line, and the like.

In one embodiment, the vehicle data may be extracted based on a statistical method, a time-frequency analysis method, or a learning algorithm thereof.

In one embodiment, the algorithm used to train the model may employ linear regression, logistic regression, support vector machines, decision trees, or random forests.

In one embodiment, in the step of evaluating the data, the data of the test set is calculated according to an evaluation algorithm to obtain an evaluation result. In one particular embodiment, the evaluation algorithm employs cross-validation, ROC validation, precision, recall, or F1-score.

In one embodiment, the vehicle assessment model is built based on a neural network algorithm, a support vector algorithm, or a random forest algorithm.

In one embodiment, in the step of model tuning, the vehicle assessment model may be further optimized through user feedback data, where the user feedback data includes user labeling information, and the information may be, for example, slow or fast speed, fast or slow acceleration, long or short braking distance, and other visual feeling data capable of illustrating the user in the automatic driving process of the vehicle.

the computer data acquisition module comprises at least one acquisition device arranged on the vehicle to be evaluated, wherein the acquisition device is used for acquiring vehicle state information when the vehicle is in an automatic driving state, and the state information can be divided into a plurality of groups according to a preset evaluation dimension;

the algorithm module is provided with a vehicle scoring model which is trained and optimized, and can evaluate the vehicle to be evaluated according to the input vehicle state information when the algorithm module operates, so as to generate evaluation information corresponding to the evaluation dimension;

In a third aspect, the present application proposes an autonomous vehicle equipped with an autonomous system, the vehicle being capable of autonomous driving, and at the same time equipped with a performance evaluation system of the autonomous vehicle proposed in any of the above embodiments.

Example 1

In this example, the vehicle evaluation model uses a neural network model.

In the aspect of constructing a model data set, the method can acquire vehicle driving data under different environments in various modes, for example, real driving data is acquired through an on-board sensor, and virtual driving data is acquired through a virtual simulation platform. These data are mixed together to construct an overall dataset, and each item of data in the dataset is then annotated for training and testing.

In order to ensure the accuracy and reliability of the evaluation result, a statistical analysis and multiple repeated tests are required to be performed on the evaluation process, for example, the statistical analysis process uses all the acquired data to count the mean value and standard deviation thereof, when the acquired data exceeds the range of 4 standard deviations of the mean value of the data type, the data is considered to have abnormality, and in order to alleviate the influence of accidental data in a single test, the accuracy and reliability of the result are improved by using a method of multiple repeated tests. In addition, the reliability of the evaluation result can be further improved by the technologies such as cross-validation and the like.

Example two

The embodiment discloses an evaluation system, which evaluates the performance of an intelligent driving function of a vehicle by analyzing images and data in the running process of the vehicle and provides support for research and development and test of the intelligent vehicle.

The system comprises the following components:

the computer data acquisition module comprises a vehicle-mounted camera and a sensor: for acquiring environmental images and data around the vehicle, such as lane lines, vehicle speed, distance of the vehicle from the object, etc.;

vehicle data processing algorithm: analyzing and processing images and data acquired by the vehicle-mounted camera and data acquired by the sensor for subsequent automatic driving function evaluation;

automatic driving function evaluation algorithm: the algorithm is realized by adopting a Python language, and is tested and verified based on the existing public data set and test scene. Specific test scenarios may include urban roads, highways, tunnels, etc. The data collected in the test comprises a plurality of indexes such as vehicle speed, acceleration, braking distance, steering angle, vehicle offset distance, user feedback information and the like, and the indexes are used for evaluating the performance of the algorithm;

user side with result display interface: and feeding the evaluation result back to a user or a tester through a display interface for evaluating the performance of the intelligent driving function of the vehicle.

Example III

The present example provides a vehicle that can collect data and evaluate in real time during the vehicle's autonomous driving process, wherein the user can choose a single evaluation dimension, or choose multiple dimensions, or choose the vehicle as a whole to evaluate. After the user knows the vehicle score, the user can evaluate the selected evaluation dimension or the vehicle data in the evaluation dimension, so that the model score is more accurate.

The remote end connected to the vehicle can also obtain the data of the vehicle and its score and evaluate the data.

In summary, the learning model provided by the embodiment of the application evaluates the driving behavior based on a plurality of evaluation dimensions, including but not limited to safety, fuel economy and comfort, and can comprehensively and objectively reflect the advantages and disadvantages of the automatic driving behavior, thereby providing a more comprehensive vehicle driving evaluation result for the user;

according to the application, the accuracy and the stability of the automatic driving system are improved through training and model tuning, the performance of the automatic driving vehicle can be more accurately evaluated, and the safety and the stability are effectively improved;

during the evaluation, all human driving data are recorded and marked in view of the collection of human driving data. These data will provide an important reference for us to help the vehicle assessment model more accurately assess the performance of the autonomous vehicle;

according to the application, through the constructed automatic driving performance evaluation model, various performance indexes of the automatic driving automobile are analyzed, and a targeted optimization suggestion is output, so that the performance and the safety of the automatic driving automobile are improved.

In addition, embodiments of the present disclosure may be embodied in the following forms: complete hardware, complete software (including firmware, resident software, micro-code, etc.), a combination of hardware and software. Furthermore, in some embodiments, the disclosed embodiments may also be implemented in the form of a computer program product in one or more computer-readable storage media, which contain computer program code.

Any combination of one or more computer-readable storage media may be employed by the computer-readable storage media described above. The computer-readable storage medium includes: an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of the computer readable storage medium include the following: portable computer diskette, hard disk, random Access Memory (RAM), read-only Memory (ROM), erasable programmable read-only Memory (EPROM), flash Memory (Flash Memory), optical fiber, compact disc read-only Memory (CD-ROM), optical storage device, magnetic storage device, or any combination thereof. In the disclosed embodiments, a computer-readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, device.

The computer program code embodied in the computer readable storage medium may be transmitted using any appropriate medium, including: wireless, wire, fiber optic cable, radio Frequency (RF), or any suitable combination thereof.

Computer program code for carrying out operations of embodiments of the present disclosure may be written in assembly instructions, instruction Set Architecture (ISA) instructions, machine-related instructions, microcode, firmware instructions, state setting data, integrated circuit configuration data, or in one or more programming languages, including an object oriented programming language such as: java, smalltalk, C ++, also include conventional procedural programming languages, such as: c language or similar programming language. The computer program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of remote computers, the remote computers may be connected via any sort of network, including: a Local Area Network (LAN) or a Wide Area Network (WAN), which may be connected to the user's computer or to an external computer.

The embodiments of the present disclosure describe the provided methods, apparatuses, and electronic devices through flowcharts and/or block diagrams.

It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer-readable program instructions. These computer-readable program instructions may be provided to a processor of a general purpose computer, special purpose computer, 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/acts specified in the flowchart and/or block diagram block or blocks.

These computer readable program instructions may also be stored in a computer readable storage medium that can cause a computer or other programmable data processing apparatus to function in a particular manner. Thus, instructions stored in a computer-readable storage medium produce an instruction means which implement the functions/acts specified in the flowchart and/or block diagram block or blocks.

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

The foregoing is merely a specific implementation of the embodiment of the present application, but the protection scope of the embodiment of the present application is not limited thereto, and any person skilled in the art can easily think about changes or substitutions within the technical scope of the embodiment of the present application, and the changes or substitutions are covered by the protection scope of the embodiment of the present application. Therefore, the protection scope of the embodiments of the present application shall be subject to the protection scope of the claims.

Claims

1. A performance evaluation method of an autonomous vehicle, comprising:

preprocessing data, namely preprocessing the vehicle parameter data;

2. The method of evaluating the performance of an autonomous vehicle according to claim 1, wherein in the step of acquiring vehicle data, the vehicle parameter data includes one or more of a vehicle speed, an acceleration, a braking force, a head angle, a GPS position, and a lane line position.

3. The method of evaluating performance of an autonomous vehicle according to claim 1, wherein the acquisition device for acquiring data includes one or more of millimeter wave radar, lidar, sensor, camera, and GPS.

4. The method of evaluating the performance of an autonomous vehicle of claim 1, wherein the data preprocessing step includes data cleansing, data smoothing, and data conversion.

5. The method of claim 1, wherein the evaluation dimension includes driving behavior, vehicle safety, vehicle performance, and environmental adaptation capability.

6. The method according to claim 5, wherein in the step of extracting data, the preprocessed data is extracted based on a statistical method, a time-frequency analysis method, or a machine learning algorithm.

7. The method of claim 1, wherein the algorithm for training the model employs linear regression, logistic regression, support vector machine, decision tree, or random forest.

8. The method for evaluating the performance of an autonomous vehicle according to claim 1, wherein the step of evaluating the data includes the test set data deriving the evaluation result according to an evaluation algorithm;

9. The method of claim 1, wherein the vehicle assessment model is built based on a neural network algorithm, a support vector algorithm, or a random forest algorithm.

10. The method for evaluating the performance of an autonomous vehicle according to claim 1, wherein the step of tuning the model further comprises: and optimizing the vehicle evaluation model according to the test set data and the user feedback data in the vehicle evaluation model.

11. A performance evaluation system of an autonomous vehicle, comprising:

12. The automated driving vehicle performance evaluation system of claim 11, wherein the acquisition device comprises one or more of radar, sensor, camera, and GPS, the acquisition device being capable of acquiring one or more of distance between the vehicle under evaluation and lane line or environment or other vehicle, vehicle offset, vehicle direction of travel, and head angle.

13. The automated driving vehicle performance evaluation system of claim 11, wherein the computer data acquisition module and/or the algorithm module is configured with a vehicle evaluation model for identifying, measuring, and evaluating objects in the surrounding environment of a vehicle to be evaluated;

14. The automated driving vehicle performance evaluation system of claim 11, wherein the computer data acquisition module and/or the algorithm module is connected with:

one or more of GPU units, distributed computer ports, cloud computing ports.

15. The automated driving vehicle performance evaluation system of claim 11, further comprising a storage module capable of storing the evaluation information and capable of outputting the evaluation information to the user side in response to a request from the user side;

16. An autonomous vehicle provided with an intelligent driving system, characterized in that the vehicle is provided with a performance evaluation system of an autonomous vehicle as claimed in any of claims 11 to 15.