Data generation method and apparatus, computer device, and storage medium

Abstract

The present application relates to a data generation method and apparatus, a computer device, and a storage medium. The method comprises: acquiring scenario data of a target scenario, wherein there are an ego vehicle and a plurality of traffic participants around the ego vehicle in the target scenario; determining the attention level of each traffic participant on the basis of the scenario data; for each target traffic participant, when the attention level of the target traffic participant is greater than an attention level threshold, using a first simulated behavior model for the target scenario to generate first simulated behavior data of the target traffic participant for the target scenario; and when the attention level of the target traffic participant is not greater than the attention level threshold, using a second simulated behavior model for the target scenario to generate second simulated behavior data of the target traffic participant for the target scenario, the degree of anthropomorphism of the simulated behavior in the target scenario represented by the first simulated behavior data being higher than that of the simulated behavior in the target scenario represented by the second simulated behavior data.

Description

Data generation methods, apparatus, computer equipment and storage media

[0001] This application claims priority to Chinese Patent Application No. 2025100422092, filed on January 10, 2025, entitled “Data Generation Method, Apparatus, Computer Equipment and Storage Medium”, the entire contents of which are incorporated herein by reference. Technical Field

[0002] This application relates to the field of vehicle technology, specifically to data generation methods, apparatus, computer equipment, and storage media. Background Technology

[0003] Constructing anthropomorphic traffic flow models for specific scenarios is a core component of simulation testing to assess the performance of algorithms running on vehicles within those scenarios. Simulating the behavior of traffic participants in these scenarios is the core of building anthropomorphic traffic flow models. The simulated behavior of traffic participants in these scenarios is represented by simulated behavioral data; in essence, simulating the behavior of traffic participants in these scenarios involves generating simulated behavioral data that represents the simulated behavior of traffic participants in those scenarios.

[0004] In related technologies, the same simulated behavior model is used to generate simulated behavior data for each traffic participant in a given scenario, resulting in high costs for constructing anthropomorphic traffic flows. Therefore, avoiding the high cost of constructing anthropomorphic traffic flows for given scenarios has become a problem that needs to be solved. Summary of the Invention

[0005] One of the purposes of this application is to provide a data generation method, apparatus, computer equipment, and storage medium to address the problem of high costs associated with constructing anthropomorphic traffic flows for corresponding scenarios.

[0006] To achieve the above objectives, the technical solution adopted in this application is as follows:

[0007] In a first aspect, embodiments of this application provide a data generation method, including:

[0008] Acquire scene data for the target scenario, which includes the vehicle and multiple traffic participants around it.

[0009] Based on the scenario data, determine the attention level of each of the multiple traffic participants, and determine the target traffic participant among the multiple traffic participants;

[0010] For each target traffic participant, when the target traffic participant's attention level is greater than the attention level threshold, a first simulated behavior data for the target traffic participant in the target scenario is generated using a first simulated behavior model for the target scenario; when the target traffic participant's attention level is not greater than the attention level threshold, a second simulated behavior data for the target traffic participant in the target scenario is generated using a second simulated behavior model for the target scenario, wherein the degree of anthropomorphism of the simulated behavior in the target scenario represented by the first simulated behavior data is higher than the degree of anthropomorphism of the simulated behavior in the target scenario represented by the second simulated behavior data.

[0011] Based on the aforementioned technical methods, the attention level of traffic participants is related to the strength of their interaction with the vehicle in the target scenario. When the attention level of the target traffic participant exceeds the attention threshold, it indicates strong interaction between the target traffic participant and the vehicle in the target scenario. A first simulated behavior model for the target scenario, with a relatively high degree of anthropomorphism, is used to generate first simulated behavior data for the target traffic participant in the target scenario. When the attention level of the target traffic participant is not greater than the attention threshold, it indicates weak interaction between the target traffic participant and the vehicle in the target scenario. A second simulated behavior model for the target traffic participant in the target scenario, with a relatively low degree of anthropomorphism, is used to generate second simulated behavior data for the target traffic participant in the target scenario.

[0012] Therefore, for target traffic participants with strong interaction with their vehicles in the target scenario, a first simulated behavior model with a relatively high degree of anthropomorphism is used to simulate their behavior in the target scenario. For target traffic participants with weak interaction with their vehicles in the target scenario, a second simulated behavior model with a relatively low degree of anthropomorphism is used to simulate their behavior in the target scenario.

[0013] Compared to simulating the behavior of every traffic participant in a given scenario using a highly anthropomorphic model, constructing a human-like traffic flow is less costly. Conversely, compared to simulating the behavior of every traffic participant in a given scenario using a less anthropomorphic model, the constructed human-like traffic flow exhibits higher realism. Therefore, it balances the cost of constructing a human-like traffic flow with its realism, avoiding the problems of high cost and low realism in either scenario.

[0014] Furthermore, based on the scenario data, determining the attention level of each of the multiple traffic participants includes:

[0015] For each traffic participant, based on the scene data, the location importance of the traffic participant's location and the type importance of the traffic participant's type are determined. Based on the location importance and the type importance, the importance parameter of the traffic participant is determined. Based on the importance parameter of the traffic participant, the attention level of the traffic participant is determined.

[0016] Furthermore, based on the importance parameter of the traffic participants, the level of attention of the traffic participants is determined, including:

[0017] The right-of-way parameters of the traffic participants are determined based on whether the traffic participants are in the target lane of the vehicle, wherein the target lane of the vehicle is indicated by the scene data.

[0018] The attention level of the traffic participants is determined based on their importance parameters and right-of-way parameters.

[0019] Furthermore, based on the importance parameters and right-of-way parameters of the traffic participants, the attention level of the traffic participants is determined, including:

[0020] When the traffic participant is a vehicle, the risk parameters of the traffic participant are determined based on the associated parameter information used to determine the risk parameters of the traffic participant. The associated parameter information includes: the relative distance between the vehicle and the traffic participant, the vehicle speed, and the traffic participant speed, wherein the associated parameter information is indicated by the scene data.

[0021] The level of attention of the traffic participants is determined based on their importance parameters, right-of-way parameters, and risk parameters.

[0022] Furthermore, based on the associated parameter information used to determine the risk parameters of the traffic participants, the risk parameters of the traffic participants are determined as follows:

[0023] Obtain the spring damping model parameters for the traffic participant, the spring damping model parameters including: virtual longitudinal spring stiffness and virtual longitudinal damping coefficient for the traffic participant, and virtual lateral spring stiffness and virtual lateral damping coefficient for the traffic participant;

[0024] Based on the spring damping model parameters for the traffic participant and the associated parameter information used to determine the risk parameters of the traffic participant, the longitudinal risk value of the traffic participant relative to the vehicle and the lateral risk value of the traffic participant relative to the vehicle are calculated.

[0025] The risk parameters of the traffic participant are determined based on the longitudinal risk value of the traffic participant relative to the vehicle and the lateral risk value of the traffic participant relative to the vehicle.

[0026] Furthermore, based on the importance parameter, right-of-way parameter, and risk parameter of the traffic participant, the attention level of the traffic participant is determined, including:

[0027] The weighted sum of the importance parameter, right-of-way parameter, and risk parameter of the traffic participant is determined as the attention level of the traffic participant.

[0028] Furthermore, the target traffic participant among the plurality of traffic participants is identified as including:

[0029] The attention levels of the multiple traffic participants are ranked from highest to lowest according to their level of concern.

[0030] A predetermined number of traffic participants from among the multiple traffic participants are identified as target traffic participants.

[0031] Furthermore, the target scene can be any one of multiple scenes; and before acquiring the scene data of the target scene, the method further includes:

[0032] Clustering is performed on the label information of multiple scene data to determine the multiple scenes and the scene to which each scene data belongs.

[0033] Secondly, embodiments of this application also provide a data generation apparatus, including:

[0034] The acquisition unit is used to acquire scene data of the target scene, wherein the target scene contains the vehicle and multiple traffic participants around the vehicle;

[0035] The determining unit is configured to determine the attention level of each of the plurality of traffic participants and the target traffic participant among the plurality of traffic participants based on the scenario data.

[0036] The generation unit is configured to, for each target traffic participant, generate first simulated behavior data of the target traffic participant in the target scenario using a first simulated behavior model for the target scenario when the target traffic participant's attention level is greater than the attention level threshold; and generate second simulated behavior data of the target traffic participant in the target scenario using a second simulated behavior model for the target scenario when the target traffic participant's attention level is not greater than the attention level threshold. The first simulated behavior data represents a higher degree of anthropomorphism in the simulated behavior in the target scenario than the second simulated behavior data represents.

[0037] Furthermore, the determining unit is further configured to, for each traffic participant, determine the location importance of the traffic participant's location and the type importance of the traffic participant's traffic participant type based on the scene data, and determine the importance parameter of the traffic participant based on the location importance and the type importance, and determine the attention level of the traffic participant based on the importance parameter of the traffic participant.

[0038] Furthermore, the determining unit is further configured to determine the right-of-way parameters of the traffic participant based on whether the traffic participant is in the target lane where the vehicle is located, wherein the target lane where the vehicle is located is indicated by the scene data; and to determine the attention level of the traffic participant based on the importance parameters and right-of-way parameters of the traffic participant.

[0039] Furthermore, the determining unit is further configured to, when the traffic participant is a vehicle, determine the risk parameters of the traffic participant based on the associated parameter information used to determine the risk parameters of the traffic participant, wherein the associated parameter information includes: the relative distance between the vehicle and the traffic participant, the vehicle speed, and the traffic participant speed, wherein the associated parameter information is indicated by the scene data; and determine the attention level of the traffic participant based on the importance parameter, the right-of-way parameter, and the risk parameter of the traffic participant.

[0040] Furthermore, the determining unit is further configured to acquire spring-damping model parameters for the traffic participant, the spring-damping model parameters including: virtual longitudinal spring stiffness and virtual longitudinal damping coefficient for the traffic participant, and virtual lateral spring stiffness and virtual lateral damping coefficient for the traffic participant; based on the spring-damping model parameters for the traffic participant and the associated parameter information used to determine the risk parameters of the traffic participant, calculate the longitudinal risk value of the traffic participant relative to the vehicle and the lateral risk value of the traffic participant relative to the vehicle; and determine the risk parameters of the traffic participant based on the longitudinal risk value of the traffic participant relative to the vehicle and the lateral risk value of the traffic participant relative to the vehicle.

[0041] Furthermore, the determining unit is further configured to determine the traffic participant's attention level by weighting the importance parameter, right-of-way parameter, and risk parameter of the traffic participant.

[0042] Furthermore, the determining unit is further used to sort the attention of the multiple traffic participants from highest to lowest according to their attention level; and to determine the top preset number of traffic participants as target traffic participants.

[0043] Furthermore, the target scenario can be any one of multiple scenarios; the data generation device also includes:

[0044] The clustering unit is used to cluster the label information of multiple scene data before acquiring the scene data of the target scene, so as to determine the multiple scenes and the scene to which each scene data belongs.

[0045] Thirdly, embodiments of this application also provide a computer device, including:

[0046] The system includes a memory and a processor, which are communicatively connected to each other. The memory stores computer instructions, and the processor executes the computer instructions to perform the above-described method.

[0047] Fourthly, embodiments of this application also provide a computer-readable storage medium storing computer instructions for causing a computer to perform the above-described methods.

[0048] Fifthly, embodiments of this application also provide a computer program product, characterized in that it includes computer instructions, which are used to cause a computer to execute the above-described method.

[0049] The beneficial effects of this application are:

[0050] Traffic participant attention is correlated with the strength of interaction between the traffic participant and the vehicle in the target scenario. When the target traffic participant's attention exceeds a threshold, it indicates strong interaction between the target traffic participant and the vehicle in the target scenario. A first simulated behavior model, which simulates the traffic participant's behavior in the target scenario with a relatively high degree of anthropomorphism, is used to generate first simulated behavior data for the target traffic participant in the target scenario. When the target traffic participant's attention is below the threshold, it indicates weak interaction between the target traffic participant and the vehicle in the target scenario. A second simulated behavior model, which simulates the traffic participant's behavior in the target scenario with a relatively low degree of anthropomorphism, is used to generate second simulated behavior data for the target traffic participant in the target scenario.

[0051] Therefore, for target traffic participants with strong interaction with their vehicles in the target scenario, a first simulated behavior model with a relatively high degree of anthropomorphism is used to simulate their behavior in the target scenario. For target traffic participants with weak interaction with their vehicles in the target scenario, a second simulated behavior model with a relatively low degree of anthropomorphism is used to simulate their behavior in the target scenario.

[0052] Compared to simulating the behavior of every traffic participant in a given scenario using a highly anthropomorphic model, constructing a human-like traffic flow is less costly. Conversely, compared to simulating the behavior of every traffic participant in a given scenario using a less anthropomorphic model, the constructed human-like traffic flow exhibits higher realism. Therefore, it balances the cost of constructing a human-like traffic flow with its realism, avoiding the problems of high cost and low realism in either scenario. Attached Figure Description

[0053] Figure 1 is a schematic diagram of an example of simulating the behavior of traffic participants using a first simulated behavior model and a second simulated behavior model.

[0054] Figure 2 is a flowchart illustrating the data generation method provided in an embodiment of this application.

[0055] Figure 3 is a schematic diagram of an example of a preset orientation used to calculate the locational importance of traffic participants.

[0056] Figure 4 is a flowchart illustrating another data generation method provided in an embodiment of this application.

[0057] Figure 5 is a schematic diagram of the hardware structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0058] The embodiments of this application will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be understood that the preferred embodiments are only for illustrating this application and are not intended to limit the scope of protection of this application.

[0059] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. Therefore, the drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0060] Referring to Figure 1, it shows a schematic diagram of an example of simulating the behavior of traffic participants using a first simulated behavior model and a second simulated behavior model.

[0061] In this example, scene data of the target scene is obtained in order to construct an anthropomorphic traffic flow of the target scene.

[0062] Figure 1 shows two target traffic participants, namely traffic participant 1 and traffic participant 2, among all target traffic participants around the vehicle. Both traffic participant 1 and traffic participant 2 are vehicles.

[0063] In this context, all target traffic participants around the vehicle are indicated by scene data of the target scenario. As an example, the target scenario is one of the following: a dangerous intrusion scenario, an intersection scenario, a ramp scenario, or an obstacle avoidance scenario.

[0064] If the attention level of traffic participant 1 is greater than the attention threshold, simulated behavior data 1 of traffic participant 1 for the target scenario is generated using the first simulated behavior model for the target scenario.

[0065] If the attention level of traffic participant 1 is not greater than the attention level threshold, the second simulated behavior model for the target scenario is used to generate simulated behavior data 2 of traffic participant 1 for the target scenario.

[0066] The first simulated behavior model for the target scenario can be a machine learning-based model. The second simulated behavior model for the target scenario can be a classic driving behavior model, such as the IDM model or the Gipps model. The first simulated behavior model for the target scenario consumes more resources than the second simulated behavior model for the target scenario.

[0067] Simulated behavior data 1 of traffic participant 1 in the target scenario represents: simulated behavior 1 of traffic participant 1 in the target scenario. Simulated behavior data 2 of traffic participant 1 in the target scenario represents: simulated behavior 2 of traffic participant 1 in the target scenario.

[0068] The degree of anthropomorphism of traffic participant 1's simulated behavior 1 in the target scenario is higher than the degree of anthropomorphism of traffic participant 1's simulated behavior 2 in the target scenario.

[0069] If the attention level of traffic participant 2 is greater than the attention threshold, simulated behavior data 3 of traffic participant 2 for the target scenario is generated using the first simulated behavior model for the target scenario.

[0070] If the attention level of traffic participant 2 is not greater than the attention level threshold, the simulated behavior data 4 of traffic participant 2 for the target scenario is generated using the second simulated behavior model for the target scenario.

[0071] Simulated behavior data 3 of traffic participant 2 in the target scenario represents: simulated behavior 3 of traffic participant 2 in the target scenario. Simulated behavior data 4 of traffic participant 2 in the target scenario represents: simulated behavior 4 of traffic participant 2 in the target scenario.

[0072] The degree of anthropomorphism in the simulated behavior 3 of traffic participant 2 in the target scenario is higher than that in the simulated behavior 4 of traffic participant 2 in the target scenario.

[0073] Referring to Figure 2, a flowchart illustrating the data generation method provided in an embodiment of this application is shown. This method is executed during the construction of a human-like traffic flow of a target scenario. The human-like traffic flow of the target scenario is used to test the performance of algorithms running on vehicles, such as autonomous driving algorithms running on vehicles, in the target scenario.

[0074] In step S201, scene data of the target scene is obtained.

[0075] The target scenario includes the vehicle itself and multiple traffic participants surrounding it. The location of the vehicle and the multiple traffic participants in the target scenario is indicated by the scenario data of the target scenario.

[0076] A self-driving vehicle can be understood as a vehicle that runs algorithms that need to be tested for performance in a target scenario.

[0077] For each of the multiple traffic participants surrounding the vehicle, that traffic participant's traffic participant type is one of several traffic participant types. As an example, the multiple traffic participant types include: people, vehicles, and other traffic participant types.

[0078] The target scene type can be any one of multiple scene types.

[0079] In one possible implementation, multiple scenario types include, but are not limited to: dangerous intrusion scenario type, intersection scenario type, ramp scenario type, obstacle avoidance scenario type, etc.

[0080] It should be noted that the scene data of the target scene is collected by the vehicle used to collect the target scene data when it is in the target scene.

[0081] In one possible implementation, the vehicle used to collect scene data is either a self-driving car or another vehicle.

[0082] The target scenario data includes, but is not limited to: data of the vehicle in the target scenario and data of traffic participants around the vehicle in the target scenario.

[0083] The data for the target scenario for the vehicle includes, but is not limited to: the vehicle's location, the vehicle's speed, etc.

[0084] Data on traffic participants around the vehicle for the target scenario includes, but is not limited to: the location of traffic participants around the vehicle, the speed of traffic participants around the vehicle, and the type of traffic participant around the vehicle.

[0085] It should be noted that the location of the vehicle in the target scene, the location of the vehicle in the target scene, all traffic participants in the target scene, and the location of all traffic participants in the target scene are indicated by the data of the target scene.

[0086] In one possible implementation, for a traffic participant in the target scenario, if the distance between the location of the traffic participant and the location of the vehicle is less than a distance threshold, then the traffic participant can be identified as a traffic participant around the vehicle.

[0087] In another possible implementation, for a traffic participant in the target scenario, if the traffic participant's position is within an area of ​​a preset shape centered on the vehicle's position, then the traffic participant can be identified as a traffic participant surrounding the vehicle. As an example, the preset shape is a square or rectangle, and the preset shape has preset side lengths in each direction.

[0088] In step S202, based on the scene data of the target scene, the attention level of each traffic participant among the multiple traffic participants around the vehicle is determined, and the target traffic participant among the multiple traffic participants around the vehicle is determined.

[0089] In one possible implementation, to determine the attention level of each traffic participant among multiple traffic participants surrounding the vehicle, the vehicle's position and the positions of the multiple traffic participants surrounding the vehicle are obtained from the scene data of the target scene. For each traffic participant among the multiple traffic participants surrounding the vehicle, the positional importance of that traffic participant is determined based on the vehicle's position and the traffic participant's position, and this positional importance is determined as the traffic participant's attention level.

[0090] As an example, determine the distance between the vehicle and each of the multiple traffic participants surrounding it. For each traffic participant, identify a preset distance interval from the vehicle where the distance between the participant and the vehicle falls within a set of preset distance intervals. The importance of this preset distance interval is then determined as the location importance of the traffic participant. Here, the right endpoint of the i-th preset distance interval is the left endpoint of the (i+1)-th preset distance interval, and the i-th distance interval is any distance interval excluding the one with the largest left endpoint value. The importance of each preset distance interval can be predetermined. A smaller left endpoint value indicates a greater importance for the preset distance interval, and a larger left endpoint value indicates a less important preset distance interval.

[0091] As another example, for each of a plurality of preset locations, the positional importance of the traffic participant closest to the vehicle at that preset location among the multiple traffic participants surrounding the vehicle is the positional importance of that preset location. The positional importance of other traffic participants surrounding the vehicle besides those with the preset location importance is the other positional importance. Wherein, the positional importance of any one preset location is greater than the other positional importance.

[0092] As an example, several preset directions include: directly in front, left front, right front, left side, right side, directly behind, left rear, and right rear.

[0093] Referring to Figure 3, it shows a schematic diagram of an example of a preset orientation used to calculate the locational importance of traffic participants.

[0094] In this example, there are nine preset directions, represented by 1, 2, 3, 4, 5, 6, 7, and 8. 1, 4, and 7 represent: directly in front, left side, and right side, respectively. 3, 6, 5, 2, and 8 represent: left front, right front and left rear, directly behind, and right rear, respectively.

[0095] As an example, for each preset orientation (directly in front, to the left, and to the right), the positional importance of the traffic participant closest to the vehicle in that preset orientation among multiple traffic participants around the vehicle is 1. For each preset orientation (left front, right front left rear, directly behind, and right rear), the positional importance of the traffic participant closest to the vehicle in that preset orientation among multiple traffic participants around the vehicle is 0.5. The positional importance of other traffic participants around the vehicle besides those with the preset orientation importance is 0.1.

[0096] In another possible implementation, to determine the attention level of each traffic participant among multiple traffic participants surrounding the vehicle, the traffic participant type of each traffic participant can be obtained from the scene data of the target scenario. For each traffic participant among the multiple traffic participants surrounding the vehicle, the type importance of that traffic participant type is determined, and this type importance is used as the attention level of that traffic participant. The type importance of each traffic participant type can be preset. As an example, the type importance of people and cyclists is 1, the type importance of small vehicles is 0.8, and the type importance of large vehicles is 0.5.

[0097] In one possible implementation, to identify the target traffic participant among multiple traffic participants surrounding the vehicle, the traffic participants can be ranked according to their level of attention, from highest to lowest. After ranking, each traffic participant around the vehicle is assigned a position; the higher the attention a traffic participant receives, the higher its position appears. Following this ranking, a predetermined number of traffic participants can be identified as the target traffic participant.

[0098] In another possible implementation, each of the multiple traffic participants around the vehicle can be identified as the target traffic participant.

[0099] In another possible implementation, to identify a target traffic participant among multiple traffic participants surrounding the vehicle, traffic participants meeting preset conditions can be identified as the target traffic participant. The preset condition is that the traffic participant's traffic participant type is the type used to determine whether it is a target traffic participant. The traffic participant type used to determine whether it is a target traffic participant is preset.

[0100] In step S203, for each target traffic participant among multiple traffic participants around the vehicle, when the attention level of the target traffic participant is greater than the attention threshold, the first simulated behavior data of the target traffic participant for the target scenario is generated using the first simulated behavior model for the target scenario; when the attention level of the target traffic participant is not greater than the attention threshold, the second simulated behavior data of the target traffic participant for the target scenario is generated using the second simulated behavior model for the target scenario.

[0101] As an example, the attention threshold is 0.8.

[0102] For a target traffic participant around the vehicle, the degree of anthropomorphism of the simulated behavior represented by the first simulated behavior data of the target traffic participant in the target scenario is higher than the degree of anthropomorphism of the simulated behavior represented by the second simulated behavior data of the target traffic participant in the target scenario.

[0103] It should be noted that the degree of anthropomorphism in the simulated behavior data generated by the simulated behavior model is directly proportional to the amount of resources consumed by the model. The higher the degree of anthropomorphism in the simulated behavior data generated by the model, the more resources the model consumes. Conversely, the lower the degree of anthropomorphism in the simulated behavior data generated by the model, the less resources it consumes.

[0104] For a target traffic participant around the vehicle, the simulated behavior represented by the first simulated behavior data of the target traffic participant in the target scenario specifically refers to the simulated behavior of the target traffic participant in the target scenario, represented by the first simulated behavior data of the target traffic participant in the target scenario.

[0105] For a target traffic participant around the vehicle, the simulated behavior represented by the second simulated behavior data of the target traffic participant in the target scenario specifically refers to the simulated behavior of the target traffic participant in the target scenario, represented by the second simulated behavior data of the target traffic participant in the target scenario.

[0106] It should be noted that the degree of anthropomorphism of the simulated behavior of the target traffic participant in the target scenario can be understood as the extent to which the simulated behavior of the target traffic participant in the target scenario approximates one of the following behaviors: the behavior performed by the traffic participant in the target scenario, or the behavior performed by the traffic participant corresponding to the operation performed by the person operating the traffic participant. Here, the behavior corresponding to the operation performed by the person operating the traffic participant is a result of the operation performed by the person operating the traffic participant in the target scenario.

[0107] As an example, the degree of anthropomorphism in the simulated behavior of a person in the target scenario can be understood as: the degree to which the simulated behavior of a person in the target scenario closely approximates the behavior actually performed by a person in the target scenario. Similarly, the degree of anthropomorphism in the simulated behavior of a vehicle in the target scenario can be understood as: the degree to which the simulated behavior of a vehicle in the target scenario closely approximates the behavior of the vehicle in response to the driving operations performed by a person driving the vehicle in the target scenario.

[0108] In one possible implementation, the first simulated behavior model for the target scenario is a machine learning-based model, for example, a model based on transfer learning, deep learning, or adversarial learning. The second simulated behavior model for the target scenario is a classic driving behavior model. The first simulated behavior model for the target scenario consumes more resources than the second simulated behavior model.

[0109] As an example, the first simulated behavior model for the target scenario is a reinforcement learning-based model for simulating the behavior of traffic participants. Traffic participants around the vehicle act as a single agent. The single agent learns and makes decisions in the environment, and its interaction with the environment follows a Markov decision process.

[0110] The interaction between a single intelligent agent and its environment is composed of multiple groups.<S,A,R,f,γ> Let S and A represent the agent's state and action spaces, respectively, and f represent the agent's state transition function, which determines the probability distribution of transitioning from state s to s' given action a. R is the reward function, which defines the instantaneous environmental reward obtained by the agent from transitioning from state s to state s' through action a. From the start time t to the end of the interaction at time T, the total environmental reward can be expressed as:

[0111] Where γ∈[0,1] is the discount factor, which is used to balance the impact of the agent's instantaneous reward and long-term reward on the total reward. The agent's learning policy can be represented as a mapping π from state to action: S→A. The goal of solving the MDP is to find the optimal policy π that maximizes the expected reward. * The expected reward is generally represented formally using the optimal state action-value function (Q-function):

[0112] Optimal strategy π * For example, optimal behavior of an intelligent agent, such as optimal driving behavior.

[0113] As an example, the second simulated behavior model for the target scenario is the cosine lane-changing model, the Intelligent Driver Model (IDM), and the John Gipps model.

[0114] The Gipps model is one of the classic driving behavior models. The cosine lane-changing model can be represented as:

[0115] Among them, a y Let d represent the lateral acceleration of the vehicle, d represent the distance between the center lines of the two lanes, L represent the longitudinal displacement generated during the lane change, and x represent the longitudinal position of the vehicle.

[0116] Referring to Figure 4, it shows a flowchart of another data generation method provided in an embodiment of this application.

[0117] In step S401, the label information of multiple scene data is clustered to determine multiple scenes and the scene to which each scene data belongs.

[0118] Clustering the label information of multiple scene data can yield multiple clustering results.

[0119] Each clustering result includes at least one scenario data point. Each clustering result corresponds to a different scenario.

[0120] For a given clustering result, each scene data belonging to that clustering result belongs to the scene corresponding to that clustering result.

[0121] For a given scenario data, the tag information for that scenario data may include: the general category tag for that scenario data, the traffic participant information tag for that scenario data, and the traffic category tag for that scenario data.

[0122] General tags for scene data include: the time of scene data collection, the location of scene data collection, the road type of the road in the scene where the vehicle used to collect the scene data was present, the road characteristics of the road in the scene where the vehicle used to collect the scene data was present, the weather conditions in the environment where the vehicle used to collect the scene data was present, and the lighting conditions in the environment where the scene data was collected. Traffic participant tags for scene data include: the type of traffic participant in the scene where the vehicle used to collect the scene data was present, the movement state of the traffic participant in the scene where the vehicle used to collect the scene data was present, the behavioral intention of the traffic participant in the scene where the vehicle used to collect the scene data was present, and the initial direction of the traffic participant in the scene where the vehicle used to collect the scene data was present. Traffic-related tags for scene data include: the level of traffic congestion at the time of scene data collection, the traffic light status in the scene where the vehicle used to collect the scene data was present, and the signs and markings in the scene where the vehicle used to collect the scene data was present.

[0123] Table 1 is an example of a general category label for scene data, Table 2 is an example of a traffic participant information label for scene data, and Table 3 is an example of a traffic category label table for scene data.

[0124] Table 1

[0125] Table 2

[0126] Table 3

[0127]

[0128] In one possible implementation, a nonlinear clustering algorithm can be used to cluster the label information of multiple scene data to determine multiple scenes and the scene to which each scene data belongs.

[0129] As an example, considering the large number of parameters and the nonlinear and non-explicit coupling relationships among them, a Gaussian mixture model with good fitting ability is used to cluster the label information of multiple scene data. The Gaussian mixture model can be represented as:

[0130] Where, f(x) GMM ;Θ) represents a Gaussian mixture model, Θ=[K,ω k ,θ k ], K represents the number of Gaussian kernels in the Gaussian mixture model, i.e., how many classes the scene data's label information can be divided into; ω kThis represents the weight of the k-th Gaussian component in the overall model, where the sum of all weights is 1. θ k =[μ k ,Σ k ] represents the mean matrix and variance matrix corresponding to the k-th component. The input to the Gaussian mixture model is the label information of multiple scene data.

[0131] In step S402, based on the scene data of the target scene, the attention level of each traffic participant among the multiple traffic participants around the vehicle is determined, and the target traffic participant among the multiple traffic participants around the vehicle is determined.

[0132] Step S402 refers to step S201.

[0133] In step S403, for each of the multiple traffic participants around the vehicle, based on the scene data of the target scene, the location importance of the traffic participant's location and the type importance of the traffic participant's type are determined. Based on the location importance of the traffic participant's location and the type importance of the traffic participant's type, the importance parameter of the traffic participant is determined. Based on the importance parameter of the traffic participant, the attention level of the traffic participant is determined.

[0134] For a traffic participant around the vehicle, the method for determining the positional importance of the traffic participant's location and the type importance of the traffic participant's type is referred to step S202.

[0135] In one possible implementation, for a traffic participant around the vehicle, determining the importance parameter of the traffic participant based on the location importance of the traffic participant's position and the type importance of the traffic participant's type includes: using the product of the location importance of the traffic participant's position and the type importance of the traffic participant's type as the importance parameter of the traffic participant.

[0136] In one possible implementation, for one of the multiple traffic participants around the vehicle, the importance parameter of that traffic participant can be used to determine the level of attention paid to that traffic participant.

[0137] In this embodiment, when determining the importance parameters of traffic participants, the influence of the location importance of the traffic participant and the type importance of the traffic participant's traffic participant type on the overall importance of the traffic participant can be considered simultaneously. This improves the accuracy of the importance parameters of traffic participants and enhances the accuracy of traffic participant attention.

[0138] In one possible implementation, step S403 includes: step S4031.

[0139] In step S4031, for one of the multiple traffic participants around the vehicle, the right-of-way parameter of the traffic participant is determined based on whether the traffic participant is in the target lane of the vehicle; the attention level of the traffic participant is determined based on the importance parameter and the right-of-way parameter of the traffic participant.

[0140] Among them, the right-of-way parameter of the traffic participant when the vehicle is in the target lane is greater than the right-of-way parameter of the traffic participant when the vehicle is not in the target lane.

[0141] In one possible implementation, for a traffic participant around the vehicle, if the traffic participant is in the target lane where the vehicle is, the traffic participant's right-of-way parameter is 1. If the traffic participant is not in the target lane where the vehicle is, the traffic participant's right-of-way parameter is 0.

[0142] Among them, the target lane in which the vehicle is located and whether the traffic participant is in the target lane in which the vehicle is located are indicated by the scene data of the target scene.

[0143] In one possible implementation, in step S4031, for one of the multiple traffic participants around the vehicle, the sum of the importance parameter of the traffic participant and the right-of-way parameter of the traffic participant is determined as the attention level of the traffic participant.

[0144] In this embodiment of the application, when determining the attention level of traffic participants, the importance parameter of traffic participants that is highly correlated with the strength of their interaction with the vehicle, and the right-of-way parameter of traffic participants that is highly correlated with the strength of their interaction with the vehicle can be considered simultaneously. This allows for a more comprehensive assessment of the strength of the interaction between traffic participants and the vehicle, thereby improving the accuracy of the attention level of traffic participants.

[0145] In one possible implementation, step S4031 includes: step S4031a-step S4031b.

[0146] In step S4031a, for a traffic participant around the vehicle, when the traffic participant is a vehicle, the risk parameters of the traffic participant are determined according to the associated parameter information used to determine the risk parameters of the traffic participant. The associated parameter information used to determine the risk parameters of the traffic participant includes: the relative distance between the vehicle and the traffic participant, the vehicle speed, and the traffic participant speed.

[0147] In step S4031b, the attention level of the traffic participant is determined based on the importance parameter, the right-of-way parameter, and the risk parameter of the traffic participant.

[0148] In one possible implementation, in step S4031a, for a traffic participant surrounding the vehicle, spring damping model parameters for that traffic participant can be obtained. These parameters include: virtual longitudinal spring stiffness and virtual longitudinal damping coefficient for the traffic participant, and virtual lateral spring stiffness and virtual lateral damping coefficient for the traffic participant. Based on the spring damping model parameters for the traffic participant and the associated parameter information used to determine the risk parameters of the traffic participant, the longitudinal risk value of the traffic participant relative to the vehicle and the lateral risk value of the traffic participant relative to the vehicle are calculated. Based on the longitudinal risk value and the lateral risk value of the traffic participant relative to the vehicle, the risk parameters of the traffic participant are determined.

[0149] To obtain the spring-damping model parameters for a traffic participant around a vehicle, a physical model of virtual springs and damping can be used to simulate the forces between vehicles, resulting in a spring-damping model for that traffic participant. This model includes virtual longitudinal spring stiffness and virtual longitudinal damping coefficient, as well as virtual lateral spring stiffness and virtual lateral damping coefficient for that traffic participant.

[0150] For a traffic participant surrounding the vehicle, the risk parameters of the traffic participant are determined based on the spring-damping model parameters for that traffic participant and the associated parameter information used to determine the risk parameters of that traffic participant. This includes: calculating the longitudinal risk value and the lateral risk value of the traffic participant relative to the vehicle based on the spring-damping model parameters for that traffic participant and the associated parameter information used to determine the risk parameters of that traffic participant; and determining the risk parameters of the traffic participant based on the longitudinal risk value and the lateral risk value of the traffic participant relative to the vehicle.

[0151] The longitudinal risk field function used to calculate the longitudinal risk value of the traffic participant relative to the vehicle can be expressed as:

[0152] U x k represents the longitudinal risk value of the traffic participant relative to the vehicle. lon The virtual longitudinal spring stiffness is represented by x, the relative distance between the vehicle and other road users is represented by x0, and the expected relative distance between the vehicle and other road users is represented by c. lon V represents the virtual longitudinal damping coefficient.p v represents the speed of a traffic participant. x This indicates the speed of the vehicle.

[0153] The lateral risk field function used to calculate the lateral risk value of the traffic participant relative to the vehicle can be expressed as:

[0154] U y c represents the lateral risk value of the traffic participant relative to the vehicle. lat V represents the virtual lateral damping coefficient. y Let y represent the lateral velocity of the vehicle, y represent the lateral distance the vehicle deviates from the centerline of the current lane, and k represent the lateral velocity of the vehicle. lat This represents the virtual lateral spring stiffness.

[0155] The risk parameter of this traffic participant can be expressed as:

[0156] Where U represents the risk parameter of the traffic participant, U x,max U represents the maximum set vertical risk. y,max This indicates the maximum set horizontal risk.

[0157] Risk parameters of traffic participants surrounding a vehicle reflect the potential danger posed by their behavior to the vehicle. Considering these risk parameters when determining the attention level of traffic participants around the vehicle allows for a more comprehensive assessment of the strength of interaction between the vehicle and these participants, thus improving the accuracy of traffic participant attention assessment.

[0158] In this embodiment of the application, when determining the attention level of traffic participants, the importance parameter of traffic participants that is highly correlated with the strength of their interaction with the vehicle, the right-of-way parameter of traffic participants that is highly correlated with the strength of their interaction with the vehicle, and the risk parameter of traffic participants that is highly correlated with the strength of their interaction with the vehicle can be considered simultaneously. This allows for a more comprehensive assessment of the strength of the interaction between traffic participants and the vehicle, thereby improving the accuracy of traffic participant attention.

[0159] In one possible implementation, in step S4031b, for one of the multiple traffic participants around the vehicle, the sum of the traffic participant's importance parameter, the traffic participant's right-of-way parameter, and the traffic participant's risk parameter is determined as the traffic participant's attention level.

[0160] In another possible implementation, in step S4031b, for a traffic participant around the vehicle, the weighted sum of the traffic participant's importance parameter, the traffic participant's right-of-way parameter, and the traffic participant's risk parameter is determined as the traffic participant's attention level.

[0161] The weighted sum of the importance parameter, right-of-way parameter, and risk parameter of the traffic participant, which is determined as the attention level of the traffic participant, can be expressed as: D=ω1I+ω2U+ω3R.

[0162] Where D represents the traffic participant's attention level, I represents the traffic participant's importance parameter, U represents the traffic participant's risk parameter, R represents the traffic participant's right-of-way parameter, and ω1, ω2, and ω3 represent the weights of the traffic participant's importance parameter, risk parameter, and right-of-way parameter, respectively. ω1 + ω2 + ω3 = 1.

[0163] It should be noted that the importance parameter of each traffic participant has the same weight, the right-of-way parameter of each traffic participant has the same weight, and the risk parameter of each traffic participant has the same weight.

[0164] In this embodiment, the weights of the importance parameter, right-of-way parameter, and risk parameter of traffic participants can be flexibly set according to the characteristics of traffic participant attention in a given scenario or testing requirements. The traffic participant attention determined using these weights aligns with the characteristics of traffic participant attention in a given scenario or testing requirements, improving both the flexibility and accuracy of determining traffic participant attention.

[0165] As an example, when ω1 is greater than 0.5, the traffic participants being monitored tend to be those with higher importance; when ω2 is greater than 0.5, the traffic participants being monitored tend to be those with higher risk; and when ω3 is greater than 0.5, the traffic participants being monitored tend to be those with higher right-of-way. In following other vehicles scenarios, to evaluate functions such as automatic emergency braking, more attention is paid to the magnitude of risk, so ω2 can be set to 1. In test scenarios involving adjacent vehicles cutting in and self-changing lanes, it is necessary to balance the risk, importance, and right-of-way of traffic participants, so ω1 = ω2 = ω3 = 1 / 3 can be set.

[0166] In step S404, for each target traffic participant among multiple traffic participants around the vehicle, when the attention level of the target traffic participant is greater than the attention threshold, the first simulated behavior data of the target traffic participant for the target scenario is generated using the first simulated behavior model for the target scenario; when the attention level of the target traffic participant is not greater than the attention threshold, the second simulated behavior data of the target traffic participant for the target scenario is generated using the second simulated behavior model for the target scenario.

[0167] Step S404 is similar to step S203, and the process of step S404 is the same as that of step S203.

[0168] This application also provides a data generation apparatus for implementing the above-described method embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "unit" can refer to a combination of software and / or hardware that performs a predetermined function. Although the apparatus described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated. The apparatus in this application is presented in the form of a functional unit, where a functional unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that execute one or more software or fixed programs, and / or other devices that can provide the above-described functions.

[0169] The data generation device includes:

[0170] The acquisition unit is used to acquire scene data of the target scene, wherein the target scene contains the vehicle and multiple traffic participants around the vehicle;

[0171] The determining unit is configured to determine the attention level of each of the plurality of traffic participants and the target traffic participant among the plurality of traffic participants based on the scenario data.

[0172] The generation unit is configured to, for each target traffic participant, generate first simulated behavior data of the target traffic participant in the target scenario using a first simulated behavior model for the target scenario when the target traffic participant's attention level is greater than the attention level threshold; and generate second simulated behavior data of the target traffic participant in the target scenario using a second simulated behavior model for the target scenario when the target traffic participant's attention level is not greater than the attention level threshold. The first simulated behavior data represents a higher degree of anthropomorphism in the simulated behavior in the target scenario than the second simulated behavior data represents.

[0173] Furthermore, the determining unit is further configured to, for each traffic participant, determine the location importance of the traffic participant's location and the type importance of the traffic participant's traffic participant type based on the scene data, and determine the importance parameter of the traffic participant based on the location importance and the type importance, and determine the attention level of the traffic participant based on the importance parameter of the traffic participant.

[0174] Furthermore, the determining unit is further configured to determine the right-of-way parameters of the traffic participant based on whether the traffic participant is in the target lane where the vehicle is located, wherein the target lane where the vehicle is located is indicated by the scene data; and to determine the attention level of the traffic participant based on the importance parameters and right-of-way parameters of the traffic participant.

[0175] Furthermore, the determining unit is further configured to, when the traffic participant is a vehicle, determine the risk parameters of the traffic participant based on the associated parameter information used to determine the risk parameters of the traffic participant, wherein the associated parameter information includes: the relative distance between the vehicle and the traffic participant, the vehicle speed, and the traffic participant speed, wherein the associated parameter information is indicated by the scene data; and determine the attention level of the traffic participant based on the importance parameter, the right-of-way parameter, and the risk parameter of the traffic participant.

[0176] Furthermore, the determining unit is further configured to acquire spring-damping model parameters for the traffic participant, the spring-damping model parameters including: virtual longitudinal spring stiffness and virtual longitudinal damping coefficient for the traffic participant, and virtual lateral spring stiffness and virtual lateral damping coefficient for the traffic participant; based on the spring-damping model parameters for the traffic participant and the associated parameter information used to determine the risk parameters of the traffic participant, calculate the longitudinal risk value of the traffic participant relative to the vehicle and the lateral risk value of the traffic participant relative to the vehicle; and determine the risk parameters of the traffic participant based on the longitudinal risk value of the traffic participant relative to the vehicle and the lateral risk value of the traffic participant relative to the vehicle.

[0177] Furthermore, the determining unit is further configured to determine the traffic participant's attention level by weighting the importance parameter, right-of-way parameter, and risk parameter of the traffic participant.

[0178] Furthermore, the determining unit is further used to sort the attention of the multiple traffic participants from highest to lowest according to their attention level; and to determine the top preset number of traffic participants as target traffic participants.

[0179] Furthermore, the target scenario can be any one of multiple scenarios; the data generation device also includes:

[0180] The clustering unit is used to cluster the label information of multiple scene data before acquiring the scene data of the target scene, so as to determine the multiple scenes and the scene to which each scene data belongs.

[0181] Referring to Figure 5, which is a schematic diagram of the hardware structure of a computer device according to an embodiment of this application, the computer device includes one or more processors 10, a memory 20, and interfaces for connecting the various components, including high-speed interfaces and low-speed interfaces. The various components communicate with each other using different buses and can be installed on a common motherboard or otherwise as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on an external input / output device (such as a display device coupled to the interface). In some alternative embodiments, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple devices can be connected, each providing some necessary operations (e.g., as a server array, a set of blade servers, or a multiprocessor system). The processor 10 can be a central processing unit, a network processor, or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip can be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device can be a complex programmable logic device (CLP), a field-programmable gate array (FPGA), a general-purpose array logic (GPRS), or any combination thereof. The memory 20 stores instructions executable by at least one processor 10 to perform the methods shown in the above embodiments. The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on vehicle usage, etc. Furthermore, the memory 20 may include high-speed random access memory and non-transient memory, such as at least one disk storage device, flash memory device, or other non-transient solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, which can be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof. The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk, or solid-state drive; the memory 20 may also include combinations of the above types of memory. The computer device also includes an input device 30 and an output device 40. The processor 10, memory 20, input device 30, and output device 40 may be connected via a bus or other means. The input device 30 can receive input digital or character information, and generate key signal inputs related to user settings and function control of the computer device, such as touch screen, keypad, mouse, trackpad, touchpad, joystick, one or more mouse buttons, trackball, joystick, etc.The output device 40 may include a display device, an auxiliary lighting device (e.g., an LED), and a haptic feedback device (e.g., a vibration motor). The display device includes, but is not limited to, liquid crystal displays, light-emitting diodes, displays, and plasma displays. In some alternative embodiments, the display device may be a touchscreen.

[0182] This application also provides a computer-readable storage medium. The methods described in this application can be implemented in hardware or firmware, or implemented as recordable on a storage medium, or implemented as computer code downloaded over a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and subsequently stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the methods shown in the above embodiments are implemented.

[0183] A portion of the embodiments of this application can be applied as a computer program product, such as computer program instructions. When executed by a computer, these instructions, through the operation of the computer, can invoke or provide the methods and / or technical solutions according to this application. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, and installation package files. Accordingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions; the computer compiling the instructions and then executing the corresponding compiled program; the computer reading and executing the instructions; or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0184] The above embodiments are merely preferred embodiments provided to fully illustrate this application, and the scope of protection of this application is not limited thereto. Equivalent substitutions or modifications made by those skilled in the art based on this application are all within the scope of protection of this application.

Claims

1. A data generation method, characterized in that: The method includes: Acquire scene data for the target scenario, which includes the vehicle and multiple traffic participants around it. Based on the scenario data, determine the attention level of each of the multiple traffic participants, and determine the target traffic participant among the multiple traffic participants; For each target traffic participant, when the target traffic participant's attention level is greater than the attention level threshold, a first simulated behavior data for the target traffic participant in the target scenario is generated using a first simulated behavior model for the target scenario; when the target traffic participant's attention level is not greater than the attention level threshold, a second simulated behavior data for the target traffic participant in the target scenario is generated using a second simulated behavior model for the target scenario, wherein the degree of anthropomorphism of the simulated behavior in the target scenario represented by the first simulated behavior data is higher than the degree of anthropomorphism of the simulated behavior in the target scenario represented by the second simulated behavior data.

2. The method according to claim 1, characterized in that: Based on the scenario data, determining the attention level of each traffic participant among the multiple traffic participants includes: For each traffic participant, based on the scene data, the location importance of the traffic participant's location and the type importance of the traffic participant's type are determined. Based on the location importance and the type importance, the importance parameter of the traffic participant is determined. Based on the importance parameter of the traffic participant, the attention level of the traffic participant is determined.

3. The method according to claim 2, characterized in that: Determining the attention level of traffic participants based on their importance parameters includes: The right-of-way parameters of the traffic participants are determined based on whether the traffic participants are in the target lane of the vehicle, wherein the target lane of the vehicle is indicated by the scene data. The attention level of the traffic participants is determined based on their importance parameters and right-of-way parameters.

4. The method according to claim 3, characterized in that: Determining the attention level of traffic participants based on their importance parameters and right-of-way parameters includes: When the traffic participant is a vehicle, the risk parameters of the traffic participant are determined based on the associated parameter information used to determine the risk parameters of the traffic participant. The associated parameter information includes: the relative distance between the vehicle and the traffic participant, the vehicle speed, and the traffic participant speed, wherein the associated parameter information is indicated by the scene data. The level of attention of the traffic participants is determined based on their importance parameters, right-of-way parameters, and risk parameters.

5. The method according to claim 4, characterized in that: The risk parameters of the traffic participants are determined based on the associated parameter information used to determine the risk parameters of the traffic participants, including: Obtain the spring damping model parameters for the traffic participant, the spring damping model parameters including: virtual longitudinal spring stiffness and virtual longitudinal damping coefficient for the traffic participant, and virtual lateral spring stiffness and virtual lateral damping coefficient for the traffic participant; Based on the spring damping model parameters for the traffic participant and the associated parameter information used to determine the risk parameters of the traffic participant, the longitudinal risk value of the traffic participant relative to the vehicle and the lateral risk value of the traffic participant relative to the vehicle are calculated. The risk parameters of the traffic participant are determined based on the longitudinal risk value of the traffic participant relative to the vehicle and the lateral risk value of the traffic participant relative to the vehicle.

6. The method according to claim 4, characterized in that: The level of attention of a traffic participant is determined based on its importance parameter, right-of-way parameter, and risk parameter, including: The weighted sum of the importance parameter, right-of-way parameter, and risk parameter of the traffic participant is determined as the attention level of the traffic participant.

7. The method according to claim 1, characterized in that: The target traffic participant among the plurality of traffic participants includes: The attention levels of the multiple traffic participants are ranked from highest to lowest according to their level of concern. A predetermined number of traffic participants from among the multiple traffic participants are identified as target traffic participants.

8. The method according to claim 1, characterized in that: The target scenario is any one of multiple scenarios; Before acquiring scene data for the target scene, the method further includes: Clustering is performed on the label information of multiple scene data to determine the multiple scenes and the scene to which each scene data belongs.

9. A data generation device, characterized in that: The device includes: The acquisition unit is used to acquire scene data of the target scene, wherein the target scene contains the vehicle and multiple traffic participants around the vehicle; The determining unit is configured to determine the attention level of each of the plurality of traffic participants and the target traffic participant among the plurality of traffic participants based on the scenario data. The generation unit is configured to, for each target traffic participant, generate first simulated behavior data of the target traffic participant in the target scenario using a first simulated behavior model for the target scenario when the target traffic participant's attention level is greater than the attention level threshold; and generate second simulated behavior data of the target traffic participant in the target scenario using a second simulated behavior model for the target scenario when the target traffic participant's attention level is not greater than the attention level threshold. The first simulated behavior data represents a higher degree of anthropomorphism in the simulated behavior in the target scenario than the second simulated behavior data represents.

10. A computer device installed in a vehicle, characterized in that, include: A memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, the processor executing the computer instructions to perform the method of any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing a computer to perform the method of any one of claims 1 to 8.

12. A computer program product, characterized in that, Includes computer instructions for causing a computer to perform the method of any one of claims 1 to 8.

Images