Driving risk scene construction and data acquisition method, system, equipment and medium

Abstract

The invention relates to a driving risk scene construction and data acquisition method, system and device and a medium, and the method comprises the following steps: carrying out the feature calibration of real traffic accident data, and carrying out the processing to obtain an accident data comparison analysis result; screening the calibrated traffic accident data based on a preset condition, and performing clustering division processing on the screened traffic accident data to obtain a typical driving risk scene; according to the accident data comparison analysis result and the typical driving risk scene, constructing a driving risk scene based on the simulated driving environment; and according to the constructed driving risk scene, carrying out a driving simulation experiment, and collecting physiological state, psychological state and driving behavior data of a driver in each driving risk scene. The method can be widely applied to the field of driving safety evaluation.

Description

Technical Field

[0001] The present invention belongs to the field of driving safety assessment, and specifically relates to a driving risk scenario construction and data collection method, system, equipment and medium based on a simulated driving environment. Background Art

[0002] According to statistics, road traffic accidents caused by drivers' risky behavior habits, abnormal driving conditions or improper operations account for more than 90% of all road traffic accidents. Therefore, the safety of drivers is the core of ensuring road traffic safety.

[0003] At present, a key driving safety assessment method is driven by driver state behavior data. This method evaluates driving safety by extracting different physiological and psychological states and driving behavior characteristics of drivers in different driving environments. Therefore, the core challenge of the safety assessment model and method driven by driver state behavior data is to obtain the driver's real physiological state, psychological state and driving behavior in driving risk scenarios.

[0004] With the development of technologies such as deep learning and the improvement of data acquisition equipment performance, data-driven driving safety assessment methods have also received widespread attention and rapid development, and many risk accident data sets and related assessment methods have been born. However, existing risk accident data sets have significant limitations. For example, data sets such as DADA collect dashcam data from the accident vehicle or surrounding vehicles before and after the accident, and then perform subsequent data calibration methods to simulate the subjective state and behavior data of the driver when the accident occurs. However, due to information loss and other reasons, this method is difficult to completely restore the driver's true physiological state, psychological state and driving behavior when the accident occurs. From an ethical point of view, it is also impossible to design natural driving risk scenarios to collect real risk accident data. Therefore, the construction of driving risk scenarios and accident data collection methods based on simulated driving environments are of great significance for accident data-driven driving safety assessment. Summary of the invention

[0005] In view of the above problems, the purpose of the present invention is to provide a driving risk scenario construction and data collection method, system, equipment and medium, which is guided by traffic safety and based on real traffic accident data and simulated driving environment to construct and generate risk accident driving scenarios, thereby collecting the driver's real physiological state, psychological state and driving behavior data when a risk accident occurs.

[0006] To achieve the above object, the present invention adopts the following technical solutions: In a first aspect, the present invention provides a method for constructing a driving risk scenario and collecting data, comprising the following steps: Perform feature calibration on real traffic accident data and process it to obtain comparative analysis results of accident data; The calibrated traffic accident data is screened based on preset conditions, and the screened traffic accident data is clustered to obtain typical driving risk scenarios; Based on the comparative analysis results of accident data and typical driving risk scenarios, a driving risk scenario based on a simulated driving environment is constructed; A simulated driving experiment is carried out based on the constructed driving risk scenarios, and the driver's physiological state, psychological state and driving behavior data are collected under each driving risk scenario.

[0007] Furthermore, the feature calibration of the real traffic accident data and processing to obtain the accident data comparison and analysis results include: Compare and analyze the types of collision objects, collision locations, and differences in the driving trajectories of both parties before the collision, and obtain accident data feature results in a unified format and standard; Calibrate the obtained accident data characteristic results according to the preset calibration rules; The calibrated accident data features are summarized and used to calculate the Hamming distance between driving risk scenarios.

[0008] Furthermore, the calibrated traffic accident data is screened based on preset conditions, and the screened traffic accident data is clustered to obtain typical driving risk scenarios, including: Screening and filtering traffic accident data in which at least one of the conflicting parties is a motor vehicle; The K-Medoids clustering method based on Hamming distance is used to cluster the screened traffic accident data and obtain typical driving risk scenarios.

[0009] Furthermore, the K-Medoids clustering method based on Hamming distance is used to perform clustering and division processing on the filtered traffic accident data to obtain typical driving risk scenarios, including: ① Take the filtered traffic accident data as the sample set and randomly determine k samples as cluster centers; ② Calculate the Hamming distance between each sample and the k cluster centers, and classify each sample into the cluster with the smallest Hamming distance; ③ Calculate the sum of the distances from all samples in each cluster to the cluster center among the k clusters; ④ In the k clusters, select each sample as the cluster center in turn, and calculate the sum of the distances from the samples in the cluster to the cluster center, and determine whether it is less than the sum of the distances calculated in step ③. If it is less, update the sample point to the new cluster center; ⑤ Repeat steps ② to ④, update the cluster centers until the k cluster centers do not change, and obtain k typical driving risk scenarios.

[0010] Furthermore, the construction of a driving risk scenario based on a simulated driving environment based on the accident data comparison and analysis results and typical driving risk scenarios includes: The experimental map and experimental vehicles are selected according to actual needs, and each experimental vehicle is deployed to the experimental map; wherein the experimental vehicles include the test vehicle and several other vehicles; Based on the typical driving risk scenarios obtained through clustering, a risk triggering method is designed to construct a risk accident driving scenario based on a simulated driving environment.

[0011] Furthermore, the risk triggering method is designed according to each typical driving risk scenario obtained by clustering, and a risk accident driving scenario based on a simulated driving environment is constructed, including: Real-time monitoring of other vehicles within a preset radius around the test vehicle; When other vehicles and the test vehicle meet the preset position and trajectory relationship conditions, other vehicles are randomly deviated from the original driving trajectory and perform preset risk behaviors; After other vehicles complete the risky behavior, they continue to perform normal autonomous driving behavior in compliance with traffic regulations.

[0012] Furthermore, the simulated driving experiment is carried out based on the constructed driving risk scenario, and the driver's physiological and psychological state and driving behavior data under the driving risk scenario in the simulated driving environment are collected, including: By simulating the cockpit, the accelerator, brake, gear, and steering wheel are mapped to the driving simulation software. The driver performs driving behaviors by operating the simulated cockpit, and the driver's behavior operations in the simulated cockpit are recorded in real time; Through driving simulation software, the driving data of the test vehicle and other vehicles in the simulated driving scene are recorded, including three-axis speed, acceleration, angular velocity, and position information; Use a desktop or glasses-type eye tracker to record the driver's attention and line of sight changes during driving and obtain the driver's attention distribution; Record the driver's heart rate, pulse, skin electricity, and respiratory physiological data during driving through physiological instruments and watches; The driver's physiological and psychological state before and after driving was recorded by questionnaire. Through near-infrared brain functional imaging and electroencephalogram equipment, the driver's brain data during driving is recorded; Through the camera, the driver's expression and posture during driving can be captured in real time, and the sound during driving can be recorded.

[0013] In a second aspect, the present invention provides a driving risk scenario construction and data collection system, comprising: The data preprocessing module is used to calibrate the features of real traffic accident data and process them to obtain comparative analysis results of the accident data; A clustering module is used to filter the calibrated traffic accident data based on preset conditions, and to perform clustering and division processing on the filtered traffic accident data to obtain typical driving risk scenarios; The scenario construction module is used to construct driving risk scenarios based on simulated driving environments according to the comparative analysis results of accident data and typical driving risk scenarios; The data collection module is used to carry out simulated driving experiments based on the constructed driving risk scenarios, and to collect the driver's physiological state, psychological state and driving behavior data under each driving risk scenario.

[0014] In a third aspect, the present invention provides a computer-readable storage medium storing one or more programs, wherein the one or more programs include instructions, which, when executed by a computing device, cause the computing device to perform any method.

[0015] In a fourth aspect, the present invention provides a computing device, comprising: one or more processors and a memory, wherein the memory stores one or more programs and is configured to be executed by the one or more processors, and the one or more programs include instructions for executing any method.

[0016] The present invention adopts the above technical solution, which has the following advantages: 1. The present invention divides and constructs driving risk scenarios based on simulated driving environments by deeply analyzing natural driving accident data, and then collects driving behavior data of drivers' physiological and psychological states when accidents occur. This method makes full use of real natural traffic accidents, summarizes the characteristics behind different types of traffic accidents through qualitative and quantitative comparative analysis, and reveals their distribution patterns and differences.

[0017] 2. The present invention uses the K-Medoids clustering method based on Hamming distance to cluster and divide driving risk scenarios, which can identify accident subsets with common characteristics from complex traffic accident data, thereby constructing representative driving risk scenarios. Compared with the existing research approach of selecting several representative scenarios based on experience, the method of constructing driving risk scenarios by calibrating and dividing real traffic accident data is more reliable and reasonable. It can not only construct driving risk scenarios with wider coverage and stronger pertinence, but also provide strong data to illustrate the necessity and importance of the constructed scenarios.

[0018] 3. When a risk accident occurs, the present invention provides scenario and data support for the driving safety assessment method driven by the driver's state and behavior data based on the driver's physiological and psychological state and driving behavior data.

[0019] The scenarios constructed and data collected by the present invention fill the gap in the current research on the subjective physiological and psychological state and driving behavior data of drivers when risk accidents occur, making the driving safety assessment method driven by driver state behavior data more scientific and reasonable, thereby reducing or even avoiding the occurrence of traffic accidents caused by driver factors and improving the level of road traffic safety. Therefore, the present invention can be widely used in the field of driving safety assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Throughout the accompanying drawings, the same reference numerals are used to represent the same components. In the accompanying drawings: Figure 1 is a flow chart of a method for constructing a driving risk scenario and collecting accident data provided by an embodiment of the present invention; Figure 2 It is a flow chart of a K-Medoids clustering implementation algorithm based on Hamming distance provided by an embodiment of the present invention; Figure 3 This is a typical driving risk scenario division result obtained through K-Medoids clustering provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the described embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention.

[0022] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0023] There are many factors and features that need to be considered when constructing a driving risk scenario (such as the type and status of the conflicting parties, the location of the conflict, the location of the conflict, the trajectories of the two parties before the conflict, etc.). Generally, the corresponding features are determined according to different research needs, and the clustering method is used to divide and construct them according to the determined features.

[0024] In some embodiments of the present invention, a driving risk scenario construction and data collection method is provided. The method performs feature calibration and comparative analysis on real natural driving accident data, performs K-Medoids clustering on driving risk scenarios according to accident characteristics, and then constructs a driving risk scenario based on a simulated driving environment based on the clustering results, thereby performing accident data collection and driving safety assessment work, providing scenario and data support for a driving safety assessment method driven by driver status behavior data.

[0025] Correspondingly, in other embodiments of the present invention, a driving risk scenario construction and data collection system, device and medium are provided.

[0026] Example 1 like Figure 1 As shown, the present invention provides a driving risk scenario construction and data collection method, which includes the following steps: Step 1: calibrate the features of real traffic accident data and process them to obtain comparative analysis results of the accident data; Step 2: Screen the calibrated traffic accident data based on preset conditions, and perform clustering on the screened traffic accident data to obtain typical driving risk scenarios; Step 3: Based on the accident data comparison and analysis results in step 1 and the typical driving risk scenarios in step 2, a driving risk scenario based on a simulated driving environment is constructed; Step 4: Carry out corresponding simulated driving experiments based on the constructed driving risk scenarios, and collect data on the driver's physiological state, psychological state and driving behavior under each driving risk scenario.

[0027] Furthermore, in the above step 1, the feature calibration of the real traffic accident data is performed, and the accident data comparison and analysis results are obtained by processing, which includes the following steps: Step 1.1: Compare and analyze the differences in the types of collision objects, collision locations, and driving trajectories of both parties before the collision, and obtain the accident data feature results in a unified format and standard; Step 1.2: Calibrate the obtained accident data characteristic results according to the preset calibration rules; Step 1.3: Summarize the calibrated accident data features to calculate the Hamming distance between driving risk scenarios for subsequent driving risk scenario clustering.

[0028] Furthermore, in the above step 2, the calibrated traffic accident data is screened based on preset conditions, and the screened traffic accident data is clustered and divided, including the following steps: Step 2.1: Filter the accident traffic data in which at least one of the conflicting parties is a motor vehicle; Step 2.2: Use the K-Medoids clustering method based on Hamming distance to cluster the filtered accident traffic data and obtain typical driving risk scenarios.

[0029] Furthermore, in the above step 2.2, if Figure 2 As shown in the figure, for the processed and feature-calibrated accident data, the K-Medoids clustering method is used to perform cluster analysis on the accident scenes, thereby dividing the typical accident scenes. K-Medoids clustering is a variant of K-Means clustering, which is suitable for processing discrete and enumerated categorical variables, is less affected by outliers, and the clustering results are more stable. Its core improvements over K-Means clustering are: calculating the center of the cluster instead of the mean, because the mean cannot be calculated for categorical variables; using a new distance metric.

[0030] K-Medoids clustering requires the selection of clustering variables, the selection of the number of clusters, and the final clustering operation. Specifically, it includes the following steps: Step 2.2.1: Take the filtered traffic accident data as the sample set and randomly determine k samples as cluster centers; Step 2.2.2: Calculate the Hamming distance between each sample and the k cluster centers, and classify each sample into the cluster with the smallest Hamming distance; Step 2.2.3: Calculate the sum of the distances from all samples in each cluster to the cluster center in the k clusters; Step 2.2.4: In the k clusters, select each sample as the cluster center in turn, and calculate the sum of the distances from the samples in the cluster to the cluster center, and determine whether it is less than the sum of the distances calculated in step 2.2.3. If so, update the sample point to the new cluster center; Step 2.2.5: Repeat steps 2.2.2 to 2.2.4, update the cluster centers until the k cluster centers do not change, and obtain k typical driving risk scenarios.

[0031] Furthermore, in the above step 3, based on the clustered typical driving risk scenarios, a risk accident driving scenario is constructed based on a simulated driving environment. Different from the natural driving scenario, which cannot be designed for ethical reasons, the typical driving risk scenario can be reproduced more realistically in the simulated driving environment, and data can be collected in real time, thereby obtaining more realistic and subjective data on the driver's physiological and psychological state and driving behavior under the driving risk scenario.

[0032] Specifically, the steps include: Step 3.1: Select appropriate experimental maps and experimental vehicles, and deploy each experimental vehicle to the experimental map.

[0033] In this embodiment, the experimental vehicles may include one or more manually driven vehicles (referred to as test vehicles) and several other vehicles, wherein the other vehicles may be controlled by the system or other experimenters, and this embodiment does not impose any restrictions on this; appropriate traffic flow parameters are set in the experimental map, and the experimental vehicles are deployed to the experimental map.

[0034] Step 3.2: Based on the typical driving risk scenarios obtained through clustering, design risk triggering methods and construct risk accident driving scenarios based on simulated driving environments.

[0035] Furthermore, the above step 3.2 may include the following steps: Step 3.2.1: Real-time monitoring of other vehicles within a preset radius around the test vehicle; Step 3.2.2: When other vehicles and the test vehicle meet the preset position and trajectory relationship conditions, randomly make other vehicles deviate from the original driving trajectory and perform risky behaviors such as sudden braking, lane changing, and merging; Step 3.2.3: After other vehicles complete the risky behavior, they continue to perform normal autonomous driving behavior in accordance with traffic regulations.

[0036] Furthermore, in the above step 4, corresponding experiments are carried out according to the constructed driving risk scenarios to collect the driver's physiological and psychological state and driving behavior data under the driving risk scenarios in the simulated driving environment, including the following steps: Step 4.1: By simulating the cockpit, the throttle, brake, gear, steering wheel, etc. are mapped to the driving simulation software. The driver performs driving behavior by operating the simulated cockpit, and the driver's behavior and operation in the simulated cockpit are recorded in real time; Step 4.2: Use driving simulation software to record the driving data of the test vehicle and other vehicles in the simulated driving scenario, which may include three-axis speed, acceleration, angular velocity, position information, etc. Step 4.3: Use a desktop or glasses-type eye tracker to record the driver's attention changes during driving and obtain the driver's attention distribution; Step 4.4: Record the driver's heart rate, pulse, skin electricity, respiration and other physiological data during driving by means of a physiological instrument, a wristwatch, etc.; Step 4.5: Record the driver's physiological and psychological state before and after driving through questionnaires and other methods; Step 4.6: Record the driver's brain data during driving through near-infrared brain functional imaging, electroencephalography and other equipment; Step 4.7: Use the camera to capture the driver's expression and posture during driving in real time, and record the sound during driving.

[0037] Example 2 This embodiment takes the DADA natural driving accident dataset as an example to further introduce the driving risk scenario construction and data collection method proposed in the present invention.

[0038] This embodiment uses the DADA natural driving accident data set, which is a large-scale driving attention data set for driving accidents within the driving field of view, containing 2,000 driving accident sequences, totaling 658,476 frames. The accident type classification in the DADA data set is very detailed, and the accident type is divided into 54 categories based on factors such as whether the vehicle is involved, the type of both parties in the accident, and the primary and secondary responsibilities of both parties in the accident. However, the DADA data set does not divide the actual accident scenes. In addition, the author of the DADA data set stated that half of the data will be retained for subsequent research, so there are only about 1,000 open source data.

[0039] The following embodiment will introduce the four steps of data processing and feature calibration, K-Medoids clustering, driving risk scenario construction based on simulated driving environment, and accident data collection based on simulated driving environment.

[0040] Step 1: Data processing and feature calibration The data processing process of the above-mentioned DADA natural driving accident data is roughly divided into two parts: accident screening and feature calibration.

[0041] First, we conduct comparative analysis based on the types of collision objects, collision locations, and driving trajectories of both parties before the collision, and finally obtain accident data feature results with unified format and standards. After processing, the following accident features can be extracted for each accident: types of collision parties, collision locations (road sections, intersections, etc.), collision locations, and actions of the two vehicles before the collision. During the specific implementation process, the implementers can extract more features for the accident based on the actual accident data.

[0042] Secondly, the accident data characteristics are calibrated.

[0043] Specifically, when calibrating the accident data features, for example: If both parties in the collision are motor vehicles, the calibration is 100; if both parties in the collision are a motor vehicle and a non-motor vehicle, the calibration is 010; if both parties in the collision are a motor vehicle and a pedestrian, the calibration is 001; If the collision occurs at an intersection, it is calibrated as 01; if the collision occurs at a road section, it is calibrated as 10; If the collision occurs at the front of the two collision bodies, the calibration is 001; if the collision occurs at the front of one collision body and the collision occurs at the side of the other collision body, the calibration is 010; if the collision occurs at the front of one collision body and the collision occurs at the back of the other collision body, the calibration is 100; If both collision entities are going straight when the collision occurs, the calibration is 0001; if the main vehicle is going straight and the participant is turning when the collision occurs, the calibration is 0010; if the main vehicle is turning and the participant is going straight when the collision occurs, the calibration is 0100; if both collision entities are turning when the collision occurs, the calibration is 1000.

[0044] Finally, the calibrated accident data features are summarized and used to calculate the Hamming distance between accident scenes for subsequent driving risk scene clustering.

[0045] For example, if two motor vehicles collide at an intersection, with the front of one vehicle hitting the side of the other vehicle, and both vehicles are moving straight when the accident occurs, the accident feature can be calibrated as 100010100001.

[0046] Step 2: K-Medoids Clustering like Figure 3 As shown in Figure 1, the typical driving risk scenario division results obtained by K-Medoids clustering.

[0047] Step 3: Construct driving risk scenarios based on simulated driving environments.

[0048] In this embodiment, carla driving simulation software is used to construct driving risk scenarios and perform driving simulation experiments.

[0049] Step 4: Accident data collection based on simulated driving environment This embodiment uses the SIMREX CDS 1.5 three-degree-of-freedom simulated driving cockpit of INNOSIMULATION to conduct simulated driving experiments and data collection. By collecting the driver's physiological, psychological state and driving behavior data during driving in real time, more realistic and subjective physiological and psychological state and driving behavior data of the driver in driving risk scenarios can be obtained.

[0050] Example 3 The above-mentioned embodiment 1 provides a driving risk scenario construction and data collection method. Correspondingly, this embodiment provides a driving risk scenario construction and data collection system. The system provided in this embodiment can implement the driving risk scenario construction and data collection method of embodiment 1, and the system can be implemented by software, hardware, or a combination of software and hardware. For example, the system may include integrated or separate functional modules or functional units to perform the corresponding steps in each method of embodiment 1. Since the system of this embodiment is basically similar to the method embodiment, the description process of this embodiment is relatively simple. For relevant matters, please refer to the partial description of embodiment 1. The embodiment of the system provided in this embodiment is only illustrative.

[0051] The driving risk scenario construction and data collection system provided in this embodiment includes: The data preprocessing module is used to calibrate the features of real traffic accident data and process them to obtain comparative analysis results of the accident data; A clustering module is used to filter the calibrated traffic accident data based on preset conditions, and to perform clustering and division processing on the filtered traffic accident data to obtain typical driving risk scenarios; The scenario construction module is used to construct driving risk scenarios based on simulated driving environments according to the comparative analysis results of accident data and typical driving risk scenarios; The data collection module is used to carry out simulated driving experiments based on the constructed driving risk scenarios, and to collect the driver's physiological state, psychological state and driving behavior data under each driving risk scenario.

[0052] Example 4 This embodiment provides a processing device corresponding to the driving risk scenario construction and data collection method provided in this embodiment 1. The processing device can be a processing device for a client, such as a mobile phone, a laptop computer, a tablet computer, a desktop computer, etc., to execute the method of embodiment 1.

[0053] The processing device includes a processor, a memory, a communication interface and a bus, and the processor, the memory and the communication interface are connected through the bus to complete mutual communication. The memory stores a computer program that can be run on the processor, and the processor executes the driving risk scenario construction and data collection method provided in this embodiment 1 when running the computer program.

[0054] Preferably, the memory may be a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory.

[0055] Preferably, the processor may be a central processing unit (CPU), a digital signal processor (DSP) or other general-purpose processors of various types, which are not limited here.

[0056] Example 5 The driving risk scenario construction and data collection method of this embodiment 1 can be specifically implemented as a computer program product, which may include a computer-readable storage medium on which computer-readable program instructions for executing the driving risk scenario construction and data collection method described in this embodiment 1 are loaded.

[0057] Computer readable storage media can be tangible devices that hold and store instructions used by instruction execution devices. Computer readable storage media can be, for example, but not limited to, electronic storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any combination thereof.

[0058] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks. These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 One or more processes and / or boxes Figure 1 A function specified in one or more boxes. These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0059] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A driving risk scenario construction and data collection method, characterized in that: The following steps are involved: Perform feature calibration on real traffic accident data and process it to obtain comparative analysis results of accident data; The calibrated traffic accident data is screened based on preset conditions, and the screened traffic accident data is clustered to obtain typical driving risk scenarios; Based on the comparative analysis results of accident data and typical driving risk scenarios, a driving risk scenario based on a simulated driving environment is constructed; A simulated driving experiment is carried out based on the constructed driving risk scenarios, and the driver's physiological state, psychological state and driving behavior data are collected under each driving risk scenario.

2. A driving risk scenario construction and data collection method as claimed in claim 1, characterized in that: The feature calibration of the real traffic accident data and processing to obtain the accident data comparison and analysis results include: Compare and analyze the types of collision objects, collision locations, and differences in the driving trajectories of both parties before the collision, and obtain accident data feature results in a unified format and standard; Calibrate the obtained accident data characteristic results according to the preset calibration rules; The calibrated accident data features are summarized and used to calculate the Hamming distance between driving risk scenarios.

3. A driving risk scenario construction and data collection method as claimed in claim 2, characterized in that: The calibrated traffic accident data is screened based on preset conditions, and the screened traffic accident data is clustered to obtain typical driving risk scenarios, including: Screening and filtering traffic accident data in which at least one of the conflicting parties is a motor vehicle; The K-Medoids clustering method based on Hamming distance is used to cluster the screened traffic accident data and obtain typical driving risk scenarios.

4. A driving risk scenario construction and data collection method as claimed in claim 3, characterized in that: The K-Medoids clustering method based on Hamming distance is used to cluster the filtered traffic accident data to obtain typical driving risk scenarios, including: ① Take the filtered traffic accident data as the sample set and randomly determine k samples as cluster centers; ② Calculate the Hamming distance between each sample and the k cluster centers, and classify each sample into the cluster with the smallest Hamming distance; ③ Calculate the sum of the distances from all samples in each cluster to the cluster center among the k clusters; ④ In the k clusters, select each sample as the cluster center in turn, and calculate the sum of the distances from the samples in the cluster to the cluster center, and determine whether it is less than the sum of the distances calculated in step ③. If it is less, update the sample point to the new cluster center; ⑤ Repeat steps ② to ④, update the cluster centers until the k cluster centers do not change, and obtain k typical driving risk scenarios.

5. The method for constructing a driving risk scenario and collecting data according to claim 1, characterized in that: The driving risk scenario based on the simulated driving environment is constructed based on the accident data comparison and analysis results and typical driving risk scenarios, including: The experimental map and experimental vehicles are selected according to actual needs, and each experimental vehicle is deployed to the experimental map; wherein the experimental vehicles include the test vehicle and several other vehicles; Based on the typical driving risk scenarios obtained through clustering, a risk triggering method is designed to construct a risk accident driving scenario based on a simulated driving environment.

6. A driving risk scenario construction and data collection method as claimed in claim 5, characterized in that: According to the typical driving risk scenarios obtained through clustering, a risk triggering method is designed to construct a risk accident driving scenario based on a simulated driving environment, including: Real-time monitoring of other vehicles within a preset radius around the test vehicle; When other vehicles and the test vehicle meet the preset position and trajectory relationship conditions, other vehicles are randomly deviated from the original driving trajectory and perform preset risk behaviors; After other vehicles complete the risky behavior, they continue to perform normal autonomous driving behavior in compliance with traffic regulations.

7. A driving risk scenario construction and data collection method as claimed in claim 5, characterized in that: The simulated driving experiment is carried out based on the constructed driving risk scenario, and the driver's physiological and psychological state and driving behavior data under the driving risk scenario in the simulated driving environment are collected, including: By simulating the cockpit, the accelerator, brake, gear, and steering wheel are mapped to the driving simulation software. The driver performs driving behaviors by operating the simulated cockpit, and the driver's behavior operations in the simulated cockpit are recorded in real time; Through driving simulation software, the driving data of the test vehicle and other vehicles in the simulated driving scene are recorded, including three-axis speed, acceleration, angular velocity, and position information; Use a desktop or glasses-type eye tracker to record the driver's attention and line of sight changes during driving and obtain the driver's attention distribution; Record the driver's heart rate, pulse, skin electricity, and respiratory physiological data during driving through physiological instruments and watches; The driver's physiological and psychological state before and after driving was recorded by questionnaire. Through near-infrared brain functional imaging and electroencephalogram equipment, the driver's brain data during driving is recorded; Through the camera, the driver's expression and posture during driving can be captured in real time, and the sound during driving can be recorded.

8. A driving risk scenario construction and data collection system, characterized in that: include: The data preprocessing module is used to calibrate the features of real traffic accident data and process them to obtain comparative analysis results of the accident data; A clustering module is used to filter the calibrated traffic accident data based on preset conditions, and to perform clustering and division processing on the filtered traffic accident data to obtain typical driving risk scenarios; The scenario construction module is used to construct driving risk scenarios based on simulated driving environments according to the accident data comparison and analysis results and typical driving risk scenarios; The data collection module is used to carry out simulated driving experiments based on the constructed driving risk scenarios, and to collect the driver's physiological state, psychological state and driving behavior data under each driving risk scenario.

9. A computer-readable storage medium storing one or more programs, characterized in that: The one or more programs include instructions, which, when executed by a computing device, cause the computing device to perform any one of the methods of claims 1 to 7.

10. A computing device, characterized in that: include: One or more processors and a memory, wherein the memory stores one or more programs and is configured to be executed by the one or more processors, wherein the one or more programs include instructions for executing any one of the methods described in claims 1 to 7.

Images