Vehicle accident early warning method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention relates to the technical field of vehicles, and discloses a vehicle accident early warning method and device, electronic equipment and a storage medium, the method comprises the following steps: obtaining running state data in a vehicle based on a preset duration, the running state data being determined based on vehicle data obtained by at least one bus source in the vehicle; determining target time sequence characteristic data in the operation state data, wherein the target time sequence characteristic data indicates a change trend, a periodic rule and a sudden change condition of the operation state data in a time dimension and a dependency relationship among the data; and inputting the target time sequence characteristic data into a risk assessment model to obtain an assessment result of the accident risk of the vehicle after the preset duration. By applying the technical scheme of the invention, more accurate vehicle early warning can be realized, and the safety driving requirement of the vehicle can be better met.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of vehicles, and in particular, to a vehicle accident warning method, device, electronic device, and storage medium. Background Art

[0002] With the gradual improvement of vehicle automation, cars are transforming from simple means of transportation to intelligent mobile terminals. This transformation makes vehicle status monitoring and accident prediction crucial for ensuring driving safety and improving traffic efficiency. Against this background, vehicle status monitoring and accident prediction technologies have emerged, becoming a key research direction in the field of automotive safety, aiming to detect potential vehicle risks in advance and prevent problems before they occur.

[0003] In the existing related technologies, there are many limitations in vehicle accident warning technologies based on fixed thresholds and simple statistical analysis, which leads to inaccuracies in vehicle warning implementation and cannot meet the requirements of safe driving of vehicles. Summary of the Invention

[0004] In view of the above problems, the embodiments of the present invention provide a vehicle accident warning method, device, electronic device, and storage medium, which are used to solve the problems in the prior art that the vehicle warning implementation is inaccurate and cannot meet the requirements of safe driving of vehicles.

[0005] According to one aspect of the embodiments of the present invention, a vehicle accident warning method is provided, and the method includes:

[0006] Obtain the operation state data of the vehicle within a preset time period, where the operation state data is determined based on the vehicle data obtained from at least one bus source in the vehicle;

[0007] Determine the target time series feature data in the operation state data, where the target time series feature data indicates the change trend, periodic law, mutation situation, and dependency relationship between data in the time dimension of the operation state data;

[0008] Input the target time series feature data into a risk assessment model to obtain an assessment result of the accident risk of the vehicle after the preset time period. The risk assessment model is determined based on at least one time series feature data and the annotation result corresponding to the at least one time series feature data. For any annotation result, the annotation result indicates the accident risk situation of the vehicle corresponding to the corresponding time series feature data.

[0009] According to another aspect of the embodiments of the present invention, a vehicle accident warning device is provided, including:

[0010] An acquisition module, configured to acquire operation state data in a vehicle within a preset duration, where the operation state data is determined based on vehicle data acquired from at least one bus source in the vehicle;

[0011] A determination module, configured to determine target timing feature data in the operation state data, where the target timing feature data indicates a change trend, a periodic pattern, a mutation condition, and a dependency relationship between data in the time dimension of the operation state data;

[0012] A processing module, configured to input the target timing feature data into a risk assessment model to obtain an assessment result of the accident risk of the vehicle after the preset duration, where the risk assessment model is determined based on at least one timing feature data and a labeling result corresponding to the at least one timing feature data, and for any labeling result, the labeling result indicates the accident risk situation of the vehicle corresponding to the corresponding timing feature data.

[0013] According to another aspect of the embodiments of the present invention, there is provided an electronic device, including: a processor, a memory, a communication interface, and a communication bus, where the processor, the memory, and the communication interface complete mutual communication through the communication bus;

[0014] The memory is used to store at least one executable instruction, and the executable instruction causes the processor to execute the operations of the above method.

[0015] According to still another aspect of the embodiments of the present invention, there is provided a computer-readable storage medium, where at least one executable instruction is stored in the storage medium, and the executable instruction causes the vehicle accident warning device / electronic device to execute the operations of the above method.

[0016] According to still another aspect of the embodiments of the present invention, there is provided a computer program product, characterized in that it includes a computer program, and when the computer program is executed by a processor, it causes the vehicle accident warning device / electronic device to execute the operations of the above method.

[0017] In an embodiment of the present invention, by obtaining the operation state data of a vehicle within a preset time period, the operation state data is determined based on vehicle data obtained from at least one bus source in the vehicle; determining target timing feature data in the operation state data, the target timing feature data indicating the change trend, periodic law, mutation situation, and dependency relationship between data of the operation state data in the time dimension; inputting the target timing feature data into a risk assessment model to obtain an assessment result of the accident risk of the vehicle after the preset time period, the risk assessment model being determined based on at least one timing feature data and at least one annotation result corresponding to the at least one timing feature data, for any annotation result, the annotation result indicating the accident risk situation of the vehicle corresponding to the corresponding timing feature data. In this technical solution, first, obtaining the operation state data of the vehicle within a preset time period from at least one bus source of the vehicle can comprehensively cover various key information of the vehicle operation; then, accurately determining the target timing feature data therein, these data deeply analyze the change trend, periodic law, mutation situation, and dependency relationship between data of the operation state data in the time dimension, and can depict a unique portrait of the vehicle operation, making the early warning result more suitable for the current vehicle itself; subsequently, inputting these key target timing feature data into a risk assessment model trained with a large number of timing feature data and corresponding annotation results, since the model is trained based on rich and diverse historical data, it can accurately identify the association between various features and accident risks, and thus can transcend the limitations of traditional methods and more sensitively and accurately assess the accident risk of the vehicle after the preset time period. Overall, it can greatly improve the timeliness and accuracy of accident risk assessment, provide a strong guarantee for the safe operation of the vehicle, and significantly reduce the possibility of traffic accidents.

[0018] The above description is only an overview of the technical solution of the embodiment of the present invention. In order to be able to understand the technical means of the embodiment of the present invention more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features, and advantages of the embodiment of the present invention more obvious and understandable, the following specifically describes the embodiments of the present invention. Brief Description of the Drawings

[0019] The drawings are only used to illustrate the embodiments and are not considered to be a limitation of the present invention. And throughout the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0020] Figure 1 Shows a schematic diagram of a vehicle accident early warning system provided by the present invention;

[0021] Figure 2 Shows a flowchart of the first embodiment of the vehicle accident early warning method provided by the present invention;

[0022] Figure 3Shows the flowchart of the second embodiment of the vehicle accident warning method provided by the present invention;

[0023] Figure 4 Schematic diagram of the vehicle dynamics state transition matrix provided by the embodiments of the present application;

[0024] Figure 5 Schematic diagram of the FFT transform analysis of the steering wheel oscillation frequency provided by the embodiments of the present application;

[0025] Figure 6 Shows the flowchart of the third embodiment of the vehicle accident warning method provided by the present invention;

[0026] Figure 7 Shows the structural schematic diagram of the embodiment of the vehicle accident warning device provided by the present invention;

[0027] Figure 8 Shows the structural schematic diagram of the embodiment of the electronic device provided by the present invention. Detailed implementation manners

[0028] The exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein.

[0029] With the gradual improvement of the vehicle automation level, the automobile is transforming from a simple means of transportation to an intelligent mobile terminal. However, this transformation makes vehicle status monitoring and accident prediction crucial for ensuring driving safety and improving traffic efficiency.

[0030] That is, on the one hand, the driver's expectation for the safety of the vehicle during driving is continuously increasing; on the other hand, the traffic management department also urgently needs to reduce the accident rate through advanced technical means. Against this background, vehicle status monitoring and accident prediction technologies have emerged and become a key research direction in the field of automotive safety, aiming to detect potential vehicle risks in advance and prevent problems before they occur.

[0031] In the current related technologies, vehicle accident warning mainly relies on the following technical paths:

[0032] 1. The method based on fixed thresholds, that is, a series of safety thresholds of vehicle operation parameters are set in advance, such as vehicle speed, tire pressure, engine temperature, etc. When the actual operation parameters of the vehicle exceed these preset fixed thresholds, the system triggers a warning;

[0033] 2. The simple statistical analysis method is to collect the operation data of the vehicle for a period of time and use basic statistical means, such as calculating the mean, variance, etc., to determine whether the vehicle operation state is within the normal range.

[0034] In the above related methods, the following technical problems exist:

[0035] 1), For the method based on a fixed threshold, its threshold setting often fails to adapt to complex and ever-changing actual driving scenarios. Different vehicles have different optimal operating parameter ranges under different working conditions, and it is difficult for a fixed threshold to take them all into account;

[0036] 2), For simple statistical analysis methods, they only rely on the basic statistical features of historical data and lack the exploration of the deep features and dynamic change laws of the data. When a vehicle suddenly has a new fault mode, since it is not within the scope of historical statistics, this method is difficult to detect and give an early warning in advance, resulting in a great reduction in the timeliness and accuracy of the early warning and being unable to effectively meet the increasingly complex vehicle safety guarantee requirements;

[0037] 3), The data source is single. The data sources of the above methods only rely on the data of the Controller Area Network (CAN) bus, resulting in limited accident prediction accuracy;

[0038] 4), The above methods usually adopt single-thread sequential processing, resulting in too high a computational load in a high-frequency CAN log environment and being unable to meet the millisecond-level real-time prediction requirements;

[0039] 5), CAN log data is usually stored and transmitted in plain text, which is easily tampered with by malicious attackers, affecting the reliability of the prediction system.

[0040] Based on the above existing technical problems, the technical concept of the present invention is as follows: Considering that the method based on a fixed threshold cannot take into account the differences in the optimal operating parameters of different vehicles under various working conditions, and the simple statistical analysis method is difficult to cope with new fault modes due to the lack of exploration of the deep features and dynamic change laws of the data, and both of them have problems such as a single data source (only relying on CAN bus data), inefficient processing methods (single-thread sequential processing cannot meet real-time requirements), and poor data security (CAN log data is easily tampered with). If it is possible to obtain the operating state data within a preset time period from at least one bus source of the vehicle to broaden the data source and avoid the limitations of a single data source; further determine the target time-series feature data in the operating state data to capture in detail various key characteristics of the data in the time dimension, which can deeply explore the data connotation, overcome the deficiencies of simple statistical analysis, and can describe the unique operating conditions of the current vehicle; finally, a risk assessment model that can accurately predict future time can be trained using multiple time-series feature data and the historical data of corresponding accidents, and the target time-series feature data is input into this risk assessment model, so as to obtain an accurate assessment of the vehicle accident risk, thereby solving the above technical problems.

[0041] The following is for the above technical conceptFigure 1 shows a schematic diagram of the vehicle accident warning system provided by the present invention. As Figure 1 shown, the schematic diagram includes: a data acquisition layer, an edge computing layer, a cloud prediction layer, and a security layer.

[0042] For the data acquisition layer, it can be arranged on the vehicle side to collect the operating state data obtained from at least one bus source. The at least one data source can be: Global Navigation Satellite System (GNSS), Inertial Measurement Unit (IMU), CAN bus, Flexible Ray (FlexRay), Ethernet.

[0043] For the edge computing layer, real-time preprocessing of the operating state data is performed, using LZ4 compression and feature extraction, such as steering angle, braking pressure, etc.

[0044] For the cloud prediction layer, it can be arranged on the electronic device side. The steering angle, braking pressure, etc. are sequentially input into a Temporal Convolutional Network (TCN), a Long-Short Term Memory (LSTM)-Gated Recurrent Unit (GRU) model to obtain an evaluation result, and the model parameters can also be dynamically updated based on whether an accident has occurred uploaded by the following vehicle.

[0045] For the security layer, SM4 encryption + blockchain evidence storage is used to protect data integrity, and DoS attack detection based on SVM (Support Vector Machine) is used to improve the anti-attack ability of the entire system.

[0046] Next, the technical solution of the present invention will be described in detail through specific embodiments. The execution entity of the present invention is an electronic device, which can specifically be a server corresponding to providing cloud services, etc., or a processing device in a vehicle, etc.

[0047] It should be noted that the following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments.

[0048] Figure 2 shows a flowchart of the first embodiment of the vehicle accident warning method provided by the present invention, and this method is executed by an electronic device. As Figure 2 shown, this method includes the following steps:

[0049] Step 21: Obtain the operation status data of the vehicle within a preset duration;

[0050] Among them, the operation status data is determined based on the vehicle data obtained from at least one bus source in the vehicle;

[0051] In this step, in order to be able to accurately determine the risk of an accident occurring to the current vehicle in the future duration in real time, the operation status data of the vehicle can be collected in real time. In order to be able to perform correlation analysis over time, the operation status data within a preset duration is used.

[0052] In a possible implementation, the preset duration can be a period of time starting from the real-time time, such as 3s, 1min, etc.

[0053] Optionally, for the operation status data of the vehicle, it can be determined based on multiple bus sources. The multiple bus sources can be the above-mentioned GNSS, IMU, CAN bus, FlexRay, Ethernet, etc.

[0054] Correspondingly, GNSS / IMU is used to collect data such as the position information, yaw angular velocity, and acceleration of the vehicle; FlexRay / Ethernet is used to collect driving data related to autonomous driving or / and advanced driver assistance in the vehicle; the CAN bus is used to collect the dynamic data of the vehicle (such as vehicle speed, acceleration, steering angle, brake pressure, etc.).

[0055] Optionally, the operation status data includes at least one of the following: steering wheel angle, brake pressure, speed data.

[0056] Correspondingly, for the obtained dynamic data, the position information of the vehicle, yaw angular velocity, acceleration, and driving data, the vehicle end performs mutual verification and analysis processing to obtain key signals such as steering wheel angle, brake pressure, and speed data.

[0057] In a possible implementation, 1), for the speed data: in the dynamic data, such as the power transmission and component motion state data when the vehicle is driving, combined with the accurate position information of the vehicle, the speed can be calculated through the position change rate. Its yaw angular velocity and acceleration data can assist in judging the dynamic characteristics of the vehicle during steering and acceleration / deceleration. At the same time, the driving data contains driver operation information. During verification and analysis, the speed calculated value of GNSS / IMU is compared with the speed data in FlexRay / Ethernet. When the deviation is too large, it is corrected according to information such as the transmission ratio in the CAN bus dynamic data.

[0058] 2), for the steering wheel angle: The steering wheel angle command operated by the driver in FlexRay / Ethernet can be compared with the actual angle feedback by the steering system in the CAN bus dynamics data, and comprehensive calibration can be performed with reference to the vehicle yaw angular velocity.

[0059] 3), for the brake pressure: The brake pressure in the CAN bus dynamics can be combined with the driver's brake operation signal in FlexRay / Ethernet, and verified and calibrated according to the vehicle acceleration change, so as to obtain more accurate steering wheel angle, brake pressure, and speed data.

[0060] Furthermore, since the above data is within a preset time period, the determined operating state data is sequential.

[0061] Optionally, a possible implementation of step 21 can be: obtaining the operating state data within a preset time period after preprocessing through the Message Queuing Telemetry Transport (MQTT) method;

[0062] wherein, the preprocessing includes at least one of the following: symmetric encryption, lossless compression, and blockchain evidence storage.

[0063] In this implementation, in order to enable the electronic device to obtain the operating state data in the vehicle more accurately, with high real-time performance, and subsequent verifiability, the operating state data can be preprocessed at the vehicle end.

[0064] For example, for symmetric encryption, the SM4 algorithm (or AES-256 algorithm) is used to divide the vehicle operating state data into fixed-length data blocks according to a specific key, and then through a series of complex encryption operations, these data blocks are converted into ciphertext form. In this way, even if the data is stolen during transmission or storage, an attacker without the correct key cannot obtain the original operating state data, thus ensuring the security and confidentiality of the data.

[0065] For lossless compression, the LZ4 algorithm can be used for compression. That is, when compressing the operating state data, the LZ4 algorithm scans the operating state data, replaces consecutive identical byte sequences with a short marker that points to the position where the sequence first appears in the data, and records the length of the sequence. In this way, without losing any original data information, the storage space occupied by the data can be reduced, and the storage and transmission efficiency of the data can be improved, making the operating state data more efficient and convenient during storage and transmission.

[0066] Regarding blockchain evidence storage, blockchain evidence storage can be a way to reliably store and prove the operation status data using blockchain technology. That is, by packaging the vehicle operation status data into blocks according to certain rules and chronological order and linking them to the previous blocks to form a chain structure, each block contains the hash value of the data, which makes it difficult to tamper with the data once it is recorded on the blockchain because any modification to the data will cause a change in the hash value and thus can be easily identified by other nodes. At the same time, the distributed nature of the blockchain ensures that the data is stored on multiple nodes, avoiding the risk of single-point failure and data loss, realizing the secure and trustworthy evidence storage of the operation status data, and providing a reliable basis for subsequent auditing, traceability, and data verification.

[0067] Furthermore, on the side of the electronic device, a method using the support vector machine (SVM) algorithm for DoS attack detection based on SVM can be applied to identify and prevent denial-of-service (DoS) attacks to ensure the anti-interference of the operation status data transmission.

[0068] Among them, the SVM constructs a classification model by analyzing a large number of known normal and abnormal data samples. In the scenario of operation status data transmission, the normal data transmission patterns and features are used as positive samples, and the data transmission patterns and features under DoS attacks are used as negative samples. The SVM algorithm will find an optimal hyperplane to separate these two types of samples as accurately as possible;

[0069] That is, when new data (such as operation status data) is transmitted, the SVM model will judge whether it belongs to a DoS attack behavior according to the characteristics of the data. For example, if it is found that a certain data transmission behavior presents abnormal traffic patterns, such as suddenly appearing a large number of repeated requests, abnormal data transmission rates, etc., which match the DoS attack characteristics in the trained SVM model, it will be identified as a potential DoS attack. In this way, DoS attacks can be detected in a timely manner, and corresponding measures can be taken, such as blocking abnormal traffic, notifying the administrator, etc., to improve the anti-attack ability of the system during the operation status data transmission process and ensure the stability and reliability of data transmission.

[0070] Step 22: Determine the target time-series feature data in the operation status data;

[0071] Among them, the target time-series feature data indicates the change trend, periodic law, mutation situation, and the dependence relationship between data in the time dimension of the operation status data;

[0072] In this step, in order to improve the accuracy of evaluating the risk of the operation status data sent by the vehicle, the operation status data can be processed by extracting time-series features.

[0073] That is, from a large amount of time-series data generated by the vehicle operating state, characteristic data that can accurately reflect the key aspects of the vehicle operating state, is representative and regular is selected and denoted as target time-series characteristic data.

[0074] Step 23: Input the target time-series characteristic data into the risk assessment model to obtain an assessment result of the accident risk of the vehicle after a preset time period;

[0075] Among them, the risk assessment model is determined based on at least one time-series characteristic data and the annotation result corresponding to at least one time-series characteristic data. For any annotation result, the annotation result indicates the accident risk situation of the vehicle corresponding to the corresponding time-series characteristic data.

[0076] In this step, the target time-series characteristic data determined from the vehicle operating state data is input into the pre-established risk assessment model. The risk assessment model will analyze and calculate based on the vehicle operating condition information contained in these target time-series characteristic data using specific algorithms and rules to output an assessment result of the accident risk of the vehicle after a preset time period. The assessment result can be the first probability of the accident risk and the second probability of no accident risk.

[0077] Optionally, before step 23, it can also be executed: Train the long short-term memory network-gated recurrent unit according to at least one time-series characteristic data and the annotation result corresponding to at least one time-series characteristic data to obtain the risk assessment model.

[0078] In this implementation, in order to construct a risk assessment model for vehicle accidents, at least one time-series characteristic data arranged in chronological order that can reflect the vehicle operating state can be collected. At the same time, for each group of time-series characteristic data, an annotation result indicating the accident risk situation of the vehicle is given. Then, using the deep learning architecture of the long short-term memory network-gated recurrent unit, these time-series characteristic data are used as inputs, and the corresponding annotation results are used as expected outputs, allowing the model to continuously learn the association pattern between the two. During the training process, the model will continuously adjust its own parameters to predict the vehicle accident risk as accurately as possible based on the time-series characteristic data to obtain a model that can be used to assess the vehicle accident risk.

[0079] Among them, the annotation result can be obtained by a technician based on the actual situation of whether an accident occurs after the time-series characteristic data and the corresponding probability of the accident occurring.

[0080] Optionally, after step 23, it can also be executed: If the first probability indicating that the vehicle has an accident risk in the assessment result is greater than the preset probability, send the assessment result to the vehicle so that the vehicle generates a warning message based on the assessment result and / or so that the vehicle controls the vehicle operation based on the assessment result;

[0081] Among them, the warning information includes at least one of the following: prompting that the current driving risk of the vehicle is relatively high and determining a driving strategy to avoid accident risks.

[0082] In this implementation, an evaluation result of the accident risk of the vehicle is obtained through a risk assessment model. When the first probability indicating the accident risk of the vehicle is greater than a preset probability, it means that the electronic device determines that the vehicle is in a high accident risk state; among them, the preset probability is a threshold set in advance and is used to define whether the risk reaches the level that requires intervention.

[0083] Furthermore, once the risk exceeds the threshold (that is, the first probability is greater than the preset probability), the evaluation result will be sent to the vehicle, and the vehicle will take corresponding measures based on this evaluation result. That is, the following implementations are possible:

[0084] 1), the warning information includes prompting that the current driving risk of the vehicle is relatively high, which can enable the driver to intuitively understand the dangerous situation of the vehicle and arouse their alertness;

[0085] 2), the warning information includes determining a driving strategy to avoid accident risks, that is, providing specific coping methods for the driver, which has strong guiding significance and helps the driver adjust their driving behavior in a timely manner and reduce the possibility of accidents.

[0086] 3), the vehicle can also directly control the vehicle operation based on the evaluation result. For example, automatically reducing the vehicle speed, adjusting the cruise control parameters, or preparing the safety assistance system in advance, etc. This further enhances the safety of the vehicle and provides additional protection for the driver through technical means.

[0087] Optionally, the above preprocessing can also be applied to the symmetric encryption, lossless compression, blockchain evidence storage, etc. of vehicle data, operating state data, target timing feature data, evaluation results, etc. corresponding to at least one bus source on the electronic device side.

[0088] The vehicle accident warning method provided by the embodiments of the present invention obtains the operation state data of the vehicle within a preset time period; determines the target time series feature data in the operation state data; and inputs the target time series feature data into a risk assessment model to obtain an assessment result of the accident risk of the vehicle after the preset time period. In this technical solution, first, the operation state data within a preset time period is obtained from at least one bus source of the vehicle, which can comprehensively cover various key information of the vehicle operation; then, the target time series feature data therein is accurately determined. These data deeply analyze the change trend, periodic law, mutation situation, and the dependence relationship between data in the time dimension of the operation state data, and can depict a unique portrait of the vehicle operation, making the warning result more suitable for the current vehicle itself; subsequently, these key target time series feature data are input into a risk assessment model trained with a large amount of time series feature data and corresponding annotation results. Since the model is trained based on rich and diverse historical data, it can accurately identify the association between various features and accident risks, and thus can transcend the limitations of traditional methods and more sensitively and accurately assess the accident risk of the vehicle after the preset time period. Overall, it can greatly improve the timeliness and accuracy of accident risk assessment, provide a strong guarantee for the safe operation of the vehicle, and significantly reduce the possibility of traffic accidents.

[0089] Based on the above embodiments, Figure 3 The flowchart of the second embodiment of the vehicle accident warning method provided by the present invention is shown, and this method is executed by the vehicle. As Figure 3 shown, step 22 may have the following possible implementations:

[0090] Step 31: Determine the target risk assessment key data corresponding to the operation state data;

[0091] In this step, by analyzing the operation state data (such as, steering wheel angle, brake pressure, speed data), the target risk assessment key data that affects the safe operation of the vehicle can be calculated.

[0092] Optionally, the risk assessment key data includes at least one of the following: acceleration data, jerk data, steering wheel oscillation frequency. The risk assessment key data includes the above items, and the following can be used for argumentation and analysis:

[0093] In a possible implementation, Figure 4 is a schematic diagram of the vehicle dynamics state transition matrix provided by the embodiments of the present application. As Figure 4 shown, the acceleration data and jerk data determined based on the speed data are directly given here. For example, some thresholds can be set first: the acceleration is 0.8g, and the change rate of the acceleration with time is 15m / s 3 , and the jerk is 30m / s 4 .

[0094] That is, when the vehicle is driving normally, if the acceleration is greater than 0.8g, it is determined as sudden acceleration / sudden braking; when the rate of change of acceleration with respect to time is greater than A (15 m / s 3 ), and / or the jerk is greater than B (30 m / s 4 ), it is considered that there is a collision risk.

[0095] In a possible implementation, Figure 5 is a schematic diagram of the fast Fourier transform (FFT) analysis of the steering wheel oscillation frequency provided by the embodiment of the present application. As Figure 5 shown:

[0096] FFT input (steering wheel signal, such as steering wheel angle); perform FFT transformation; analyze the frequency distribution in the range of 0 - 10 Hz; determine whether there is an abnormality in the frequency range of 4 - 6 Hz. If there is an abnormality in this frequency band, it indicates that there may be abnormal oscillation of the steering wheel, and then enter the "detected abnormal oscillation" branch. If there is no abnormality, enter the "normal state" branch.

[0097] That is, the steering wheel oscillation frequency can adopt the oscillation frequency - related data in the frequency range of 4 - 6 Hz after FFT transformation.

[0098] The above has explained the roles of acceleration data, jerk data, and steering wheel oscillation frequency in risk assessment. Then, proceed with the following.

[0099] Step 32: Input the target risk assessment key data into the temporal convolutional network to obtain the target temporal feature data.

[0100] Among them, the temporal convolutional network is determined based on at least one risk assessment key data and at least one temporal feature data corresponding to the at least one risk assessment key data.

[0101] In this step, the pre - trained temporal convolutional network is used to process the input target risk assessment key data. It can be processed through multiple TCN convolutional layers, attention mechanism, fully - connected layers, etc., and output the target temporal feature data indicating the change trend, periodic law, mutation situation, and the dependence relationship between data in the time dimension of the operation state data.

[0102] The vehicle accident warning method provided by the embodiment of the present invention determines the target risk assessment key data corresponding to the operating state data, inputs the target risk assessment key data into a temporal convolutional network to obtain target temporal feature data, and the temporal convolutional network is determined based on at least one risk assessment key data and the temporal feature data corresponding to at least one risk assessment key data. In this technical solution, by determining the target risk assessment key data corresponding to the operating state data, the core information during vehicle operation is accurately focused, avoiding interference from redundant data. Furthermore, the TCN can effectively capture the local and global dependencies of data in the time dimension, deeply mine the potential patterns and rules in the target risk assessment key data, and then output accurate target temporal feature data.

[0103] Based on the above embodiment, the vehicle type of the vehicle is the first type; correspondingly, Figure 6 The flowchart of the third embodiment of the vehicle accident warning method provided by the present invention is shown, as Figure 6 shown, step 23 may have the following possible implementations:

[0104] Step 61: Determine the risk assessment model corresponding to the first type in the risk assessment models corresponding to at least one vehicle type;

[0105] Among them, for the risk assessment model corresponding to each vehicle type, the risk assessment model corresponding to the vehicle type is determined based on at least one temporal feature data of the vehicle corresponding to the vehicle type and the annotation result corresponding to at least one temporal feature data of the vehicle corresponding to the vehicle type;

[0106] In this solution, in order to improve the accuracy of vehicle risk assessment, based on different vehicle types, the risk assessment model can be determined according to the training data corresponding to the respective vehicle types (at least one temporal feature data and the annotation result corresponding to at least one temporal feature data of the vehicle corresponding to the vehicle type), and the risk assessment model corresponding to each vehicle type is obtained.

[0107] When implementing this step, the corresponding risk assessment model can be determined based on the vehicle type being the first type.

[0108] Step 62: Input the target temporal feature data into the risk assessment model corresponding to the first type to obtain the assessment result of the accident risk of the vehicle after a preset duration.

[0109] The vehicle accident warning method provided by the embodiments of the present invention determines the risk assessment model corresponding to the first type in at least one risk assessment model corresponding to a vehicle type; wherein, for each risk assessment model corresponding to a vehicle type, the risk assessment model corresponding to the vehicle type is determined based on at least one time-series feature data of the vehicle corresponding to the vehicle type and the annotation result corresponding to at least one time-series feature data of the vehicle corresponding to the vehicle type; the target time-series feature data is input into the risk assessment model corresponding to the first type to obtain an assessment result of the accident risk of the vehicle after a preset duration. In this technical solution, risk assessment models are constructed separately for different vehicle types, and each model is trained based on a large amount of time-series feature data of the corresponding vehicle type and its annotation results, which enables each model to deeply fit the operating characteristics of a specific vehicle type and accurately capture its unique risk patterns, that is, to obtain a more accurate assessment result of the current vehicle's accident risk in this embodiment.

[0110] Figure 7 The structural schematic diagram of the embodiment of the vehicle accident warning device provided by the present invention is shown. As Figure 7 shown, the device includes:

[0111] An acquisition module 71, configured to acquire the operation state data of the vehicle within a preset duration, and the operation state data is determined based on the vehicle data acquired from at least one bus source in the vehicle;

[0112] A determination module 72, configured to determine the target time-series feature data in the operation state data, and the target time-series feature data indicates the change trend, periodic law, mutation situation, and dependency relationship between data in the time dimension of the operation state data;

[0113] A processing module 73, configured to input the target time-series feature data into a risk assessment model to obtain an assessment result of the accident risk of the vehicle after a preset duration, and the risk assessment model is determined based on at least one time-series feature data and the annotation result corresponding to at least one time-series feature data, and for any annotation result, the annotation result indicates the accident risk situation of the vehicle corresponding to the corresponding time-series feature data.

[0114] In one or more embodiments, the determination module 72 is specifically configured to:

[0115] Determine the target risk assessment key data corresponding to the operation state data;

[0116] Input the target risk assessment key data into a temporal convolutional network to obtain the target time-series feature data, and the temporal convolutional network is determined based on at least one risk assessment key data and the time-series feature data corresponding to at least one risk assessment key data.

[0117] In one or more embodiments, before inputting the target time-series feature data into the risk assessment model to obtain the assessment result of the accident risk of the vehicle after a preset duration, the determination module 72 is further configured to:

[0118] Train the long short-term memory network-gated recurrent unit according to at least one time-series feature data and the annotation result corresponding to at least one time-series feature data to obtain a risk assessment model.

[0119] In one or more embodiments, the sending module is configured to:

[0120] If the first probability that the assessment result indicates that the vehicle has an accident risk is greater than the preset probability, send the assessment result to the vehicle, so that the vehicle generates a warning message based on the assessment result and / or so that the vehicle controls the vehicle operation based on the assessment result;

[0121] Wherein, the warning message includes at least one of the following: prompting that the current driving risk of the vehicle is relatively high, determining a driving strategy to avoid accident risk.

[0122] In one or more embodiments, the acquisition module 71 is specifically configured to:

[0123] Obtain the operation state data within a preset duration after preprocessing through the message queue telemetry transmission method;

[0124] Wherein, the preprocessing includes at least one of the following: symmetric encryption, lossless compression, blockchain evidence storage.

[0125] In one or more embodiments, the vehicle type of the vehicle is the first type;

[0126] Correspondingly, the processing module 73 is specifically configured to:

[0127] Determine the risk assessment model corresponding to the first type in the risk assessment models corresponding to at least one vehicle type. For the risk assessment model corresponding to each vehicle type, the risk assessment model corresponding to the vehicle type is determined based on at least one time-series feature data of the vehicle corresponding to the vehicle type and the annotation result corresponding to at least one time-series feature data of the vehicle corresponding to the vehicle type;

[0128] Input the target time-series feature data into the risk assessment model corresponding to the first type to obtain the assessment result of the accident risk of the vehicle after a preset duration.

[0129] In one or more embodiments, the operation state data includes at least one of the following: steering wheel angle, brake pressure, speed data;

[0130] The risk assessment key data includes at least one of the following: acceleration data, jerk data, steering wheel oscillation frequency.

[0131] It should be noted that it should be understood that the division of each module of the above device is only a division of logical functions. In actual implementation, it can be fully or partially integrated into a physical entity, or physically separated. And these modules can all be implemented in the form of software called by processing elements. They can also all be implemented in the form of hardware. It is also possible that some modules are implemented in the form of software called by processing elements, and some modules are implemented in the form of hardware. In addition, all or part of these modules can be integrated together or independently implemented. Here, the processing element can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each of the above modules can be completed by the integrated logic circuit in the processor element or the instructions in the form of software.

[0132] As can be seen from the above, the vehicle accident warning device provided by the embodiment of the present invention can comprehensively cover various key information of vehicle operation, and based on this information, it can depict a unique portrait of vehicle operation, making the warning result more in line with the current vehicle itself. Then, based on the target time series feature data for risk assessment, it can greatly improve the timeliness and accuracy of accident risk assessment, provide a strong guarantee for the safe operation of the vehicle, and significantly reduce the possibility of traffic accidents.

[0133] Figure 8 The structural schematic diagram of the embodiment of the electronic device provided by the present invention is shown, as Figure 8 shown, the electronic device may include: a processor 82, a communication interface 84, a memory 86, and a communication bus 88.

[0134] Among them: the processor 82, the communication interface 84, and the memory 86 communicate with each other through the communication bus 88. The communication interface 84 is used to communicate with network elements of other devices such as clients or other servers. The processor 82 is used to execute the program 80, and specifically can execute the relevant steps in the above-mentioned embodiment of the anti-collision method for the electronic device.

[0135] Specifically, the program 80 may include program code, and the program code includes computer-executable instructions.

[0136] The processor 82 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention. One or more processors included in the electronic device may be of the same type of processor, such as one or more CPUs. They may also be of different types of processors, such as one or more CPUs and one or more ASICs.

[0137] A memory 86 for storing a program 80. The memory 86 may include a high-speed RAM memory and may also include a non-volatile memory, such as at least one disk memory.

[0138] Specifically, the program 80 can be called by the processor 82 to cause the electronic device to perform the following operations:

[0139] Obtain the operation state data of the vehicle within a preset time period, where the operation state data is determined based on the vehicle data obtained from at least one bus source in the vehicle;

[0140] Determine the target timing feature data in the operation state data, where the target timing feature data indicates the change trend, periodic law, mutation situation, and dependency relationship between data of the operation state data in the time dimension;

[0141] Input the target timing feature data into a risk assessment model to obtain an assessment result of the accident risk of the vehicle after the preset time period. The risk assessment model is determined based on at least one timing feature data and the annotation result corresponding to the at least one timing feature data. For any annotation result, the annotation result indicates the accident risk situation of the vehicle corresponding to the corresponding timing feature data.

[0142] In one or more embodiments, determining the target timing feature data in the operation state data includes:

[0143] Determine the target risk assessment key data corresponding to the operation state data;

[0144] Input the target risk assessment key data into a temporal convolutional network to obtain the target timing feature data. The temporal convolutional network is determined based on at least one risk assessment key data and the timing feature data corresponding to the at least one risk assessment key data.

[0145] In one or more embodiments, before inputting the target timing feature data into the risk assessment model to obtain the assessment result of the accident risk of the vehicle after the preset time period, the following is also performed:

[0146] Train the long short-term memory network-gated recurrent unit based on the at least one timing feature data and the annotation result corresponding to the at least one timing feature data to obtain the risk assessment model.

[0147] In one or more embodiments, the following is also performed:

[0148] If the first probability that the evaluation result indicates that the vehicle has an accident risk is greater than a preset probability, send the evaluation result to the vehicle, so that the vehicle generates a warning message based on the evaluation result and / or so that the vehicle controls the operation of the vehicle based on the evaluation result;

[0149] Wherein, the warning message includes at least one of the following: indicating that the current driving risk of the vehicle is relatively high, determining a driving strategy to avoid accident risk.

[0150] In one or more embodiments, the obtaining of the operation state data within a preset duration in the vehicle includes:

[0151] Obtain the operation state data within the preset duration after preprocessing through the message queue telemetry transmission method;

[0152] Wherein, the preprocessing includes at least one of the following: symmetric encryption, lossless compression, blockchain evidence deposit.

[0153] In one or more embodiments, the vehicle type of the vehicle is the first type;

[0154] Correspondingly, the inputting of the target timing feature data into the risk assessment model to obtain the evaluation result of the accident risk of the vehicle after the preset duration includes:

[0155] Determine the risk assessment model corresponding to the first type in the risk assessment models corresponding to at least one vehicle type. For the risk assessment model corresponding to each vehicle type, the risk assessment model corresponding to the vehicle type is determined based on at least one timing feature data of the vehicle corresponding to the vehicle type and the annotation result corresponding to the at least one timing feature data of the vehicle corresponding to the vehicle type;

[0156] Input the target timing feature data into the risk assessment model corresponding to the first type to obtain the evaluation result of the accident risk of the vehicle after the preset duration.

[0157] In one or more embodiments, the operation state data includes at least one of the following: steering wheel angle, brake pressure, speed data;

[0158] The key data for risk assessment includes at least one of the following: acceleration data, jerk data, and steering wheel oscillation frequency.

[0159] As can be seen from the above, the electronic device provided by the embodiment of the present invention can comprehensively cover various key information of vehicle operation, and based on this information, can depict a unique portrait of vehicle operation, making the warning result more suitable for the current vehicle itself. Then, based on the target time-series feature data for risk assessment, it can greatly improve the timeliness and accuracy of accident risk assessment, provide a strong guarantee for the safe operation of the vehicle, and significantly reduce the possibility of traffic accidents.

[0160] The embodiment of the present invention provides a computer-readable storage medium, and the storage medium stores at least one executable instruction. When the executable instruction runs on a vehicle accident warning device / electronic device, the vehicle accident warning device / electronic device is caused to execute the vehicle accident warning method in any of the above method embodiments.

[0161] The executable instruction can specifically be used to cause the vehicle accident warning device / electronic device to perform the following operations:

[0162] Obtain the operation state data of the vehicle within a preset time period, where the operation state data is determined based on the vehicle data obtained from at least one bus source in the vehicle;

[0163] Determine the target time-series feature data in the operation state data, where the target time-series feature data indicates the change trend, periodic law, mutation situation, and dependence relationship between data in the time dimension of the operation state data;

[0164] Input the target time-series feature data into a risk assessment model to obtain an assessment result of the accident risk of the vehicle after the preset time period. The risk assessment model is determined based on at least one time-series feature data and the annotation result corresponding to the at least one time-series feature data. For any annotation result, the annotation result indicates the accident risk situation of the vehicle corresponding to the corresponding time-series feature data.

[0165] In one or more embodiments, the determining the target time-series feature data in the operation state data includes:

[0166] Determine the target key data for risk assessment corresponding to the operation state data;

[0167] Input the target key data for risk assessment into a temporal convolutional network to obtain the target time-series feature data. The temporal convolutional network is determined based on at least one key data for risk assessment and the time-series feature data corresponding to the at least one key data for risk assessment.

[0168] In one or more embodiments, before inputting the target time series feature data into the risk assessment model to obtain the assessment result of the accident risk of the vehicle after the preset duration, the following is also performed:

[0169] Train the long short-term memory network-gated recurrent unit according to the at least one time series feature data and the annotation result corresponding to the at least one time series feature data to obtain the risk assessment model.

[0170] In one or more embodiments, the following is also performed:

[0171] If the first probability that the assessment result indicates that the vehicle has an accident risk is greater than the preset probability, send the assessment result to the vehicle, so that the vehicle generates a warning message based on the assessment result and / or so that the vehicle controls the operation of the vehicle based on the assessment result;

[0172] Wherein, the warning message includes at least one of the following: indicating that the current driving risk of the vehicle is relatively high, determining a driving strategy to avoid accident risk.

[0173] In one or more embodiments, the obtaining of the operation state data of the vehicle within a preset duration includes:

[0174] Obtain the operation state data within the preset duration after preprocessing through the message queue telemetry transmission method;

[0175] Wherein, the preprocessing includes at least one of the following: symmetric encryption, lossless compression, blockchain evidence storage.

[0176] In one or more embodiments, the vehicle type of the vehicle is the first type;

[0177] Correspondingly, the inputting of the target time series feature data into the risk assessment model to obtain the assessment result of the accident risk of the vehicle after the preset duration includes:

[0178] Determine the risk assessment model corresponding to the first type in the risk assessment models corresponding to at least one vehicle type. For each risk assessment model corresponding to a vehicle type, the risk assessment model corresponding to the vehicle type is determined based on at least one time series feature data of the vehicle corresponding to the vehicle type and the annotation result corresponding to the at least one time series feature data of the vehicle corresponding to the vehicle type;

[0179] Input the target time series feature data into the risk assessment model corresponding to the first type to obtain the assessment result of the accident risk of the vehicle after the preset duration.

[0180] In one or more embodiments, the operating state data includes at least one of the following: steering wheel angle, brake pressure, speed data;

[0181] The key risk assessment data includes at least one of the following: acceleration data, jerk data, steering wheel oscillation frequency.

[0182] As can be seen from the above, the vehicle accident warning device / electronic device provided by the embodiments of the present invention can comprehensively cover various key information of vehicle operation, and based on this information, a unique portrait of vehicle operation can be depicted, making the warning result more suitable for the current vehicle itself. Then, based on the target time series feature data for risk assessment, the timeliness and accuracy of accident risk assessment can be greatly improved, providing a strong guarantee for the safe operation of the vehicle and significantly reducing the possibility of traffic accidents.

[0183] The embodiments of the present invention provide a computer program product, including a computer program, which when executed by a processor implements the operations of the above-mentioned vehicle accident warning method.

[0184] Its implementation principle and technical effects are as disclosed above.

[0185] The descriptions of the above embodiments tend to emphasize the differences between the embodiments. The same or similar parts can be referred to each other. For the sake of brevity, they are not described in detail in this article.

[0186] The methods disclosed in the method embodiments provided by the present invention can be arbitrarily combined without conflict to obtain new method embodiments.

[0187] The features disclosed in the product embodiments provided by the present invention can be arbitrarily combined without conflict to obtain new product embodiments.

[0188] The features disclosed in the method or device embodiments provided by the present invention can be arbitrarily combined without conflict to obtain new method embodiments or device embodiments.

[0189] It should be noted that the above computer-readable storage medium may be a ROM, a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a ferromagnetic random access memory (FRAM), a flash memory, a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD-ROM), etc. It may also be various vehicles including one or any combination of the above memories.

[0190] It should be noted that in this document, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "including a..." does not exclude the existence of additional identical elements in the process, method, article or device including such element.

[0191] The serial numbers of the above embodiments of the present invention are only for description and do not represent the superiority or inferiority of the embodiments.

[0192] Through the description of the above embodiments, those skilled in the art can clearly understand that the above embodiment methods can be implemented by means of software plus necessary general hardware nodes. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation. Based on such an understanding, the technical solution of the present invention, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disc) and includes several instructions for causing a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, a vehicle-mounted terminal or a network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0193] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses, devices, and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows 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 the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing device produce a means for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0194] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including an instruction means that implements the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0195] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks. The algorithms or displays provided herein are not inherently related to any particular computer, virtual system, or other device. In addition, embodiments of the present invention are not directed to any particular programming language.

[0196] It should be noted that the above embodiments illustrate the present invention rather than limit the present invention, and those skilled in the art can design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention can be implemented by means of hardware including several different elements and by means of a suitably programmed computer. In a unit claim listing several devices, several of these devices can be embodied by the same item of hardware. The use of the words first, second, and third, etc. does not denote any order. These words can be interpreted as names. The steps in the above embodiments, unless otherwise specified, should not be construed as limiting the order of execution.

Claims

1. A vehicle accident early warning method, characterized in that: The method comprises: Acquiring running status data in a vehicle based on a preset time period, wherein the running status data is determined based on vehicle data acquired by at least one bus source in the vehicle; Determine target time series feature data in the running status data, where the target time series feature data indicates a change trend, periodicity, mutation, and dependency between data of the running status data in a time dimension; The target time series feature data is input into a risk assessment model to obtain an assessment result of the accident risk of the vehicle after the preset time period. The risk assessment model is determined based on at least one time series feature data and a labeling result corresponding to the at least one time series feature data. For any labeling result, the labeling result indicates the accident risk of the vehicle corresponding to the corresponding time series feature data.

2. The method according to claim 1, characterized in that The determining of the target time series characteristic data in the running status data comprises: Determine the target risk assessment key data corresponding to the operating status data; The target risk assessment key data is input into a temporal convolutional network to obtain the target temporal feature data, wherein the temporal convolutional network is determined based on at least one risk assessment key data and the temporal feature data corresponding to the at least one risk assessment key data.

3. The method according to claim 1, characterized in that Before inputting the target time series feature data into the risk assessment model to obtain the assessment result of the accident risk of the vehicle after the preset time period, the method further includes: According to the at least one time series feature data and the labeling result corresponding to the at least one time series feature data, the long short-term memory network-gated recurrent unit is trained to obtain the risk assessment model.

4. The method according to any one of claims 1 to 3, characterized in that: The method further comprises: If the evaluation result indicates that the first probability of the risk of an accident occurring to the vehicle is greater than a preset probability, the evaluation result is sent to the vehicle so that the vehicle generates a warning message based on the evaluation result, and / or the vehicle controls the operation of the vehicle based on the evaluation result; The warning information includes at least one of the following: prompting that the vehicle is currently at a high risk of driving and determining a driving strategy to avoid accident risks.

5. The method according to any one of claims 1 to 3, characterized in that: The obtaining of the running status data of the vehicle based on a preset time period includes: Obtaining the pre-processed running status data within the preset time period through a message queue telemetry transmission method; Among them, the preprocessing includes at least one of the following: symmetric encryption, lossless compression, and blockchain evidence storage.

6. The method according to any one of claims 1 to 3, characterized in that: The vehicle type of the vehicle is a first type; Accordingly, the target time series feature data is input into the risk assessment model to obtain the assessment result of the accident risk of the vehicle after the preset time period, including: Determining a risk assessment model corresponding to the first type from among the risk assessment models corresponding to at least one vehicle type, wherein for each risk assessment model corresponding to the vehicle type, the risk assessment model corresponding to the vehicle type is determined based on at least one time series feature data of a vehicle corresponding to the vehicle type and a labeling result corresponding to at least one time series feature data of the vehicle corresponding to the vehicle type; The target time series feature data is input into the risk assessment model corresponding to the first type to obtain an assessment result of the accident risk of the vehicle after the preset time period.

7. The method according to claim 2, characterized in that The operating status data includes at least one of the following: steering wheel angle, brake pressure, speed data; The key risk assessment data includes at least one of the following: acceleration data, jerk data, and steering wheel oscillation frequency.

8. A vehicle accident warning device, characterized in that: The device comprises: An acquisition module, configured to acquire running status data in a vehicle based on a preset time period, wherein the running status data is determined based on vehicle data acquired by at least one bus source in the vehicle; A determination module, used to determine target time series feature data in the running status data, wherein the target time series feature data indicates a change trend, periodicity, mutation, and dependency relationship between the running status data in a time dimension; A processing module is used to input the target time series feature data into a risk assessment model to obtain an assessment result of the accident risk of the vehicle after the preset time period. The risk assessment model is determined based on at least one time series feature data and a labeling result corresponding to the at least one time series feature data. For any labeling result, the labeling result indicates the accident risk of the vehicle corresponding to the corresponding time series feature data.

9. An electronic device, characterized in that: include: A processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus; The memory is used to store at least one executable instruction, and the executable instruction enables the processor to execute the operation of the vehicle accident warning method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: The storage medium stores at least one executable instruction, and when the executable instruction is executed on the vehicle accident warning device / electronic device, the electronic device / vehicle accident warning device executes the operation of the vehicle accident warning method as described in any one of claims 1-7.

Images