Vehicle accident handling method, device, equipment and product

Abstract

The invention provides a vehicle accident handling method, device, equipment and product, which are applied to an accident vehicle, the accident vehicle deploys a pre-trained motion prediction model and a pre-trained damage prediction model, and the vehicle accident handling method comprises the following steps: acquiring collision data of the accident vehicle and position information of passengers in the accident vehicle; according to the collision data, the motion condition of the body part of the passenger in the accident vehicle in the accident collision is predicted through a motion prediction model, and predicted motion time sequence characteristics of the body part are obtained; according to the predicted motion time sequence characteristics, the damage of the body part in the accident collision is predicted through a damage prediction model, and a predicted damage result of the body part is obtained; generating a passenger injury evaluation report according to the position information of the passenger and the injury prediction result; and sending a passenger injury assessment report to a rescue party. Therefore, the passenger injury assessment report of the accident vehicle is actively provided for the rescue party, and the efficiency of effectively rescuing the passengers of the accident vehicle is improved.

Description

Technical Field

[0001] This application relates to the field of vehicle safety, and in particular to a method, apparatus, equipment and product for handling vehicle accidents. Background Technology

[0002] Vehicle safety is one of the important indicators for measuring vehicle performance, and deploying vehicle safety systems on vehicles can effectively improve vehicle safety.

[0003] The on-board accident emergency call system (AECS) is a vehicle safety system designed to respond to traffic accidents. When a traffic accident occurs, the AECS provides accident information such as the accident location and airbag deployment status to a remote assistance center.

[0004] However, if occupants in a traffic accident lose consciousness or are unable to speak, it is difficult to confirm their physical condition through voice communication between the occupants and the AECS, resulting in the occupants not receiving timely and effective rescue. Summary of the Invention

[0005] Based on the above-mentioned technological status, this application provides a vehicle accident handling method, device, equipment, and product that can improve the efficiency of effective rescue of accident vehicles.

[0006] To achieve the above-mentioned technical objectives, this application proposes the following technical solution: According to a first aspect of this application, a vehicle accident handling method is provided, applied to an accident vehicle, wherein the accident vehicle is equipped with a pre-trained motion prediction model and a pre-trained damage prediction model. The vehicle accident handling method includes: acquiring collision data of the accident vehicle and the position information of the occupants in the accident vehicle; predicting the motion of the occupants' body parts in the accident collision using the motion prediction model based on the collision data, obtaining predicted motion temporal features of the body parts; predicting the damage suffered by the body parts in the accident collision using the damage prediction model based on the predicted motion temporal features, obtaining predicted damage results of the body parts; generating an occupant damage assessment report based on the occupant's position information and the predicted damage results of the body parts; and sending the occupant damage assessment report to the rescue party.

[0007] In some implementations, the step of predicting the motion of the occupant's body parts during a collision using the motion prediction model based on the collision data to obtain the predicted motion timing features of the body parts includes: extracting the occupant's safety constraint information and the vehicle's acceleration information from the collision data; the safety constraint information includes one or more of the following: seatbelt exit position, seatbelt force limiter level, seatbelt deployment time, or airbag deployment time; and predicting the motion of the body parts during a collision using the motion prediction model based on the safety constraint information and the vehicle acceleration information to obtain the predicted motion timing features.

[0008] In some implementations, the vehicle body acceleration information includes a vehicle body acceleration waveform, and the predicted motion timing features include a predicted acceleration curve. The step of predicting the motion of the body part during a collision using the motion prediction model based on the safety constraint information and the vehicle body acceleration information to obtain the predicted motion timing features includes: obtaining input feature data for the motion prediction model based on the safety constraint information and the vehicle body acceleration waveform; inputting the input feature data into the motion prediction model; and predicting the motion curve of the body part during the collision based on the input feature data in the motion prediction model to obtain the predicted acceleration curve of the body part.

[0009] In some implementations, the motion prediction model and the damage prediction model are obtained as follows: simulation data under various collision conditions are obtained through a vehicle collision simulation platform; a training dataset is obtained from the simulation data; a first initial model and a second initial model are trained using the training dataset to obtain a trained first initial model and a trained second initial model; and model lightweighting is performed on the trained first initial model and the trained second initial model to obtain the motion prediction model and the damage prediction model.

[0010] In some implementations, before predicting the motion of the occupant's body parts during the collision using the motion prediction model based on the collision data to obtain the predicted motion timing features of the body parts, the method further includes: determining that the occupant has obstacles to remote distress calls.

[0011] In some implementations, determining that the occupant has a remote distress signal obstacle includes: determining that the occupant cannot answer a call; before determining that the occupant of the accident vehicle cannot answer a call, the method further includes: assessing the accident that occurred in the accident vehicle to obtain the severity of the accident; if the severity of the accident is greater than a preset threshold, activating the vehicle-mounted emergency call system and the occupant injury prediction system, wherein the occupant injury prediction system includes the motion prediction model and the injury prediction model; and establishing a voice call between the accident vehicle and a remote emergency call service platform through the vehicle-mounted emergency call system.

[0012] In some implementations, the rescue party includes a vehicle alliance rescue platform, and the vehicle accident handling method further includes: sending one or more of the following information about the accident vehicle to the vehicle alliance rescue platform: vehicle location, vehicle status, surrounding road information and / or vehicle identification, so as to contact vehicles that meet the rescue conditions through the vehicle alliance rescue platform to rescue the accident vehicle.

[0013] According to a second aspect of this application, a vehicle accident handling device is provided, applied to an accident vehicle, wherein the accident vehicle is equipped with a pre-trained motion prediction model and a pre-trained damage prediction model. The vehicle accident handling device includes: an acquisition unit, configured to acquire collision data of the accident vehicle and position information of occupants in the accident vehicle; a motion prediction unit, configured to predict the motion of the occupants' body parts during the accident collision based on the collision data and the motion prediction model, obtaining predicted motion temporal features of the body parts; a damage prediction unit, configured to predict the damage suffered by the body parts during the accident collision based on the predicted motion temporal features and the damage prediction model, obtaining predicted damage results of the body parts; a report generation unit, configured to generate an occupant damage assessment report based on the occupant's position information and the predicted damage results of the body parts; and a sending unit, configured to send the occupant damage assessment report to the rescue party.

[0014] According to a third aspect of this application, an in-vehicle device is provided, including a memory and a processor; the memory is connected to the processor and is used to store a program; the processor is used to implement the vehicle accident handling method as described in the first aspect or any implementation thereof by running the program in the memory.

[0015] According to a fourth aspect of this application, a vehicle is provided, the vehicle including the on-board equipment as described in the third aspect.

[0016] According to a fifth aspect of this application, a storage medium is provided that stores a computer program, which, when executed by a processor, implements the vehicle accident handling method as described in the first aspect or any implementation thereof.

[0017] According to a sixth aspect of this application, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the vehicle accident handling method as described in the first aspect or any implementation thereof.

[0018] This application provides a method, apparatus, equipment, and product for handling vehicle accidents. The method involves deploying a motion prediction model and an injury prediction model on the accident vehicle. The motion prediction model obtains the predicted motion time-series characteristics of the occupants' body parts. Based on these characteristics, the injury prediction model obtains the predicted injury results for each body part. Based on the occupants' location information and the predicted injury results, an occupant injury assessment report is generated and sent to the rescue team. This proactively provides rescue teams with information on occupant injuries, facilitating effective rescue efforts and improving the efficiency and success rate of rescuing occupants from accident vehicles. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0020] Figure 1 A flowchart illustrating a vehicle accident handling method provided in this application embodiment. Figure 1 .

[0021] Figure 2 The diagram shows an example of the input and output of the motion prediction model and the injury prediction model.

[0022] Figure 3 A flowchart illustrating a vehicle accident handling method provided in this application embodiment. Figure 2 .

[0023] Figure 4 This is a flowchart of the occupant injury prediction system.

[0024] Figure 5 This is an example diagram of a rescue communication group in a traffic accident scenario.

[0025] Figure 6 This is an example diagram of a rescue communication group in a vehicle breakdown scenario.

[0026] Figure 7 This is a schematic diagram of the structure of a vehicle accident handling device provided in an embodiment of this application.

[0027] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0028] The technical solution proposed in this application is applicable to vehicle accident scenarios, aiming to enable the accident vehicle to proactively provide detailed and reliable information on occupant injuries to the rescue party, so that the rescue party can provide timely and effective rescue for the accident vehicle, thereby improving the efficiency and success rate of rescuing occupants in the accident vehicle.

[0029] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0030] Traditional vehicles rely primarily on passive safety features such as seat belts and airbags to protect occupants during a collision. After a collision, occupants request assistance from a remote rescue center through active safety systems. However, if an occupant is unconscious, their physical condition cannot be confirmed through voice communication between the occupant and rescue personnel, resulting in the occupant not receiving timely and accurate assistance.

[0031] To address the aforementioned issues, embodiments of this application provide a vehicle accident handling method, apparatus, equipment, and product. Based on collision data of the accident vehicle, a motion prediction model deployed on the accident vehicle is used to predict the motion of the occupants' body parts during the collision, obtaining the predicted motion time-series characteristics of the body parts. Based on the predicted motion time-series characteristics of the body parts, a damage prediction model deployed on the accident vehicle is used to predict the damage suffered by the body parts during the collision, obtaining the predicted damage results of the body parts. Combining the occupants' location information and the predicted damage results of the occupants' body parts, an occupant injury assessment report is proactively provided to the rescue team, enabling the rescue team to provide timely and effective rescue to the accident vehicle based on the occupant injury assessment report, thereby improving the efficiency and success rate of rescuing occupants in accident vehicles.

[0032] Exemplary methods Figure 1 A flowchart illustrating a vehicle accident handling method provided in this application embodiment. Figure 1 .like Figure 1 As shown, the vehicle accident handling method provided in this embodiment includes the following steps S101 to S105: S101, acquire collision data of the accident vehicle and the location information of the occupants of the accident vehicle.

[0033] The vehicle accident handling method provided in this application embodiment can be executed by in-vehicle equipment, such as vehicle-mounted equipment or control units related to vehicle safety.

[0034] The collision data of the accident vehicle is related to the physical impact it receives during the collision. The collision data may include one or more of the following: parameter information of the passive safety features inside the accident vehicle, information on the vehicle's positional changes during the collision, information on the vehicle's attitude changes during the collision, the collision location of the accident vehicle, or the impact force at the collision location.

[0035] Passive safety features may include seat belts and / or airbags.

[0036] The location information of the occupants of the accident vehicle may include the occupants' location information before the accident and / or the occupants' location information after the accident.

[0037] In this embodiment, the physical impact of the accident vehicle during the collision can be detected by the vehicle sensors of the accident vehicle, and the collision data of the accident vehicle can be obtained; the position information of the occupants of the accident vehicle can also be detected by the vehicle sensors of the accident vehicle.

[0038] In one example, the vehicle sensors used to collect collision data may include one or more of the following: an acceleration sensor, a pressure sensor, an angular velocity sensor, a displacement sensor, an airbag sensor, or a seatbelt status sensor. An acceleration sensor can be used to collect the instantaneous acceleration of the vehicle during a collision; a pressure sensor can be used to detect the pressure changes caused by the deformation of the vehicle's body structure during a collision, obtaining the collision location and the impact force at that location; an angular velocity sensor can be used to collect the rotational angular velocity of the vehicle during a collision; a displacement sensor can be used to collect the displacement of the vehicle during a collision; an airbag sensor can be used to collect airbag parameters during a collision, such as airbag acceleration, airbag pressure, and / or airbag deployment time, where the airbag deployment time can be the airbag ignition (gas generator ignition) time; a seatbelt status sensor can be used to collect seatbelt parameters during a collision, such as seatbelt insertion status, seatbelt force limiter rating, and / or seatbelt deployment time, where the seatbelt deployment time can be the seatbelt pretensioner trigger time. Therefore, by using one or more of the following sensors: acceleration sensor, pressure sensor, angular velocity sensor, displacement sensor, airbag sensor, or seat belt status sensor, the accuracy of collision data acquisition can be improved.

[0039] In one example, vehicle sensors used to collect occupant location information may include one or more of the following: seat pressure sensors, cameras, infrared sensors, radar sensors, or seat belt sensors. The location information of occupants in an accident vehicle can be detected using one or more of these sensors, improving the accuracy of occupant location information.

[0040] S102, Based on the collision data, the motion prediction model is used to predict the motion of the occupant's body parts during the collision, and the predicted motion time sequence characteristics of the body parts are obtained.

[0041] Among them, the accident vehicle is equipped with a pre-trained motion prediction model and a pre-trained damage prediction model. Both the motion prediction model and the damage prediction model can adopt neural network models, thereby utilizing the powerful modeling capabilities of neural network models to improve the accuracy of the motion prediction model in predicting the motion of body parts in an accident collision, and to improve the accuracy of the damage prediction model in predicting the damage of body parts in an accident collision.

[0042] A collision is a dynamic process that begins when a vehicle comes into contact with a collision object and ends when the two separate, a process that lasts for a very short period. The predicted motion time-series characteristics of body parts can indicate the predicted motion changes of those body parts during the collision period, such as predicted speed changes and predicted displacement changes.

[0043] The predicted motion changes of body parts during the collision period reflect the physical impact received by the body parts, and the physical impact determines the damage suffered by the body parts. For example, the more dramatic the predicted motion changes of a body part, the greater the physical impact and the greater the damage. Therefore, we can first obtain the temporal characteristics of the predicted motion of body parts through a motion prediction model, and then provide these temporal characteristics to a damage prediction model. The damage prediction model can then obtain the predicted damage results for the body parts. By utilizing the relationship between the predicted motion changes of body parts, the physical impact received by the body parts, and the damage suffered by the body parts, we can improve the accuracy of predicting the damage suffered by body parts in a collision.

[0044] Among them, there can be one or more body parts; the body parts can be key body parts of the occupant, such as the head, chest, shoulders, neck, abdomen, hands, and legs.

[0045] In this embodiment, the collision data of the accident vehicle can be input into the motion prediction model, or data related to the movement of the occupant's body parts can be obtained from the collision data of the accident vehicle and input into the motion prediction model. In the motion prediction model, the movement of the occupant's body parts in the accident collision is predicted based on the input data to obtain the predicted motion time sequence characteristics of the occupant's body parts.

[0046] In cases where the vehicle involved in the accident has multiple occupants, the predicted motion temporal features of each occupant's body parts can be obtained. When there are multiple body parts, the predicted motion temporal features corresponding to each of the multiple body parts can be obtained.

[0047] S103, Based on the predicted motion time sequence characteristics, the damage to body parts during an accident collision is predicted using a damage prediction model, and the predicted damage results for the body parts are obtained.

[0048] The predicted damage results for body parts can include the predicted damage value of the body part. The predicted damage value is related to the degree of damage; for example, the higher the predicted damage value, the more severe the damage.

[0049] Alternatively, the predicted damage results for body parts can be described in natural language to indicate the degree of damage, such as no damage, minor damage, moderate damage, or severe damage.

[0050] In this embodiment, the predicted motion time-series features of body parts can be input into the injury prediction model, or the predicted motion time-series features of body parts can be preprocessed before being input into the injury prediction model. The injury prediction model can predict the damage to body parts during a collision based on the input data, thus obtaining the predicted injury result for the body parts.

[0051] In cases where the vehicle involved in the accident has multiple occupants, the predicted injury results for each occupant's body parts can be obtained. When there are multiple body parts, predicted injury results for each body part can be obtained separately. Furthermore, considering the correlation between multiple body parts—for example, a violent head movement can lead to neck injury—the predicted motion time-series features corresponding to multiple body parts can be input together into the injury prediction model. The injury prediction model then predicts the injuries suffered by multiple body parts in a collision, obtaining predicted injury results for each body part separately. In this process, the injury prediction model utilizes the learned correlations between body parts to improve the accuracy of injury prediction for multiple body parts.

[0052] S104. Generate an occupant injury assessment report based on the occupant's location information and the predicted injury results of the occupant's body parts.

[0053] In one example, the occupant's location information, the name of the occupant's body part, and the predicted injury result of the occupant's body part can be combined to obtain an occupant injury assessment report. The occupant injury report may include one or more sets of occupant information, with different sets of occupant information corresponding to different occupants. Each set of occupant information may include the occupant's location information, the predicted injury result of occupant A's head, the predicted injury result of occupant A's chest, etc.

[0054] In another example, based on the predicted injury results of the occupant's body parts, the location and severity of the injury can be determined. The occupant's location information, the name of the injured body part, and the severity of the injury are combined to generate an occupant injury assessment report. This report can include one or more sets of occupant information, with different sets corresponding to different occupants. Each set of information may include the occupant's location, the name of the injured body part, and the severity of the injury. This approach avoids including redundant information (information about uninjured body parts) in the occupant injury report, ensuring that it contains the necessary information for accurate rescue while remaining concise and clear, thus improving the efficiency with which rescuers can glean important information from the report.

[0055] S105, send the occupant injury assessment report to the rescue team.

[0056] In this embodiment, an occupant injury assessment report is sent to the rescue team. Based on this report, the rescue team can promptly and specifically prepare professional first aid plans, prepare professional first aid resources (e.g., if a passenger's head injury is predicted, a rescue unit equipped with neurosurgical emergency equipment will be prioritized), and / or provide remote self-rescue guidance to the occupants. First aid resources may include medical and / or fire-fighting resources. Preparing first aid resources includes arranging ambulances, medical personnel, medical equipment, fire trucks, and firefighters. Preparing professional first aid resources includes instructing occupants on the correct posture for evacuation from the vehicle, instructing occupants to remain still, instructing occupants on how to stop bleeding from injured areas, and informing occupants of the exact waiting time for rescue.

[0057] In one example, an on-board accident emergency call system (AECS) is deployed on the accident vehicle. The rescue party may include the remote call center of the AECS and send the occupant injury assessment report to the remote call center of the AECS so that the remote call center can prepare a timely and targeted emergency plan and / or emergency resources according to the occupant injury assessment report.

[0058] Optionally, the occupant injury assessment report can be forwarded to the emergency center of the medical institution and / or the rescue platform of the traffic accident rescue organization through the remote call center of AECS, so that the emergency center of the medical institution and / or the rescue platform of the traffic accident rescue organization can obtain the injury information of the occupants in the accident vehicle in advance, make targeted rescue preparations, and improve rescue efficiency and occupant survival rate.

[0059] In another example, the rescue party may include the emergency center of a medical institution and / or the rescue platform of a traffic accident rescue organization. The accident vehicle may send an occupant injury assessment report to the emergency center of the medical institution and / or the rescue platform of the traffic accident rescue organization so that the emergency center of the medical institution and / or the rescue platform of the traffic accident rescue organization can obtain the injury information of the occupants in the accident vehicle in advance, make targeted rescue preparations, and improve rescue efficiency and occupant survival rate.

[0060] In this embodiment, the correlation between the movement of body parts during a collision and the injury status of body parts during a collision is utilized. By combining motion prediction models and injury prediction models, motion prediction and injury prediction of body parts of occupants in a collision are performed. This improves the accuracy of the predicted injury results for body parts, thereby improving the accuracy of the occupant injury assessment report. This allows rescuers to make targeted rescue preparations and / or rescue guidance based on the occupant injury assessment report, improving the efficiency of rescuing passengers in collision vehicles and increasing the survival rate of occupants in traffic accidents.

[0061] Below, we provide some possible implementation methods for predicting movement in different body parts.

[0062] In one possible implementation, safety constraint information of the occupants and vehicle acceleration information are extracted from the collision data of the accident vehicle. The safety constraint information includes one or more of the following: seatbelt exit position, seatbelt force limiter level, seatbelt deployment time, or airbag deployment time. Based on the safety constraint information and vehicle acceleration information, a motion prediction model is used to predict the motion of body parts during the collision, obtaining the predicted motion timing characteristics of the body parts. The occupant's safety constraint information reflects the triggering status of passive safety facilities at the occupant's location and the occupant's body posture. The vehicle acceleration information is crucial data related to the motion of the occupant's body parts. Providing the occupant's safety constraint information and the vehicle acceleration information to the motion prediction model can effectively improve the accuracy of the motion prediction model in predicting the motion of body parts during a collision.

[0063] The location of the seatbelt exit point is related to the seatbelt design. For example, some seatbelts have their exit point located above the seat back, while others have it located on the side of the seat cushion. The seatbelt exit point affects the occupant's initial body posture and the direction in which the seatbelt exerts restraining force on the occupant's body during a sudden change in body posture during a collision.

[0064] The seatbelt force limiter rating is a measure of the strength of the restraining force exerted by the seatbelt on the occupant's body due to sudden changes in body position during a collision. The strength of the restraining force exerted by the seatbelt on the occupant's body affects the movement of parts of the occupant's body.

[0065] The seatbelt deployment time and airbag deployment time can be referred to the explanation in the foregoing embodiments, and will not be repeated here. The seatbelt deployment time and airbag deployment time also affect the movement of the occupant's body parts: after the seatbelt deployment time, the restraining force of the seatbelt on the occupant's body parts will slow down the movement of the occupant's body parts; after the airbag deployment time, the resistance of the airbag on the occupant's body parts will also slow down the movement of the occupant's body parts.

[0066] In this implementation, the safety restraint information of the occupants and the vehicle's acceleration information are identified and extracted from the collision data of the accident vehicle. Both safety restraint information and vehicle acceleration information are temporal information. Safety restraint information includes one or more of the following during the collision time period: seatbelt exit position, seatbelt force limiter level, seatbelt deployment time, or airbag deployment time. Vehicle acceleration information includes the vehicle's acceleration during the collision time period, such as the B-pillar acceleration of the accident vehicle. During a collision, vehicle acceleration causes changes in occupant posture, and safety restraint information constrains these changes. Therefore, by combining vehicle acceleration information and safety restraint information, a motion prediction model is used to predict the motion of body parts during a collision, obtaining the predicted temporal characteristics of the body parts' motion. This effectively improves the accuracy of predicting the motion of body parts during a collision and enhances the accuracy of the predicted temporal characteristics of the body parts' motion.

[0067] Optionally, the vehicle body acceleration information includes the vehicle body acceleration waveform, and the predicted motion timing features of body parts include the predicted acceleration curves of body parts. Based on the occupant's safety constraint information and the vehicle body acceleration information, a motion prediction model is used to predict the motion of the occupant's body parts during the collision, obtaining the predicted motion timing features of the occupant's body parts. This includes: obtaining input feature data for the motion prediction model based on the occupant's safety constraint information and the vehicle body acceleration waveform; inputting the input feature data into the motion prediction model, and then predicting the motion curves of the occupant's body parts during the collision based on the input feature data, thus obtaining the predicted acceleration curves of the occupant's body parts. Therefore, the predicted acceleration curves of body parts more vividly and concretely demonstrate the changes in acceleration of body parts during a collision. These changes in acceleration reflect the changes in impact force experienced by the body parts, which in turn determine whether the body parts are damaged and the degree of damage, thus improving the accuracy of subsequent damage prediction.

[0068] Among them, the vehicle body acceleration waveform is, for example, the B-pillar acceleration curve.

[0069] In this optional approach, both the occupant safety restraint information and the vehicle's acceleration waveform are time-series information. Features of the occupant safety restraint information and the vehicle's acceleration waveform can be concatenated according to the timeline corresponding to the collision to obtain the input feature data for the motion prediction model. For example, features of the seatbelt exit position, seatbelt force limiter level, seatbelt deployment time, airbag deployment time, and B-pillar acceleration at the same moment can be concatenated according to the timeline corresponding to the collision to obtain multi-dimensional feature vectors corresponding to multiple moments. The input feature data of the motion prediction model includes these multi-dimensional feature vectors corresponding to these multiple moments. Then, the input feature data is input into the motion prediction model, which predicts the motion curves of body parts during the collision based on the input feature data, obtaining the predicted acceleration curves of the body parts.

[0070] Below, we provide some possible implementation methods for the application process of the injury prediction model, as well as the training process of the motion prediction model and the injury prediction model.

[0071] In one possible implementation, when the predicted motion time-series features of a body part include its predicted acceleration curve, the predicted acceleration curve can be input into a damage prediction model. The damage prediction model then extracts features from the predicted acceleration curve and predicts the damage to the body part in a collision based on these extracted features, thus obtaining the predicted damage result. Therefore, by utilizing the predicted acceleration curve of the body part, the accuracy of the damage prediction model in predicting the damage to the body part in a collision can be improved.

[0072] As an example, Figure 2 The diagram shows an example of the input and output of a motion prediction model and a damage prediction model. Figure 2 As shown, safety constraint information, including seat belt exit location, seat belt force limiter level, seat belt deployment time and / or airbag deployment time, and vehicle acceleration information can be input into the motion prediction model. The motion prediction model then predicts the acceleration curves of the occupant's head, chest, and shoulder, among other body parts. The acceleration curves of the body parts are then input into the injury prediction model, which predicts the degree of head injury, chest injury, and shoulder injury, among other injury results.

[0073] In one possible implementation, the damage prediction model adopts the Transformer model, which leverages the Transformer model's ability to capture long-term dependencies when processing time-series data to improve the damage prediction model's ability to process the time-series features of body parts and improve the accuracy of damage prediction.

[0074] In one possible implementation, the motion prediction model and damage prediction model are obtained as follows: Simulation data under various collision conditions is obtained through a vehicle collision simulation platform; a training dataset is acquired from the simulation data; a first initial model and a second initial model are trained using the training dataset to obtain the trained first initial model and the trained second initial model; the trained first initial model and the trained second initial model are then subjected to model lightweighting processing to obtain the motion prediction model and the damage prediction model. On the one hand, the simulation platform provides rich and reliable training data for the motion prediction model and the damage prediction model, improving the model training effect; on the other hand, the model lightweighting process reduces the size, computational load, and memory consumption of the motion prediction model and the damage prediction model, enabling them to be deployed and run in vehicles.

[0075] The first initial model is the initial model corresponding to the motion prediction model. The motion prediction model is obtained by training the first initial model and performing model lightweighting. The second initial model is the initial model corresponding to the injury prediction model. The injury prediction model is obtained by training the second initial model and performing model lightweighting.

[0076] Multiple collision conditions can refer to various collision scenarios, where the conditions under which vehicles collide differ, such as the different objects involved, vehicle speeds, and / or collision locations. A vehicle collision simulation platform can be used to simulate multiple collision conditions, obtaining simulation data under various scenarios and improving the richness of the simulation data.

[0077] The first initial model and the second initial model can be trained separately to focus on improving the motion prediction accuracy of the motion prediction model and the damage prediction accuracy of the damage prediction model.

[0078] In this implementation, simulated collision data of the simulated vehicle, the simulated motion temporal features of the simulated occupants' body parts, and the simulated injury results of the simulated occupants' body parts can be obtained from simulation data under various collision conditions. The training dataset includes simulated collision data of the simulated vehicle, the simulated motion temporal features of the simulated occupants' body parts, and the simulated injury results of the simulated occupants' body parts. The training dataset can be split into two datasets: a first dataset and a second dataset. The first dataset includes simulated collision data of the simulated vehicle and the simulated motion temporal features of the simulated occupants' body parts, while the second dataset includes the simulated motion temporal features of the simulated occupants' body parts and the simulated injury results of the simulated occupants' body parts. Based on the first dataset, a first initial model can be trained in a supervised manner to obtain a trained first initial model. In this supervised training, the simulated collision data of the simulated vehicle serves as the training samples, and the simulated motion temporal features of the simulated occupants' body parts serve as the sample labels. Subsequently, the trained first initial model undergoes model lightweighting processing to obtain a motion prediction model. Based on the second dataset, a second initial model can be trained in a supervised manner to obtain a trained second initial model. In this supervised training, the simulated motion temporal features of the simulated occupants' body parts in the simulated vehicle serve as training samples, and the simulated damage results of the simulated occupants' body parts in the simulated vehicle serve as sample labels. Subsequently, the trained second initial model undergoes model lightweighting to obtain a damage prediction model.

[0079] Optionally, during the acquisition of the training dataset, traffic accident case data can also be collected. Combining the simulation data and the traffic accident case data yields the training dataset. On the one hand, traffic accident case data provides realistic and reliable training data; on the other hand, simulation data increases the volume and richness of the training data.

[0080] Optionally, the random forest algorithm is used to train the first and second initial models. During training, the random forest algorithm is used to extract features, perform variable importance analysis, and conduct multi-decision tree ensemble learning on the training dataset, which effectively improves the generalization ability and prediction accuracy of the trained first and second initial models for complex collision scenarios.

[0081] Optionally, the model lightweighting process includes: converting the floating-point parameters in the first initial model and the second initial model from a first precision representation to a second precision representation to reduce the model storage overhead of the first initial model and the second initial model, wherein the data precision of the first precision representation is higher than that of the second precision representation; and / or performing pruning operations on the first initial model and the second initial model to reduce the number of model parameters of the first initial model and the second initial model.

[0082] In some embodiments, before predicting the motion of occupants' body parts during a collision using a motion prediction model based on the collision data of the accident vehicle, and obtaining the predicted motion sequence characteristics of the body parts, it can be determined that the occupants of the accident vehicle face obstacles to remote assistance. Therefore, in scenarios where the occupants of the accident vehicle face obstacles to remote assistance, the aforementioned vehicle accident handling method proactively provides occupant injury assessment reports to the rescue team, enabling the occupants to receive timely and effective assistance.

[0083] Among these, the occupants' inability to remotely call for help may include situations such as the occupants being unconscious, confused, unable to speak, and / or being far from the active distress call device of the accident vehicle.

[0084] In one example, the occupants of the accident vehicle may be unable to remotely call for help. In such cases, none of the occupants can provide information about their own or others' injuries to the rescuers. Therefore, it is necessary to proactively provide the rescuers with an occupant injury assessment report through the aforementioned vehicle accident handling methods.

[0085] In another example, the occupants of the accident vehicle may have remote distress obstacles. This could mean that any occupant of the accident vehicle has a remote distress obstacle. For the occupant with a remote distress obstacle, other occupants may also be trapped in their seats and the injury status of the occupant with a remote distress obstacle cannot be determined. In this case, the occupant injury assessment report can be proactively provided to the rescue party through the vehicle accident handling method described above.

[0086] Figure 3 A flowchart illustrating a vehicle accident handling method provided in this application embodiment. Figure 2 .like Figure 3 As shown, the vehicle accident handling method includes the following steps S301 to S309: S301, obtain collision data of the accident vehicle and the location information of the occupants of the accident vehicle.

[0087] The implementation principle and technical effects of S301 are the same as those in the aforementioned embodiments, and will not be repeated here.

[0088] S302 assesses the severity of an accident involving a vehicle.

[0089] In this embodiment, the severity of the accident can be assessed based on the vehicle speed during the collision, the vehicle's state after the collision (such as the state of the vehicle body structure, engine state, and airbag state), and / or the collision object of the vehicle.

[0090] In one example, the acceleration of the airbag control unit (ACU) in the accident vehicle can be obtained. The acceleration of the ACU indicates the severity of the accident, thereby improving the accuracy of accident severity assessment for the accident vehicle.

[0091] In this example, the ACU can be pre-calibrated using the peak acceleration method. During a collision, the ACU controls the airbag deployment time through an ignition control algorithm: the ACU filters its own acceleration along the X and Y axes to obtain its own acceleration, where the X and Y axes are the coordinate axes in the vehicle's coordinate system. If the ACU's acceleration exceeds a preset acceleration threshold, the airbag is deployed. Therefore, the ACU's acceleration accurately reflects the degree of collision, i.e., the severity of the accident.

[0092] S303: When the severity of an accident exceeds a preset threshold, establish a voice call between the accident vehicle and a remote emergency call service platform.

[0093] In this embodiment, if the severity of the accident exceeds a preset threshold, a call can be made to a remote emergency call service platform through the vehicle-mounted terminal on the accident vehicle. After the call is successful, a voice call is established between the accident vehicle and the remote emergency call service platform.

[0094] Optionally, an accident severity level exceeding a preset threshold includes situations where the ACU's acceleration exceeds a preset acceleration threshold. This utilizes ACU acceleration to improve the accuracy of accident severity assessment.

[0095] Optionally, after establishing a voice call between the accident vehicle and the remote emergency call service platform, the accident vehicle can send its location and status to the emergency call service platform so that the emergency call platform can promptly carry out rescue based on the vehicle's status and location, such as dispatching a fire truck to the rescue if the vehicle is severely deformed.

[0096] S304, Did the passengers in the vehicle involved in the accident respond to calls?

[0097] Whether the passengers in the accident vehicle responded to calls reflects whether they had any obstacles to remotely seeking help, such as whether they were unconscious, confused, or unable to speak.

[0098] In this embodiment, it can be determined whether passengers in the accident vehicle have responded to calls by detecting the voices of passengers inside the accident vehicle and / or the voices of staff sent by the emergency call service platform. For example, if the voices of passengers inside the accident vehicle are detected and the content of their speech meets the response requirements, it is determined that the passengers in the accident vehicle have responded to calls; otherwise, it is determined that the passengers in the accident vehicle have not responded to calls. Similarly, if the voices of staff sent by the emergency call service platform indicate that no response has been received from passengers, it is determined that the passengers in the accident vehicle have not responded to calls.

[0099] In this embodiment, if the passengers of the accident vehicle respond to a call, it is determined that the passengers of the accident vehicle do not have any obstacles to remote distress, and S309 can be executed; otherwise, it is determined that the passengers of the accident vehicle have obstacles to remote distress, and S305 can be executed.

[0100] Optionally, an onboard emergency call system (AECS) and an occupant injury prediction system are deployed on the accident vehicle. When the severity of the accident exceeds a preset threshold, both the AECS and the occupant injury prediction system are activated. The occupant injury prediction system includes a motion prediction model and an injury prediction model. A voice call is established between the accident vehicle and a remote emergency call service platform via the AECS. Thus, when the severity of the accident exceeds the preset threshold, both the AECS and the occupant injury prediction system are activated simultaneously, achieving a two-in-one emergency response mechanism. If the AECS detects that passengers are not responding to calls, the motion prediction model and injury prediction model in the occupant injury prediction system are used to predict occupant injuries in a timely manner, improving the efficiency of rescuing passengers in the accident vehicle.

[0101] The occupant injury prediction system can perform tasks related to the motion prediction model and the injury prediction model, which can be referred to in the aforementioned embodiments.

[0102] As an example, Figure 4 A flowchart of the occupant injury prediction system, such as Figure 4 As shown, the workflow of the occupant injury prediction system includes: obtaining a training dataset based on simulation data and traffic accident case data; training a motion prediction model and an injury prediction model based on the training dataset; using the motion prediction model to predict the acceleration curves of various body parts of the occupant; providing the acceleration curves of various body parts of the occupant to the injury prediction model; using the injury prediction model to predict the injury status of various body parts of the occupant; and finally generating an occupant injury assessment report; based on the occupant injury assessment report, preparing rescue plans, preparing rescue resources, and providing occupant self-rescue guidance.

[0103] S305. Based on the collision data, a motion prediction model is used to predict the motion of the occupants' body parts in the accident vehicle during the collision, and the predicted motion time sequence characteristics of the body parts are obtained.

[0104] S306, Based on the predicted motion time sequence characteristics of body parts, the damage to body parts in an accident collision is predicted using a damage prediction model, and the predicted damage results of the body parts are obtained.

[0105] S307 generates an occupant injury assessment report based on the occupant's location information and predicted injury results for body parts.

[0106] S308 sent an occupant injury assessment report to the rescue team.

[0107] The implementation principles and technical effects of S304 to S308 are the same as those in the aforementioned embodiments and will not be repeated here.

[0108] S309, through voice communication between the accident vehicle and a remote emergency call service platform, reports the occupant injury situation to the remote emergency call service platform.

[0109] In this embodiment, when the occupants of the accident vehicle respond to a call, the occupants and the staff of the emergency call service platform can assess the occupants' injuries through voice interaction, and can also communicate and coordinate the optimal rescue plan in real time.

[0110] In this embodiment, by assessing the severity of the accident, a voice call is promptly established between the accident vehicle and a remote emergency call platform. It is determined whether the passenger responds to the call during the voice call. If the passenger is unable to respond, a passenger injury assessment report is generated through motion prediction model and injury prediction model, and the passenger injury assessment report is proactively provided to the rescue party, so that the passenger can receive timely and effective assistance and improve the passenger rescue success rate.

[0111] In some embodiments, the vehicle location can be sent to the rescue party so that the rescue party can accurately reach the accident site to carry out the rescue based on the vehicle location.

[0112] In some embodiments, vehicle status can be sent to the rescue party so that the rescue party can arrange rescue personnel, rescue equipment and rescue plans according to the vehicle status, thereby improving rescue efficiency and success rate.

[0113] In some embodiments, the rescuer includes a vehicle alliance rescue platform, and further includes sending one or more of the following information about the accident vehicle to the vehicle alliance rescue platform: vehicle location, vehicle status, surrounding road information and / or vehicle identification, so as to contact vehicles that meet the rescue conditions through the vehicle alliance rescue platform to rescue the accident vehicle, thereby improving the efficiency of the accident vehicle obtaining effective rescue and the success rate of rescue.

[0114] Among them, surrounding road information refers to road information near the location of the accident vehicle, which may include road closure information, road congestion information, road hazard information, etc.

[0115] Vehicle identification may include license plate number, body color, vehicle brand, and vehicle logo.

[0116] Vehicles eligible for assistance may include those registered on the vehicle alliance rescue platform. The vehicle alliance rescue platform may include rescue characteristic information for multiple registered vehicles, indicating the rescue resources that the registered vehicle can provide.

[0117] In this embodiment, the vehicle alliance rescue platform can determine the necessary rescue resources for the accident vehicle based on the occupant injury assessment report, and then contact registered vehicles that meet the rescue criteria to provide assistance. And / or, the vehicle alliance rescue platform can contact nearby vehicles to provide assistance based on the accident vehicle's location information.

[0118] In one example, the vehicle alliance rescue platform can search for vehicles that meet the rescue conditions among multiple registered vehicles based on the rescue resources needed by the accident vehicle and the rescue feature information corresponding to multiple registered vehicles. It can then contact the vehicles that meet the rescue conditions to provide rescue for the accident vehicle, thereby finding registered vehicles that can provide the corresponding rescue resources so that the accident vehicle can be rescued in a timely manner.

[0119] In this example, by matching the rescue resources needed by the accident vehicle with the rescue feature information corresponding to multiple registered vehicles, a registered vehicle possessing at least some of the rescue resources required by the accident vehicle can be identified. This registered vehicle possessing at least some of the rescue resources required by the accident vehicle can be identified as a vehicle meeting the rescue criteria; alternatively, the current location of the registered vehicle possessing at least some of the rescue resources required by the accident vehicle can be obtained, and based on this current location and the location of the accident vehicle, a vehicle meeting the rescue criteria can be identified from among the registered vehicles possessing at least some of the rescue resources required by the accident vehicle, for example, a vehicle whose distance from the accident vehicle is less than a distance threshold can be selected as a vehicle meeting the rescue criteria.

[0120] Optionally, rescue characteristic information may include one or more of the following: medical resources (such as first aid kits for external injuries, medical emergency rescue equipment) configured on the registered vehicle, fire-fighting resources (such as fire extinguishers) configured on the registered vehicle, vehicle repair resources configured on the registered vehicle, and rescue qualifications (such as medical qualifications, fire-fighting qualifications, and first aid qualifications) possessed by the owner of the registered vehicle. The rescue resources required by the accident vehicle include one or more of the following: medical resources, fire-fighting resources, vehicle repair resources, and rescue personnel. One or more of the following: medical resources, fire-fighting resources, vehicle repair resources, and rescue personnel required by the accident vehicle can be matched with the above-mentioned rescue characteristic information of multiple registered vehicles to determine vehicles that meet the rescue conditions.

[0121] In this optional approach, the vehicle alliance rescue platform can solicit the consent of registered vehicle owners in advance, inquiring whether they are willing to provide roadside assistance when conditions permit. Upon obtaining the owner's consent, the platform can collect the vehicle's rescue characteristic information and store it in a cloud database. The vehicle alliance rescue platform can also request authorization from the registered vehicle owner to obtain the vehicle's location information when the vehicle meets the rescue conditions, thereby improving rescue efficiency.

[0122] In another example, a vehicle alliance rescue platform can contact nearby vehicles to provide assistance based on the location of the accident vehicle using local area broadcast technology. Local area broadcast technology includes, for example, vehicle-to-everything (V2X) technology.

[0123] Optionally, after identifying a vehicle that meets the rescue criteria, the vehicle alliance rescue platform sends a rescue request to that vehicle. The rescue request may include the location of the accident vehicle and the rescue resources required by the accident vehicle. After receiving the rescue request, the vehicle that meets the rescue criteria may send a response message to the vehicle alliance rescue platform to agree to rescue or refuse rescue, so that the vehicle alliance rescue platform can promptly understand the situation of vehicles that can go to rescue.

[0124] Optionally, if the vehicle alliance rescue platform receives a response message from a vehicle that meets the rescue conditions and agrees to rescue, it can establish a rescue communication group. The rescue communication group can include the identification information of the vehicle agreeing to rescue, the rescue resources provided by the vehicle agreeing to rescue, and the communication methods provided by the vehicle agreeing to rescue (such as a call button, a call window, or a temporary communication number). The vehicle alliance rescue platform can send the rescue communication group to the vehicle agreeing to rescue. After receiving the rescue communication group, the vehicle agreeing to rescue can display the rescue communication group on its in-vehicle terminal interface, so that the vehicles agreeing to rescue can communicate with each other through the rescue communication group, thereby improving rescue efficiency.

[0125] It should be noted that, in addition to traffic accident scenarios, the vehicle alliance rescue platform can also be used for vehicle breakdown scenarios, providing assistance to vehicles that have broken down.

[0126] Figure 5 This is an example diagram of a rescue communication group in a traffic accident scenario, such as... Figure 5 As shown, the rescue communication group includes the license plate numbers of the vehicles that agree to the rescue (such as license plate number XXXXX, license plate number YYYYY, license plate number ZZZZZ) and the accident rescue resources of the vehicles that agree to the rescue (such as emergency medical equipment, reflective vests, first aid kits, fire extinguishers, first aiders, etc.).

[0127] Figure 6 Example diagram of a rescue communication group in a vehicle breakdown scenario, such as Figure 6 As shown, the rescue communication group includes the license plate number of the vehicle that agrees to the rescue (such as license plate number XXXXX, license plate number YYYYY, license plate number ZZZZZ) and the breakdown rescue resources of the vehicle that agrees to the rescue (such as jump starter, tripod, jack, fire extinguisher, spare tire, tire removal tools, etc.).

[0128] Optionally, the vehicle alliance rescue platform will send information about the roads surrounding the accident vehicle to public communication platforms in the surrounding area, so as to remind nearby vehicles to pay attention to safety and avoid similar accidents from happening.

[0129] Optionally, the vehicle alliance rescue platform will send information about the roads surrounding the accident vehicle to the map service provider so that the map service provider can replan the travel routes of users in the vicinity.

[0130] Exemplary device Corresponding to the above-described vehicle accident handling method, this application also provides a vehicle accident handling device.

[0131] Figure 7 This is a schematic diagram of a vehicle accident handling device provided in an embodiment of this application. The vehicle accident handling method is applied to an accident vehicle, and the accident vehicle is equipped with a pre-trained motion prediction model and a pre-trained damage prediction model. Figure 7 As shown, the vehicle accident handling device 700 provided in this application embodiment includes: an acquisition unit 701, a motion prediction unit 702, a damage prediction unit 703, a report generation unit 704, and a sending unit 705. The acquisition unit 701 is used to acquire collision data of the accident vehicle and the position information of the occupants in the accident vehicle; the motion prediction unit 702 is used to predict the motion of the occupants' body parts in the accident collision based on the collision data and through a motion prediction model, and obtain the predicted motion time sequence characteristics of the body parts; the injury prediction unit 703 is used to predict the injury suffered by the body parts in the accident collision based on the predicted motion time sequence characteristics and through an injury prediction model, and obtain the predicted injury results of the body parts; the report generation unit 704 is used to generate an occupant injury assessment report based on the occupant's position information and the predicted injury results of the body parts; and the sending unit 705 is used to send the occupant injury assessment report to the rescue party.

[0132] In some implementations, the motion prediction unit 702 is specifically used to: extract occupant safety constraint information and vehicle body acceleration information from collision data. The safety constraint information includes one or more of the following: seat belt exit position, seat belt force limit level, seat belt deployment time or airbag deployment time; based on the safety constraint information and vehicle body acceleration information, predict the motion of body parts in the accident collision using a motion prediction model to obtain predicted motion timing features.

[0133] In some implementations, the vehicle acceleration information includes the vehicle acceleration waveform, and the predicted motion timing features of the body parts include the predicted acceleration curves of the body parts; the motion prediction unit 702 is specifically used to: obtain the input feature data of the motion prediction model based on the safety constraint information and the vehicle acceleration waveform; input the input feature data into the motion prediction model, and predict the motion curves of the body parts in the accident collision based on the input feature data in the motion prediction model to obtain the predicted acceleration curves of the body parts.

[0134] In some implementations, the motion prediction model and the damage prediction model are obtained as follows: simulation data under various collision conditions are obtained through a vehicle collision simulation platform; a training dataset is obtained from the simulation data; the first initial model and the second initial model are trained using the training dataset to obtain the trained first initial model and the trained second initial model; the trained first initial model and the trained second initial model are subjected to model lightweighting processing to obtain the motion prediction model and the damage prediction model.

[0135] In some implementations, the vehicle accident handling device also includes a remote distress obstacle determination unit (not shown in the figure) for determining whether there is a remote distress obstacle for the occupants.

[0136] In some implementations, the remote distress obstacle determination unit is specifically used to: determine that the occupants cannot respond to calls; the vehicle accident handling device also includes: an accident assessment unit (not shown in the figure), used to assess the accident involving the vehicle and obtain the severity of the accident; a system activation unit (not shown in the figure), used to activate the vehicle-mounted emergency call system and the occupant injury prediction system when the severity of the accident exceeds a preset threshold, the occupant injury prediction system including a motion prediction model and an injury prediction model; and a call establishment unit (not shown in the figure), used to establish a voice call between the vehicle and the remote emergency call service platform through the vehicle-mounted emergency call system.

[0137] In some implementations, the rescue party includes a vehicle alliance rescue platform, and the sending unit 705 is also used to send one or more of the following information about the accident vehicle to the vehicle alliance rescue platform: vehicle location, vehicle status, surrounding road information and / or vehicle identification, so as to contact vehicles that meet the rescue conditions through the vehicle alliance rescue platform to rescue the accident vehicle.

[0138] The vehicle accident handling device provided in this embodiment belongs to the same concept as the vehicle accident handling method provided in the above embodiments of this application. It can execute the vehicle accident handling method provided in any of the above embodiments of this application and has the corresponding functional modules and beneficial effects for executing the vehicle accident handling method. Technical details not described in detail in this embodiment can be found in the specific processing content of the vehicle accident handling method provided in the above embodiments of this application, and will not be repeated here.

[0139] The functions implemented by the acquisition unit 701, motion prediction unit 702, damage prediction unit 703, report generation unit 704, and sending unit 705 can be implemented by the same or different processors, and this application embodiment does not limit them.

[0140] It should be understood that the units in the above device can be implemented by a processor calling software. For example, the device includes a processor connected to a memory containing instructions. The processor calls the instructions stored in the memory to implement any of the above methods or to implement the functions of each unit in the device. The processor can be a general-purpose processor, such as a CPU or microprocessor, and the memory can be internal or external to the device. Alternatively, the units in the device can be implemented as hardware circuits. By designing the hardware circuits, some or all of the unit functions can be implemented. The hardware circuits can be understood as one or more processors. For example, in one implementation, the hardware circuit is an ASIC, and the functions of some or all of the above units are implemented by designing the logical relationships between the components within the circuit. In another implementation, the hardware circuit can be implemented using a PLD, such as an FPGA, which can include a large number of logic gates. The connection relationships between the logic gates are configured through configuration files to implement the functions of some or all of the above units. All units in the above device can be implemented entirely by a processor calling software, entirely by hardware circuits, or partially by a processor calling software with the remaining parts implemented by hardware circuits.

[0141] In this application embodiment, a processor is a circuit with signal processing capabilities. In one implementation, the processor can be a circuit with instruction reading and execution capabilities, such as a CPU, microprocessor, GPU, or DSP. In another implementation, the processor can implement certain functions through the logical relationships of hardware circuits. These logical relationships are fixed or reconfigurable. For example, the processor may be a hardware circuit implemented as an ASIC or PLD, such as an FPGA. In a reconfigurable hardware circuit, the process of the processor loading a configuration document and configuring the hardware circuit can be understood as the processor loading instructions to implement the functions of some or all of the above units. Furthermore, it can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as an NPU, TPU, or DPU.

[0142] As can be seen, each unit in the above device can be one or more processors (or processing circuits) configured to implement the above methods, such as: CPU, GPU, NPU, TPU, DPU, microprocessor, DSP, ASIC, FPGA, or a combination of at least two of these processor forms.

[0143] Furthermore, the units in the above devices can be integrated in whole or in part, or they can be implemented independently. In one implementation, these units are integrated together and implemented in the form of a System-on-Chip (SoC). The SoC may include at least one processor for implementing any of the above methods or implementing the functions of the units in the device. The at least one processor may be of different types, such as CPU and FPGA, CPU and artificial intelligence processor, CPU and GPU, etc.

[0144] Exemplary System This application provides an electronic device, see [link to relevant documentation] Figure 8 As shown, the electronic device includes a memory 800 and a processor 810; wherein the memory 800 is connected to the processor 810 and is used to store programs; the processor 810 is used to implement the vehicle accident handling method disclosed in any of the above embodiments by running the programs stored in the memory 800.

[0145] Specifically, the aforementioned electronic device may also include: a bus, a communication interface 820, an input device 830, and an output device 840.

[0146] The processor 810, memory 800, communication interface 820, input device 830, and output device 840 are interconnected via a bus. Among them: A bus can include a pathway for transmitting information between various components of a computer system.

[0147] The processor 810 can be a general-purpose processor, such as a general-purpose central processing unit (CPU), a microprocessor, etc., or an application-specific integrated circuit (ASIC), or one or more integrated circuits used to control the execution of the program of the present invention. It can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), an off-the-shelf programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0148] The processor 810 may include a main processor, as well as a baseband chip, modem, etc.

[0149] The memory 800 stores a program that executes the technical solution of this invention, and may also store an operating system and other key business functions. Specifically, the program may include program code, which includes computer operation instructions. More specifically, the memory 800 may include read-only memory (ROM), other types of static storage devices capable of storing static information and instructions, random access memory (RAM), other types of dynamic storage devices capable of storing information and instructions, disk storage, flash memory, etc.

[0150] Input device 830 may include a device for receiving user input data and information, such as a keyboard, mouse, camera, scanner, light pen, voice input device, touch screen, pedometer, or gravity sensor.

[0151] Output device 840 may include devices that allow information to be output to a user, such as a display screen, printer, speaker, etc.

[0152] The communication interface 820 may include a device that uses any transceiver to communicate with other devices or communication networks, such as Ethernet, Radio Access Network (RAN), Wireless Local Area Network (WLAN), etc.

[0153] The processor 810 executes the program stored in the memory 800 and calls other devices, which can be used to implement each step of any of the vehicle accident handling methods provided in the above embodiments of this application.

[0154] Optionally, the electronic device is an in-vehicle device.

[0155] This application also proposes a vehicle equipped with the aforementioned on-board equipment.

[0156] This application also proposes a chip, which includes a processor and a data interface. The processor reads and runs a program stored in the memory through the data interface to execute the vehicle accident handling method described in any of the above embodiments. For the specific processing procedure and its beneficial effects, please refer to the embodiments of the above vehicle accident handling method.

[0157] Exemplary computer program products and storage media In addition to the methods and devices described above, embodiments of this application may also be computer program products, which include computer program instructions that, when executed by a processor, cause the processor to perform the steps in the vehicle accident handling methods according to various embodiments of this application as described in any of the above embodiments of this specification.

[0158] The computer program product can be written in any combination of one or more programming languages ​​to perform the operations of the embodiments of this application. The programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0159] Furthermore, embodiments of this application may also be storage media storing computer programs, which are executed by a processor to perform the steps of the vehicle accident handling methods according to various embodiments of this application described in any of the above embodiments of this specification, specifically implementing the steps of the vehicle accident handling methods as described above.

[0160] For the foregoing method embodiments, in order to simplify the description, they are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to this application, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0161] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For apparatus embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0162] The steps in the methods of the various embodiments of this application can be adjusted, merged, or deleted in order according to actual needs, and the technical features described in each embodiment can be replaced or combined.

[0163] The units of the apparatus in the various embodiments of this application can be merged, divided, and deleted according to actual needs.

[0164] It should be understood that the disclosed terminals, devices, and methods can be implemented in other ways, given the several embodiments provided in this application. For example, the terminal embodiments described above are merely illustrative. For instance, the division of modules or sub-modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple sub-modules or modules may be combined or integrated into another module, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or modules, and may be electrical, mechanical, or other forms.

[0165] The modules or submodules described as separate components may or may not be physically separate. The components that constitute a module or submodule may or may not be physical modules or submodules; that is, they may be located in one place or distributed across multiple network modules or submodules. Some or all of the modules or submodules can be selected to achieve the purpose of this embodiment's solution, depending on actual needs.

[0166] Furthermore, the functional modules or sub-modules in the various embodiments of this application can be integrated into one processing module, or each module or sub-module can exist physically separately, or two or more modules or sub-modules can be integrated into one module. The integrated modules or sub-modules described above can be implemented in hardware or in the form of software functional modules or sub-modules.

[0167] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0168] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software unit executed by a processor, or a combination of both. The software unit can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0169] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0170] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for handling vehicle accidents, characterized in that, Applied to accident vehicles, the accident vehicles are equipped with pre-trained motion prediction models and pre-trained damage prediction models, and the vehicle accident handling method includes: Obtain the collision data of the accident vehicle and the location information of the occupants in the accident vehicle; Based on the collision data, the motion prediction model is used to predict the motion of the occupant's body parts during the collision, thereby obtaining the predicted motion temporal features of the body parts. Based on the predicted motion time sequence characteristics, the damage to the body part in the accident collision is predicted by the damage prediction model to obtain the predicted damage result of the body part. Based on the occupant's location information and the predicted injury results of the body parts, an occupant injury assessment report is generated; Send the aforementioned occupant injury assessment report to the rescue team.

2. The vehicle accident handling method according to claim 1, characterized in that, The step of predicting the motion of the occupant's body parts during the collision using the motion prediction model based on the collision data, and obtaining the predicted motion temporal features of the body parts, includes: From the collision data, the occupant's safety restraint information and the vehicle's acceleration information are extracted. The safety restraint information includes one or more of the following: seat belt exit position, seat belt force limiter level, seat belt deployment time or airbag deployment time. Based on the safety constraint information and the vehicle acceleration information, the motion prediction model is used to predict the motion of the body parts during the accident collision, thereby obtaining the predicted motion time sequence features.

3. The vehicle accident handling method according to claim 2, characterized in that, The vehicle body acceleration information includes a vehicle body acceleration waveform, and the predicted motion time-series features include a predicted acceleration curve. The process of predicting the motion of the body parts during a collision using the motion prediction model based on the safety constraint information and the vehicle body acceleration information, to obtain the predicted motion time-series features, includes: Based on the safety constraint information and the vehicle body acceleration waveform, the input feature data of the motion prediction model is obtained; The input feature data is input into the motion prediction model, and the motion curve of the body part in the accident collision is predicted based on the input feature data to obtain the predicted acceleration curve of the body part.

4. The vehicle accident handling method according to any one of claims 1 to 3, characterized in that, The motion prediction model and the injury prediction model are obtained in the following ways: Simulation data under various collision conditions were obtained through a vehicle collision simulation platform. Obtain the training dataset from the simulation data; Using the training dataset, the first initial model and the second initial model are trained to obtain the trained first initial model and the trained second initial model. The trained first initial model and the trained second initial model are subjected to model lightweighting to obtain the motion prediction model and the damage prediction model.

5. The vehicle accident handling method according to any one of claims 1 to 3, characterized in that, Before predicting the motion sequence characteristics of the occupant's body parts during the collision using the motion prediction model based on the collision data, the method further includes: It was determined that the occupants had a remote distress signaling obstacle.

6. The vehicle accident handling method according to claim 5, characterized in that, The determination that the occupant has a remote distress signaling obstacle includes: It was determined that the occupant was unable to answer the call; Before determining that the occupants of the accident vehicle are unable to answer calls, the method further includes: The severity of the accident involving the vehicle was assessed. If the severity of the accident exceeds a preset threshold, the vehicle-mounted emergency call system and the occupant injury prediction system are activated. The occupant injury prediction system includes the motion prediction model and the injury prediction model. The vehicle-mounted emergency call system establishes a voice call between the accident vehicle and a remote emergency call service platform.

7. The vehicle accident handling method according to any one of claims 1 to 3, characterized in that, The rescue providers include the Vehicle Alliance Rescue Platform, and also include: Send one or more of the following information about the accident vehicle to the vehicle alliance rescue platform: vehicle location, vehicle status, surrounding road information and / or vehicle identification, so as to contact vehicles that meet the rescue conditions through the vehicle alliance rescue platform to rescue the accident vehicle.

8. A vehicle accident handling device, characterized in that, Applied to accident vehicles, the accident vehicles are equipped with pre-trained motion prediction models and pre-trained damage prediction models. The vehicle accident handling device includes: The acquisition unit is used to acquire the collision data of the accident vehicle and the location information of the occupants in the accident vehicle; The motion prediction unit is used to predict the motion of the occupant's body parts during the collision based on the collision data and the motion prediction model, so as to obtain the predicted motion time sequence features of the body parts. The damage prediction unit is used to predict the damage to the body part in the accident collision based on the predicted motion time sequence characteristics and the damage prediction model, so as to obtain the predicted damage result of the body part. The report generation unit is used to generate an occupant injury assessment report based on the occupant's location information and the predicted injury results of the body parts; The sending unit is used to send the occupant injury assessment report to the rescue team.

9. A vehicle-mounted device, characterized in that, Including memory and processor; The memory is connected to the processor and is used to store programs; The processor is used to implement the vehicle accident handling method as described in any one of claims 1 to 7 by running the program in the memory.

10. A storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the vehicle accident handling method as described in any one of claims 1 to 7.

11. A computer program product, characterized in that, The system includes a computer program that, when executed by a processor, implements the vehicle accident handling method as described in any one of claims 1 to 7.

Images