CN-121980951-A - Vehicle safety assessment method and device based on accident data

Abstract

The application provides a vehicle safety assessment method and device based on accident data, wherein the method comprises the steps of obtaining a historical accident data set corresponding to a vehicle historical accident of a target vehicle, and carrying out data fusion on the historical accident data set to generate target accident data corresponding to the vehicle historical accident; and simulating the risk accident scene by utilizing a simulation model corresponding to the target vehicle aiming at each risk accident scene so as to evaluate the performance of the safety configuration of the target vehicle under the risk accident scene and determine the safety configuration evaluation result of the target vehicle. By the method and the device, before a real accident occurs in a large scale, the safety performance of the vehicle safety configuration under the extreme working condition can be evaluated, and the pertinence, the foresight and the reality coverage of the safety evaluation are obviously improved.

Inventors

• LIU JUNTING
• LV YING
• LIU QIUZHENG

Assignees

• 中国第一汽车股份有限公司

Dates

Publication Date

20260505

Application Date

20260202

Claims (10)

1. 1. A vehicle safety assessment method based on accident data, characterized in that the vehicle safety assessment method comprises: Acquiring a historical accident data set corresponding to a vehicle historical accident of a target vehicle, and carrying out data fusion on the historical accident data set to generate target accident data corresponding to the vehicle historical accident; performing cluster analysis on the plurality of target accident data to identify at least one risk accident scene; And simulating the risk accident scene by using a simulation model corresponding to the target vehicle aiming at each risk accident scene so as to evaluate the performance of the safety configuration of the target vehicle in the risk accident scene and determine the safety configuration evaluation result of the target vehicle.
2. 2. The vehicle security evaluation method according to claim 1, characterized in that after determining a security configuration evaluation result of the target vehicle, the vehicle security evaluation method further comprises: And inputting the safety configuration evaluation result into a pre-constructed vehicle engineering knowledge graph corresponding to the target vehicle so as to determine a safety optimization suggestion corresponding to the safety configuration.
3. 3. The vehicle safety assessment method according to claim 1, wherein the historical accident data set includes accident scene investigation text, accident scene images, vehicle operation time sequence data and accident occurrence weather data, wherein the data fusion is performed on the historical accident data set to generate target accident data corresponding to the vehicle historical accident, and the method comprises the following steps: Carrying out named entity identification and relation extraction on the accident scene investigation text by using a large language model, and determining an accident semantic feature set corresponding to the vehicle history accident; Carrying out vehicle damage identification on the accident scene image by utilizing an image detection model, and determining a vehicle damage characteristic set corresponding to the vehicle history accident; And carrying out time window alignment on the vehicle operation time sequence data and the accident occurrence weather data, and correlating the time window alignment with the accident semantic feature set and the vehicle breakage feature set to obtain the target accident data.
4. 4. The vehicle safety assessment method according to claim 1, wherein the performing cluster analysis on the plurality of target accident data identifies at least one risk accident scenario, comprises: Converting each target accident data into corresponding high-dimensional feature vectors, and clustering and grouping each high-dimensional feature vector by using a clustering algorithm to obtain an accident feature cluster with similar accident feature combinations; for each accident feature cluster, counting the occurrence frequency of the accident feature cluster in a plurality of historical accident data sets, and determining the risk score of the accident feature cluster by combining the accident influence results of the vehicle historical accidents corresponding to the accident feature cluster; and when the risk score is greater than or equal to a preset score threshold value, taking the accident feature cluster as the risk accident scene.
5. 5. The vehicle safety assessment method according to claim 1, wherein the step of using the simulation model corresponding to the target vehicle to simulate the risk accident scenario to assess the performance of the safety configuration of the target vehicle in the risk accident scenario, and determining the safety configuration assessment result of the target vehicle comprises: Converting the risk accident scene into initial conditions corresponding to a simulation environment, calling the simulation model to perform simulation operation in the simulation environment, and determining an intervention result and an accident simulation result of the safety configuration; And randomly adjusting at least one variable in the initial condition within a preset range by utilizing a Monte Carlo simulation algorithm, executing repeated simulation for a plurality of times, and counting the effective intervention rate and the failure probability of the safety configuration in the risk accident scene so as to obtain a safety configuration evaluation result corresponding to the safety configuration.
6. 6. The vehicle safety assessment method according to claim 5, wherein said counting the effective intervention rate and failure probability of the safety arrangement in the risk accident scenario comprises: Analyzing the simulation results of multiple accidents, and determining the effective intervention times of the safety configuration started in a preset time window and realizing the control effect, and the failure times of the safety configuration started beyond the preset time window and not started beyond the preset time window; and determining the effective intervention rate based on the effective intervention times and the simulation times, and determining the failure probability based on the failure times and the simulation times.
7. 7. A vehicle safety evaluation device based on accident data, characterized in that the vehicle safety evaluation device comprises: the accident data acquisition module is used for acquiring a historical accident data set corresponding to a vehicle historical accident of a target vehicle, and carrying out data fusion on the historical accident data set to generate target accident data corresponding to the vehicle historical accident; the risk accident scene determining module is used for carrying out cluster analysis on the plurality of target accident data and identifying at least one risk accident scene; the safety evaluation module is used for simulating each risk accident scene by utilizing a simulation model corresponding to the target vehicle so as to evaluate the performance of the safety configuration of the target vehicle in the risk accident scene and determine the safety configuration evaluation result of the target vehicle.
8. 8. The vehicle safety evaluation device according to claim 7, further comprising an optimization suggestion determination module that, after determining a safety configuration evaluation result of the target vehicle, is configured to: And inputting the safety configuration evaluation result into a pre-constructed vehicle engineering knowledge graph corresponding to the target vehicle so as to determine a safety optimization suggestion corresponding to the safety configuration.
9. 9. An electronic device comprising a processor, a memory and a bus, the memory storing machine-readable instructions executable by the processor, the processor and the memory in communication over the bus when the electronic device is in operation, the machine-readable instructions being executable by the processor to perform the steps of the accident data-based vehicle safety assessment method according to any one of claims 1 to 6.
10. 10. A computer-readable storage medium, characterized in that the computer-readable storage medium has stored thereon a computer program which, when executed by a processor, performs the steps of the accident data-based vehicle safety assessment method according to any one of claims 1 to 6.

Description

Vehicle safety assessment method and device based on accident data Technical Field The application relates to the technical field of vehicle safety evaluation, in particular to a vehicle safety evaluation method and device based on accident data. Background With the rapid development of the automotive industry, vehicle safety performance has become a central concern in the development process. Traditional safety development mainly relies on laboratory standard tests, such as standardized crash tests like new vehicle evaluation regulations. The tests evaluate the safety configurations of a body stabilizing system, an automatic emergency braking system and the like of a vehicle by simulating collision working conditions such as front collision, side collision and the like in a fixed scene. In addition, limited field testing is often performed to supplement laboratory testing. However, traffic accident scenarios in the real world have a high degree of complexity and unpredictability, involving varying road conditions, weather factors, vehicle interaction patterns, and driver behavior differences. Laboratory standard tests, while capable of covering part of typical operating conditions, do not adequately simulate high risk scenarios such as multi-car chain collisions in extreme weather, accidents in complex road topologies, or sudden interactions of traffic participants. Due to the lack of systematic collection and analysis of real accident data, traditional safety development methods have difficulty in dynamically identifying these potentially high risk conditions, resulting in a vehicle safety assessment that tends to lag behind changes in actual road risk. Disclosure of Invention Therefore, the application aims to provide a vehicle safety assessment method and device based on accident data, which actively identify potential high-risk scenes which cannot be covered by standard tests by deep mining of massive multidimensional accident data, and carry out simulation tests on the identified high-risk scenes in a vehicle digital model, so that the safety performance of the vehicle safety configuration under extreme working conditions can be assessed before a real accident occurs in a large scale, and the pertinence, the prospective and the actual coverage of the safety assessment are obviously improved. In a first aspect, an embodiment of the present application provides a vehicle security assessment method based on accident data, where the vehicle security assessment method includes: Acquiring a historical accident data set corresponding to a vehicle historical accident of a target vehicle, and carrying out data fusion on the historical accident data set to generate target accident data corresponding to the vehicle historical accident; performing cluster analysis on the plurality of target accident data to identify at least one risk accident scene; And simulating the risk accident scene by using a simulation model corresponding to the target vehicle aiming at each risk accident scene so as to evaluate the performance of the safety configuration of the target vehicle in the risk accident scene and determine the safety configuration evaluation result of the target vehicle. Further, after determining the safety configuration evaluation result of the target vehicle, the vehicle safety evaluation method further includes: And inputting the safety configuration evaluation result into a pre-constructed vehicle engineering knowledge graph corresponding to the target vehicle so as to determine a safety optimization suggestion corresponding to the safety configuration. Further, the historical accident data set comprises accident scene investigation text, accident scene images, vehicle operation time sequence data and accident occurrence weather data, and the data fusion is carried out on the historical accident data set to generate target accident data corresponding to the vehicle historical accident, which comprises the following steps: Carrying out named entity identification and relation extraction on the accident scene investigation text by using a large language model, and determining an accident semantic feature set corresponding to the vehicle history accident; Carrying out vehicle damage identification on the accident scene image by utilizing an image detection model, and determining a vehicle damage characteristic set corresponding to the vehicle history accident; And carrying out time window alignment on the vehicle operation time sequence data and the accident occurrence weather data, and correlating the time window alignment with the accident semantic feature set and the vehicle breakage feature set to obtain the target accident data. Further, the performing cluster analysis on the plurality of target accident data identifies at least one risk accident scenario, including: Converting each target accident data into corresponding high-dimensional feature vectors, and clustering and grouping each high-dimensional