A method, device, electronic device, vehicle and product for predicting meeting trajectory

Abstract

The present disclosure provides a method, device, electronic device, vehicle, and product for predicting a meeting trajectory. The method includes: in response to detecting an oncoming vehicle, traversing behavior samples in a pre-constructed driving paradigm behavior space, wherein the behavior samples include the vehicle's longitudinal acceleration and lateral offset parameters; each time a behavior sample is traversed, trajectory deduction is performed for the vehicle and the oncoming vehicle, respectively, based on the behavior sample and the current motion parameters of the vehicle and the oncoming vehicle, to obtain a trajectory pair for the vehicle and the oncoming vehicle; quantitative evaluation is performed on the plurality of trajectory pairs obtained after the traversal to determine a target trajectory pair that is suitable for the vehicle; and the vehicle is controlled to travel according to the deduced trajectory of the vehicle in the target trajectory pair. Accordingly, the present disclosure achieves trajectory prediction during meeting by traversing the driving paradigm behavior space and performing trajectory deduction, thereby improving the safety and traffic efficiency of autonomous vehicles during meeting.

Description

Technical Field

[0001] One or more embodiments of the present disclosure relate to the technical field of trajectory prediction, and in particular, to a method, device, electronic device, and storage medium for predicting a meeting trajectory. Background Art

[0002] With the rapid development of autonomous driving technology, vehicle trajectory prediction and decision-making in complex traffic scenarios have become a research hotspot. Especially in oncoming traffic scenarios, where the interactive game relationships between vehicles are complex, accurately predicting vehicle trajectories and making reasonable decisions are crucial to ensuring driving safety.

[0003] Currently, trajectory prediction in meeting scenarios requires independently predicting the future trajectory of one vehicle and then planning a path for the other vehicle based on this prediction. For example, the trajectory of the oncoming vehicle is first independently predicted, and then a feasible path is generated for the ego vehicle based on the space occupied by its trajectory. This approach separates the decision-making processes of the ego vehicle and the oncoming vehicle, ignoring the dynamic game relationship in which the behaviors of both vehicles influence and co-evolve during actual meeting processes. This approach fails to capture the interactive nature of real traffic scenarios. As a result, the resulting decisions are often limited to local optima and lack global coordination. This makes them prone to decision failure when faced with unexpected or complex vehicle behavior, which in turn impacts driving safety and efficiency. Summary of the Invention

[0004] To address the problem in related art where trajectory prediction methods fail to consider the mutual influence of the behaviors of the ego vehicle and the oncoming vehicle during the meeting process, resulting in a lack of global coordination in the generated predicted trajectory, the present disclosure provides a meeting trajectory prediction method, comprising:

[0005] In response to detecting an oncoming vehicle, traversing behavior samples in a pre-constructed driving paradigm behavior space, wherein the behavior samples include longitudinal acceleration and lateral offset parameters of the vehicle;

[0006] Each time a behavior sample is traversed, the trajectory of the ego vehicle and the oncoming vehicle are deduced based on the behavior sample and the current motion parameters of the ego vehicle and the oncoming vehicle, respectively, to obtain a trajectory pair of the ego vehicle and the oncoming vehicle;

[0007] Performing quantitative evaluation on the plurality of trajectory pairs obtained after the traversal to determine a target trajectory pair suitable for the vehicle;

[0008] The ego vehicle is controlled to travel according to the ego vehicle deduced trajectory in the target trajectory.

[0009] Optionally, before traversing the behavior samples in the pre-constructed driving paradigm behavior space, the method further includes:

[0010] Acquire a vehicle real driving data set, wherein each data item in the vehicle real driving data set includes longitudinal acceleration and lateral offset parameters of the vehicle meeting;

[0011] Filter each item of data in the real driving data set of the vehicle to retain only representative item data of different driving modes, and construct a driving paradigm behavior space based on the representative item data of different driving modes.

[0012] Optionally, each time a behavior sample is traversed, trajectory deduction is performed on the ego vehicle and the oncoming vehicle based on the behavior sample and current motion parameters of the ego vehicle and the oncoming vehicle to obtain a trajectory pair of the ego vehicle and the oncoming vehicle, including:

[0013] Each time a behavior sample is traversed, the trajectory of the ego vehicle and the oncoming vehicle is deduced within a corresponding time period after the current moment based on the behavior sample and the current motion parameters of the ego vehicle and the oncoming vehicle.

[0014] The behavior samples in the driving paradigm behavior space are retraversed. When another behavior sample is traversed, the above trajectory deduction process is repeated until the deduced trajectory pair is sufficient for the vehicle and the oncoming vehicle to complete the meeting.

[0015] Optionally, the method further includes:

[0016] estimating the time required for the vehicle and the oncoming vehicle to complete the meeting, and determining that at least a preset number of deduction processes need to be performed based on the corresponding time of one deduction and the time required to complete the meeting;

[0017] If the deduced trajectory pair has completed a preset number of deduction processes, it is determined that the deduced trajectory pair is sufficient for the ego vehicle and the oncoming vehicle to complete the meeting.

[0018] Optionally, after controlling the ego vehicle to travel according to the ego vehicle deduced trajectory in the target trajectory pair, the method further includes:

[0019] If the meeting process between the ego vehicle and the oncoming vehicle has not yet been completed, predicting the ideal position of the oncoming vehicle based on the trajectory of the oncoming vehicle aligned with the target trajectory;

[0020] Detecting the deviation between the actual position of the oncoming vehicle and the ideal position, and if the deviation exceeds a preset deviation threshold, retraversing the behavior samples in the pre-constructed driving paradigm behavior space to re-determine a target trajectory pair that is suitable for the ego vehicle;

[0021] If the deviation does not exceed the preset deviation threshold, the ego vehicle continues to be controlled to travel according to the ego vehicle deduced trajectory in the target trajectory.

[0022] Optionally, controlling the ego vehicle to travel according to the ego vehicle deduced trajectory in the target trajectory pair includes: controlling the ego vehicle to travel according to the ego vehicle deduced trajectory in the target trajectory pair, and accumulating a number of decision steps;

[0023] After controlling the ego vehicle to travel according to the ego vehicle deduced trajectory in the target trajectory pair, the method further includes:

[0024] If the meeting process between the vehicle and the oncoming vehicle has not been completed, the accumulated number of decision steps is counted;

[0025] If the accumulated number of decision steps reaches a preset step threshold, the behavior samples in the pre-constructed driving paradigm behavior space are traversed again to determine a target trajectory pair that is adapted to the vehicle;

[0026] If the accumulated number of decision steps does not reach the preset step threshold, the ego vehicle continues to be controlled to travel according to the ego vehicle deduced trajectory in the target trajectory pair, and the number of decision steps is accumulated once.

[0027] Optionally, before performing quantitative evaluation on the plurality of trajectory pairs obtained after the traversal to determine a target trajectory pair suitable for the vehicle, the method further includes:

[0028] According to the preset risk assessment rules, collision risk assessment is performed on several trajectory pairs obtained after the traversal;

[0029] According to the collision risk assessment result, only a number of trajectory pairs with qualified collision risk assessment results are retained.

[0030] The present disclosure also provides a device for predicting a vehicle trajectory, the device comprising:

[0031] A traversal unit, configured to traverse behavior samples in a pre-constructed driving paradigm behavior space in response to detecting an oncoming vehicle, wherein the behavior samples include longitudinal acceleration and lateral offset parameters of the vehicle;

[0032] a deduction unit configured to, after traversing a behavior sample each time, perform trajectory deduction on the ego vehicle and the oncoming vehicle based on the behavior sample and current motion parameters of the ego vehicle and the oncoming vehicle, respectively, to obtain a trajectory pair of the ego vehicle and the oncoming vehicle;

[0033] an evaluation unit, configured to perform quantitative evaluation on the plurality of trajectory pairs obtained after the traversal to determine a target trajectory pair suitable for the ego vehicle;

[0034] A control unit is used to control the vehicle to travel according to the deduced trajectory of the vehicle in the target trajectory.

[0035] The present disclosure also provides a vehicle, comprising:

[0036] processor;

[0037] a memory for storing processor-executable instructions;

[0038] The processor implements the above method by running the executable instructions.

[0039] The present disclosure also provides a computer program product, comprising a computer program / instruction, which implements the above method when executed by a processor.

[0040] In response to detecting an oncoming vehicle, the disclosed embodiments traverse a pre-constructed driving paradigm behavior space containing vehicle longitudinal acceleration and lateral offset parameters. For each behavior sample, trajectory deduction is performed based on the motion parameters of the current ego vehicle and the oncoming vehicle, generating a series of possible trajectory pairs for the ego vehicle and the oncoming vehicle. These trajectory pairs are then quantitatively evaluated, and a target trajectory pair suitable for the ego vehicle is selected as the basis for the final decision. Finally, the ego vehicle is controlled to follow the trajectory of the selected target trajectory pair.

[0041] Through the above approach, compared with the defect of traditional step-by-step decision-making that cannot achieve global optimization, the technical solution disclosed in the present invention regards the vehicle and the oncoming vehicle as equal game subjects, can dynamically balance the behavior of all vehicles, and achieve more reasonable decision-making and planning in complex meeting scenarios, avoiding the decision-making limitations brought by unilateral predictions, significantly improving the accuracy and reliability of meeting trajectory predictions, enhancing the system's ability to cope with dynamic environmental changes, and effectively ensuring driving safety and efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0043] Figure 1 is a flowchart of a method for predicting a vehicle trajectory according to an exemplary embodiment;

[0044] Figure 2 This is a schematic diagram of vehicle trajectory deduction in a meeting scenario, shown in an exemplary embodiment;

[0045] Figure 3 is a flowchart of another method for predicting a vehicle trajectory according to an exemplary embodiment;

[0046] Figure 4 is a hardware structure diagram of an electronic device shown in an exemplary embodiment;

[0047] Figure 5 The figure is a block diagram of a device for predicting a vehicle trajectory according to an exemplary embodiment. DETAILED DESCRIPTION

[0048] In order to enable those skilled in the art to better understand the technical solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present disclosure.

[0049] It should be noted that in other embodiments, the steps of the corresponding method are not necessarily performed in the order shown and described in this disclosure. In some other embodiments, the method may include more or fewer steps than those described in this disclosure. In addition, a single step described in this disclosure may be broken down into multiple steps for description in other embodiments; and multiple steps described in this disclosure may be combined into a single step for description in other embodiments.

[0050] With the rapid development of autonomous driving technology, vehicle trajectory prediction and decision-making in complex traffic scenarios have become a research hotspot. Especially in oncoming traffic scenarios, where the interactive game relationships between vehicles are complex, accurately predicting vehicle trajectories and making reasonable decisions are crucial to ensuring driving safety.

[0051] Currently, trajectory prediction in meeting scenarios requires independently predicting the future trajectory of one vehicle and then planning a path for the other vehicle based on this prediction. For example, the trajectory of the oncoming vehicle is first independently predicted, and then a feasible path is generated for the ego vehicle based on the space occupied by its trajectory. This approach separates the decision-making processes of the ego vehicle and the oncoming vehicle, ignoring the dynamic game relationship in which the behaviors of both vehicles influence and co-evolve during actual meeting processes. This approach fails to capture the interactive nature of real traffic scenarios. As a result, the resulting decisions are often limited to local optima and lack global coordination. This makes them prone to decision failure when faced with unexpected or complex vehicle behavior, which in turn impacts driving safety and efficiency.

[0052] In view of this, the present disclosure aims to propose a technical solution for meeting trajectory prediction that fully considers the interactive game relationship between vehicles.

[0053] This technical solution first responds to the detection of an oncoming vehicle by traversing a pre-constructed behavioral space of driving paradigms, which includes vehicle longitudinal acceleration and lateral offset parameters. Furthermore, for each behavioral sample, trajectory deduction is performed based on the motion parameters of the current ego vehicle and the oncoming vehicle, generating a series of possible trajectory pairs for the ego vehicle and the oncoming vehicle. These trajectory pairs are then quantitatively evaluated, and a target trajectory pair suitable for the ego vehicle is selected as the basis for the final decision. Finally, the ego vehicle is controlled to follow the trajectory in the selected target trajectory pair.

[0054] For example, a vehicle (vehicle A) is traveling at 60 km / h on a two-lane, two-way highway. It detects an oncoming vehicle (vehicle B) traveling at 70 km / h in the opposite lane, 100 meters away. Vehicle A's system triggers the oncoming vehicle trajectory prediction method, traversing behavior samples from a pre-built driving paradigm behavior space. Assume that this behavior space contains the following behavior samples: Sample 1: Longitudinal acceleration = 0 m / s² (maintaining a constant speed), lateral offset = 0.2 m (0.2 m offset to the right of the lane); Sample 2: Longitudinal acceleration = 0.1 m / s² (acceleration), lateral offset = -0.1 m (0.1 m offset to the left of the lane); Sample 3: Longitudinal acceleration = -0.3 m / s² (deceleration), lateral offset = -0.3 m (0.3 m offset to the left of the lane). After traversing Sample 1, the trajectories of vehicles A and B are deduced based on Sample 1 and the current motion parameters of vehicles A and B. As can be seen, vehicle A then maintains a constant speed of 60 km / h while offsetting 0.2 meters to the right of the lane; vehicle B maintains a constant speed of 70 km / h with a lateral offset of 0 (remaining in its original lateral position in the lane). Based on the above trajectory deduction process, trajectory pair 1 for vehicles A and B is obtained. After traversing samples 2 and 3, a similar process is repeated to obtain trajectory pairs 2 and 3 for vehicles A and B. The system then quantitatively evaluates the three trajectory pairs obtained based on safety, comfort, and traffic efficiency. Sample 1 scored the highest and was identified as the target trajectory pair for vehicle A. Finally, the system controls vehicle A to follow the deduced trajectory from the target trajectory pair: maintaining a constant speed of 60 km / h while offsetting 0.2 meters to the right of the lane, thereby completing the passing maneuver safely, comfortably, and efficiently.

[0055] Through the above approach, compared with the defect of traditional step-by-step decision-making that cannot achieve global optimization, the technical solution disclosed in the present invention regards the vehicle and the oncoming vehicle as equal game subjects, can dynamically balance the behavior of all vehicles, and achieve more reasonable decision-making and planning in complex meeting scenarios, avoiding the decision-making limitations brought by unilateral predictions, significantly improving the accuracy and reliability of meeting trajectory predictions, enhancing the system's ability to cope with dynamic environmental changes, and effectively ensuring driving safety and efficiency.

[0056] The present disclosure is described below through specific embodiments in combination with specific application scenarios.

[0057] See Figure 1 , Figure 1 FIG. 1 is a flowchart of a method for predicting a vehicle trajectory according to an exemplary embodiment. The method may include the following steps:

[0058] Step 102: In response to detecting an oncoming vehicle, traverse behavior samples in a pre-constructed driving paradigm behavior space, wherein the behavior samples include longitudinal acceleration and lateral offset parameters of the vehicle.

[0059] For example, a vehicle (vehicle A) is traveling at 60 km / h on a two-lane, two-way highway. It detects an oncoming vehicle (vehicle B) traveling at 70 km / h in the opposite lane, 100 meters away. Vehicle A's system triggers its oncoming trajectory prediction method, traversing behavior samples from a pre-built driving paradigm behavior space. Assume that this behavior space contains the following behavior samples: Sample 1: Longitudinal acceleration = 0 m / s² (maintaining a constant speed), lateral offset = 0.2 m (0.2 m offset to the right of the lane); Sample 2: Longitudinal acceleration = 0.1 m / s² (acceleration), lateral offset = -0.1 m (0.1 m offset to the left of the lane); Sample 3: Longitudinal acceleration = -0.3 m / s² (deceleration), lateral offset = -0.3 m (0.3 m offset to the left of the lane).

[0060] During driving, the onboard sensor systems (such as cameras, radar, and lidar) detect an oncoming vehicle in the opposite lane, triggering the oncoming vehicle trajectory prediction process. Before the prediction process begins, the system has constructed a driving paradigm behavior space using a large amount of real-world driving data. Specifically, the system collects a large amount of driving data from vehicles in oncoming scenarios. This data includes the vehicle's longitudinal acceleration (indicating the degree to which the vehicle accelerates or decelerates along the lane) and lateral offset parameters (indicating the vehicle's lateral position change relative to the lane centerline).

[0061] In one embodiment shown, before traversing the behavior samples in the pre-constructed driving paradigm behavior space, the method further includes: obtaining a vehicle real driving data set, each data item in the vehicle real driving data set including longitudinal acceleration and lateral offset parameters of the vehicle meeting; filtering each data item in the vehicle real driving data set to retain only representative data items of different driving modes, and constructing a driving paradigm behavior space based on the representative data items of different driving modes.

[0062] For example, the system uses cluster analysis to filter the driving data of a large number of vehicles in oncoming scenarios (the data includes the vehicle's longitudinal acceleration and lateral offset parameters), obtains several cluster centers, and then selects the data near the cluster center points as representative data. For example, the three representative data items are longitudinal acceleration = 0 m / s² (maintaining a constant speed), lateral offset = 0.2m (0.2m offset to the right of the lane); longitudinal acceleration = 0.1 m / s² (acceleration), lateral offset = -0.1m (0.1m offset to the left of the lane); longitudinal acceleration = -0.3 m / s² (deceleration), lateral offset = -0.3m (0.3m offset to the left of the lane). Each representative data item may represent a driving style, such as aggressive, conservative, smooth, etc. Based on each representative data item, a driving paradigm behavior space is constructed.

[0063] The driving paradigm behavior space is constructed by acquiring real-world driving data sets and filtering representative data from different driving modes. This real-world driving data includes longitudinal acceleration and lateral offset parameters during vehicle passing, effectively reflecting the vehicle's behavioral characteristics under different driving styles. By selecting and retaining representative data, the constructed driving paradigm behavior space covers various driving modes while maintaining data simplicity, improving the efficiency of subsequent traversal and deduction. Cluster analysis groups a large number of similar data points into clusters. This process first defines the number of clusters to be generated. Then, a clustering algorithm is applied, attempting to minimize the distance between data points within the same cluster and maximize the distance between data points in different clusters. For each cluster, its geometric center, the cluster centroid, is calculated, representing the average value of all data points within that cluster. The system then selects data points close to these cluster centroids as representative items to ensure that the selected data accurately reflects the key behavioral characteristics of the driving mode.

[0064] It should be noted that in addition to filtering the various data items in the real vehicle driving dataset through cluster analysis, the data items in the real vehicle driving dataset can also be filtered through manually formulated rules. For example, threshold ranges for different driving modes can be set based on traffic regulations, driving habits, or human observation. Data items in the original dataset that meet the above rules are classified into corresponding driving style categories, and typical samples are selected from each category as representative data. In addition, the various data items in the real vehicle driving dataset can also be filtered through deep learning methods. For example, the original driving data (including longitudinal acceleration and lateral offset) is constructed into a time series or state vector form and input into a neural network. The neural network uses a supervised deep learning model to classify driving modes, and for each category, a typical sample is selected as representative data. The present disclosure does not limit the method for filtering representative data for different driving modes.

[0065] In related technologies, it is usually difficult to strike a balance between sampling completeness and computational efficiency when predicting the trajectory of vehicles during oncoming traffic, because fully enumerating all possible combinations of vehicle behaviors will lead to excessive consumption of computing resources. If the computing power burden is reduced and the number of samples is reduced, key scenarios may be missed.

[0066] Through the technical methods in this embodiment, the data in the vehicle's real driving data set are screened to construct a behavior space with sparse and complete characteristics that conforms to the driving paradigm. Only the necessary representative data is retained, which helps to identify and distinguish different driving styles, effectively reduces redundant data, and improves computing efficiency.

[0067] Step 104: Each time a behavior sample is traversed, the trajectory of the ego vehicle and the oncoming vehicle are deduced based on the behavior sample and the current motion parameters of the ego vehicle and the oncoming vehicle to obtain a trajectory pair of the ego vehicle and the oncoming vehicle.

[0068] For example, after traversing to sample 1, the trajectories of vehicles A and B are deduced based on sample 1 and their current motion parameters. It can be seen that vehicle A then maintains a constant speed of 60 km / h, while deviating 0.2m to the right of the lane; vehicle B maintains a constant speed of 70 km / h, and its lateral offset is 0 (remaining in its original lateral position in the lane). Based on this trajectory deduction process, trajectory pair 1 for vehicles A and B is obtained. After traversing to samples 2 and 3, the above process is repeated to obtain trajectory pairs 2 and 3 for vehicles A and B.

[0069] Among them, when an oncoming vehicle is detected, the system begins to traverse the behavior samples in the driving paradigm behavior space. Each behavior sample contains specific longitudinal acceleration and lateral offset parameters, representing a possible driving behavior. The system will deduce the movement trajectory of the vehicle and the oncoming vehicle in the future based on the current motion state (such as position, speed, etc.) of the vehicle and the oncoming vehicle, combined with the traversed behavior samples. The deduction method can be based on the target vehicle speed. Combined with the kinematic equation. In order to meet safely, the target speed of the vehicle is usually a lower speed. If the vehicle's current speed , the vehicle needs to decelerate and select the longitudinal acceleration from the driving paradigm behavior space , the deduction step length is , then The speed of the time step , No. The position of the time step When decelerating to the target passing speed When the vehicle enters the uniform speed meeting phase. In rare cases, if the vehicle's current speed is , the vehicle needs to accelerate, and the longitudinal acceleration is selected from the driving paradigm behavior space The deduction process is similar to the above. If the vehicle's current speed is , then the vehicle moves at a uniform speed until the meeting is completed.

[0070] In deriving the After the position of the vehicle in the first time step is obtained, the position can be derived based on the lateral offset parameters (including lateral offset or lateral offset angle). The lateral offset of the vehicle in the first time step can be directly obtained if the lateral offset parameter is the lateral offset. The lateral offset of the time step. If the lateral offset parameter is the lateral offset angle, then The time step is relative to the The lateral offset of the time step ,in, is the longitudinal position deviation between two adjacent time steps, , The vehicle is The vertical position of the time step, The vehicle is The vertical position of the time step, It is The lateral deviation angle at each time step is selected from the driving paradigm behavior space.

[0071] It should be noted that in addition to using the aforementioned kinematic calculation formulas, trajectory prediction can also utilize a vehicle dynamics model, combining the vehicle's current state (speed, acceleration, steering angle, etc.) with environmental constraints (road boundaries, obstacle locations), and observed behavior samples to solve the vehicle's future trajectory through numerical integration methods. Alternatively, deep learning models can be used to predict the vehicle's future trajectory. This disclosure does not limit the method of performing trajectory prediction for both the ego vehicle and oncoming vehicles based on behavior samples and their current motion parameters.

[0072] Step 106: Perform quantitative evaluation on the plurality of trajectory pairs obtained after the traversal to determine a target trajectory pair that is suitable for the vehicle.

[0073] For example, the system performs a weighted quantitative evaluation on the three sets of trajectory pairs obtained at the end of the traversal from the dimensions of safety, comfort, and traffic efficiency, using the safety evaluation function, comfort evaluation function, and efficiency evaluation function as well as the weight of each function. After the evaluation, the trajectory pair of sample 1 scored the highest and was determined to be the target trajectory pair suitable for vehicle A.

[0074] By traversing all behavior samples in the driving paradigm behavior space, the system can obtain multiple possible trajectory pairs. Each trajectory pair contains the deduced trajectory of the ego vehicle and the deduced trajectory of the oncoming vehicle. Quantitative evaluation is based on multiple indicators related to vehicle meeting, such as safety, comfort, and efficiency. A comprehensive score for each trajectory pair is calculated through weighted calculation, thereby selecting the target trajectory pair that best suits the ego vehicle. Quantitative evaluation is based on multiple indicators related to meeting, such as safety, comfort, and efficiency. Each indicator has a corresponding evaluation function and weight. The evaluation function can be a cost function or a reward function. The reward function is usually the inverse of the cost function. Cost functions may include, but are not limited to: lateral safety cost function with oncoming vehicles, lateral offset cost function, passability cost function, obstacle distance cost function, acceleration comfort cost function, and acceleration rate comfort cost function.

[0075] For example, the evaluation function for safety might be proportional to the minimum distance to oncoming vehicles; the evaluation function for comfort might be inversely proportional to the rate of change of acceleration; and the evaluation function for efficiency might be inversely proportional to the time it takes to complete a meeting. The weight of each indicator reflects its relative importance. The weighting of each indicator may vary in different driving scenarios. For example, safety may be weighted more heavily on narrow roads, while efficiency may be weighted more heavily on wide roads.

[0076] It should be noted that quantitative evaluation is not limited to evaluation functions of different dimensions. It can also be achieved by manually formulating a series of rules and thresholds based on principles such as driving safety, comfort, and efficiency, or by scoring behaviors through reinforcement learning. This disclosure does not limit the specific evaluation methods and weighting methods.

[0077] Step 108: Control the ego vehicle to travel according to the ego vehicle deduced trajectory in the target trajectory.

[0078] For example, the system controls vehicle A to drive according to the self-vehicle deduced trajectory centered on the target trajectory. That is, vehicle A maintains a constant speed of 60 km / h while offsetting 0.2 m to the right of the lane, thereby completing the meeting safely, comfortably and efficiently.

[0079] The target trajectory pair is a suitable trajectory pair determined after quantitative evaluation among multiple possible trajectory pairs. This suitable trajectory pair not only considers safety (such as collision avoidance), but also comfort (such as smooth acceleration and lateral deviation) and efficiency (such as minimizing driving time and energy consumption). The ego vehicle's predicted trajectory is the future motion path of the ego vehicle based on the current driving situation and selected behavioral samples. Specifically, it includes motion parameters such as the ego vehicle's speed, acceleration, and lateral deviation over a period of time.

[0080] In one embodiment shown, each time a behavior sample is traversed, trajectory deduction is performed on the ego vehicle and the oncoming vehicle respectively based on the behavior sample and the current motion parameters of the ego vehicle and the oncoming vehicle to obtain a trajectory pair of the ego vehicle and the oncoming vehicle, including: each time a behavior sample is traversed, trajectory deduction is performed on the ego vehicle and the oncoming vehicle respectively within a corresponding time length after the current moment based on the behavior sample and the current motion parameters of the ego vehicle and the oncoming vehicle; the behavior samples in the driving paradigm behavior space are re-traversed, and when another behavior sample is traversed, the above trajectory deduction process is repeated until the deduced trajectory pair is sufficient for the ego vehicle and the oncoming vehicle to end the meeting.

[0081] For example, vehicle A is traveling at a constant speed of 60 km / h, while vehicle B is traveling at a constant speed of 70 km / h. The two vehicles are currently 30 meters apart. The system traverses a behavior sample from the pre-built driving paradigm behavior space. This sample has a longitudinal acceleration of 0 m / s² (maintaining a constant speed) and a lateral offset of -0.3 m (0.3 m offset to the left of the lane). Based on this sample, the system begins the first trajectory extrapolation step. Within a future extrapolation step of 0.1 seconds, vehicle A maintains a speed of 60 km / h and offsets 0.3 m to the left of the lane. Simultaneously, vehicle B also maintains a speed of 70 km / h and offsets 0.3 m to the left of the lane. At the end of the 0.1 second extrapolation step, the distance between the two vehicles has been reduced to approximately 26.39 meters (taking into account relative speed and time). The system then continues to traverse the next behavior sample from the pre-built driving paradigm behavior space and repeats the trajectory extrapolation process until the distance between vehicle A and vehicle B reaches 0 m or less, at which point the two vehicles are considered to have completed their meeting.

[0082] See Figure 2 , Figure 2 FIG. 1 is a schematic diagram of a vehicle trajectory deduction in a meeting scenario, shown as an exemplary embodiment. Figure 2 The figure shows a scenario of an autonomous vehicle driving in a complex traffic environment. The figure shows two vehicles traveling through a busy area surrounded by several stationary vehicles. The two vehicles are performing a passing maneuver to ensure safe passage. The arrows surrounding the vehicles in the figure represent the direction of travel of the ego vehicle (the vehicle on the left) and the detection range of the onboard sensors. The onboard sensors can detect the distance between the two vehicles and the distance between the ego vehicle and other obstacles, which are used for consideration in trajectory planning and decision-making. The figure shows three trajectory simulations, represented by three shaded blocks in the direction of travel ahead of each vehicle. After the three trajectory simulations are completed, the ego vehicle and the oncoming vehicle have completed the passing process.

[0083] The duration of one deduction can be referred to as a deduction step. When the rear ends of the ego vehicle and the oncoming vehicle are offset after the meeting, the meeting is considered complete. Trajectory deduction uses a step-by-step approach, deducing trajectories within a time step (e.g., 0.1s) at a time. Specifically, the system first calculates the position, velocity, and acceleration of the ego vehicle and the oncoming vehicle for the next time step based on their current motion states and the longitudinal acceleration and lateral offset parameters in the traversed behavior samples. The system then retraces the behavior samples, selects the next behavior sample, and continues to deduce the trajectory for the next time step. This process continues until the deduced trajectory covers the entire meeting process. This step-by-step deduction approach more accurately simulates the dynamic changes during the meeting process and improves the accuracy of trajectory prediction.

[0084] In one embodiment shown, the method further includes: estimating the time required for the ego vehicle and the oncoming vehicle to complete the meeting, so as to determine that at least a preset number of deduction processes need to be performed based on the corresponding time of one deduction and the time required to complete the meeting; if the deduced trajectory pair has completed the preset number of deduction processes, then it is determined that the deduced trajectory pair is sufficient for the ego vehicle and the oncoming vehicle to complete the meeting.

[0085] For example, to determine the duration of a simulation, the system estimates the time required to complete the meeting based on factors such as the current speed, position, and road width of the ego vehicle and the oncoming vehicle. For example, if the relative speed of the ego vehicle and the oncoming vehicle is 20m / s and the current distance is 100m, the meeting will take approximately 5s to complete. If the time step of a simulation is 0.1s, at least 50 simulations are required to cover the entire meeting process. If the trajectory pair obtained by the simulation has completed 50 simulations, it can be considered to cover the entire meeting process.

[0086] To ensure that the simulated trajectory pairs cover the entire meeting process, this method estimates the time required for the ego vehicle and the oncoming vehicle to complete the meeting. Based on the duration of each simulation, the method calculates the minimum number of simulations required. Once the preset number of simulations are complete, the simulated trajectory pairs are considered sufficient for the ego vehicle and the oncoming vehicle to complete the meeting.

[0087] Considering that the process of meeting vehicles usually involves deceleration, the actual meeting time is longer than the normal driving time when the distance between the two vehicles is constant, and there may be other unexpected situations or calculation errors. Therefore, the system generally adds a certain amount of redundancy to the calculated number of simulations. For example, the number of simulations can be increased from 50 to 100 or more to ensure that the simulated trajectory pairs can cover the entire meeting process.

[0088] In one embodiment shown, after controlling the ego vehicle to travel according to the ego vehicle deduced trajectory in the target trajectory pair, the method further includes: if the meeting process between the ego vehicle and the oncoming vehicle has not been completed, predicting the ideal position of the oncoming vehicle according to the oncoming vehicle deduced trajectory in the target trajectory pair; detecting the deviation between the actual position and the ideal position of the oncoming vehicle, and if the deviation exceeds a preset deviation threshold, re-traversing the behavior samples in the pre-constructed driving paradigm behavior space to re-determine the target trajectory pair that is suitable for the ego vehicle; if the deviation does not exceed the preset deviation threshold, continuing to control the ego vehicle to travel according to the ego vehicle deduced trajectory in the target trajectory pair.

[0089] For example, based on a comprehensive scoring process, the system selects the highest-scoring trajectory pair as the target trajectory pair and controls vehicle A to follow the ego vehicle's predicted trajectory from the target trajectory pair. During control, the system continuously monitors the deviation between the oncoming vehicle's actual position and its predicted ideal position. If the deviation exceeds a preset threshold (e.g., 0.5m), indicating that the oncoming vehicle's actual behavior is inconsistent with the prediction, the system re-traverses the behavior sample from the pre-constructed driving paradigm behavior space and, based on this sample and the vehicle's current motion parameters, re-deduces the trajectory to adapt to the oncoming vehicle's behavior changes.

[0090] While controlling the ego vehicle to follow the predicted trajectory centered on the target trajectory, this method also continuously monitors the deviation between the oncoming vehicle's actual position and the predicted ideal position. If the deviation exceeds a preset threshold, indicating a significant difference between the oncoming vehicle's actual trajectory and the predicted trajectory, the method re-traverses the behavioral samples and performs trajectory deduction to adapt to the oncoming vehicle's behavior changes. If the deviation is within an acceptable range, the vehicle continues to follow the original trajectory.

[0091] In related technologies, setting the trajectory prediction frequency for autonomous vehicles often faces a dilemma. High-frequency decision-making increases the load on computing resources and leads to frequent vehicle behavior switching, reducing driving comfort and safety. However, blindly reducing the decision-making frequency can cause delayed environmental response and decision lags. In this embodiment, trajectory prediction is re-performed when the deviation between the predicted ideal position of an oncoming vehicle and its actual position exceeds a preset threshold, thereby maintaining an appropriate decision-making frequency and maintaining driving safety and comfort.

[0092] In one embodiment shown, controlling the ego vehicle to travel according to the ego vehicle-derived trajectory in the target trajectory pair includes: controlling the ego vehicle to travel according to the ego vehicle-derived trajectory in the target trajectory pair, and accumulating a number of decision steps once; after controlling the ego vehicle to travel according to the ego vehicle-derived trajectory in the target trajectory pair, the method further includes: if the meeting process between the ego vehicle and the oncoming vehicle has not been completed, counting the accumulated number of decision steps; if the accumulated number of decision steps reaches a preset step threshold, re-traversing the behavior samples in the pre-constructed driving paradigm behavior space to again determine the target trajectory pair that is suitable for the ego vehicle; if the accumulated number of decision steps does not reach the preset step threshold, continuing to control the ego vehicle to travel according to the ego vehicle-derived trajectory in the target trajectory pair, and accumulating a number of decision steps once.

[0093] For example, see Figure 3 , Figure 3 FIG. 1 is a flow chart of another method for predicting a vehicle trajectory according to an exemplary embodiment. Figure 3As shown in the flowchart, the ego vehicle, after detecting an oncoming vehicle, predicts and executes the optimal trajectory through a series of steps to safely and efficiently complete the meeting process. The process begins with detection of the oncoming vehicle and the initialization of the number of decision steps, j, to 0. Next, the system performs trajectory prediction, including behavior sampling, trajectory deduction, trajectory pruning, and trajectory selection. During the trajectory prediction phase, the system deduces trajectories based on behavior samples traversed in a pre-constructed driving paradigm behavior space and the current motion parameters of the ego vehicle and the oncoming vehicle, generating multiple possible trajectory pairs. Trajectory pruning is then performed on all possible trajectory pairs to remove those with collision risk. The remaining trajectory pairs are then quantitatively evaluated to determine the optimal trajectory pair suitable for the ego vehicle. The optimal trajectory pair is then executed for the ego vehicle, and the number of decision steps, j+1, is initialized. During the ego vehicle's driving, a determination is made as to whether the meeting process with the oncoming vehicle has been completed. If the ego vehicle fails to meet the target, the system checks whether the initial number of decision steps, j, has reached the preset number, N. It then collects the actual position of the oncoming vehicle and compares it with the predicted position, determined from the derived trajectory, to determine whether the position deviation exceeds a threshold. If either of these two conditions is met, the currently derived trajectory is no longer suitable for continued use, and the system re-predicts the trajectory to adapt to potentially changing traffic conditions. If neither condition is met, the system continues executing the currently derived trajectory and increases the number of decision steps again. This process continues until the ego vehicle meets the target, at which point the system returns the ego vehicle to cruise mode and continues to detect the presence of oncoming vehicles.

[0094] This embodiment introduces the concept of decision steps. Each time the system controls the vehicle to follow a simulated trajectory for a time step, it accumulates a decision step. When the cumulative number of decision steps reaches a preset threshold, the trajectory simulation is re-run to ensure the timeliness and accuracy of the control strategy. This dynamic adjustment mechanism enables the system to flexibly respond to complex traffic conditions and improve the safety and efficiency of meeting vehicles.

[0095] In one embodiment shown, the method further includes: before performing a quantitative evaluation on the several trajectory pairs obtained at the end of the traversal to determine the target trajectory pair that is adapted to the vehicle, the method further includes: performing a collision risk evaluation on the several trajectory pairs obtained at the end of the traversal according to preset risk assessment rules; and based on the collision risk assessment results, only retaining the trajectory pairs with qualified collision risk assessment results.

[0096] For example, at the end of the system traversal, three trajectory pairs are obtained: trajectory pair A, trajectory pair B, and trajectory pair C. Each trajectory pair contains different possible driving paths that the ego vehicle and the oncoming vehicle could take during the meeting process. Before performing a quantitative assessment, the system first performs a collision risk assessment on these trajectory pairs to screen out those with safety hazards. This collision risk assessment can be performed based on a minimum safe distance risk assessment rule: the system sets a minimum safe distance of 0.25m between vehicles and between vehicles and the road boundary. If, during the meeting process, the distance between the ego vehicle and the oncoming vehicle traveling along trajectory pair A, or the distance between either vehicle and the road boundary, is less than 0.25m, and the distance between the ego vehicle and the oncoming vehicle traveling along trajectory pair B or trajectory pair C, or the distance between either vehicle and the road boundary, is greater than 0.25m, then trajectory pair A is deemed unsafe, and only trajectory pairs B and C are retained for the next quantitative assessment step.

[0097] Before performing a quantitative assessment, the system first evaluates the collision risk of each trajectory pair, eliminating those with potential safety hazards to ensure that the final selected trajectory pair ensures driving safety. This collision risk assessment is based on pre-set risk assessment rules. These rules may include: minimum safe distances between vehicles, minimum safe distances between vehicles and road boundaries, and maximum permissible acceleration and deceleration. The system checks each trajectory pair for violations of these safety rules throughout the entire meeting process. If so, the trajectory pair is marked as unsafe and will not be considered further.

[0098] In addition, this embodiment also introduces a learning mechanism. The system will record the data and control effects of each meeting process, and use this data to continuously optimize the driving paradigm behavior space and evaluation method model, so that the system can continuously adapt to different driving environments and driving styles.

[0099] Through the above steps, the meeting trajectory prediction method provided in this embodiment can accurately predict the trajectories of the own vehicle and the oncoming vehicle during the meeting process, and control the own vehicle to complete the meeting safely, comfortably and efficiently based on the prediction results.

[0100] Corresponding to the above-mentioned embodiment of the method for predicting a meeting trajectory, the present disclosure further provides an embodiment of a device for predicting a meeting trajectory.

[0101] See Figure 4 , Figure 4This is a hardware structure diagram of an electronic device shown in an exemplary embodiment. At the hardware level, the device includes a processor 402, an internal bus 404, a network interface 406, a memory 408, and a non-volatile memory 410, and of course may also include other required hardware. One or more embodiments of the present disclosure can be implemented based on software, such as the processor 402 reading the corresponding computer program from the non-volatile memory 410 into the memory 408 and then running it. Of course, in addition to software implementation, one or more embodiments of the present disclosure do not exclude other implementation methods, such as logic devices or a combination of software and hardware, etc., that is, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0102] See Figure 5 , Figure 5 FIG. 5 is a block diagram of a vehicle-meeting trajectory prediction device according to an exemplary embodiment. The vehicle-meeting trajectory prediction device 500 can be applied to Figure 4 The electronic device shown in the figure is used to implement the technical solution of the present disclosure. The device includes:

[0103] A traversal unit 502 is configured to traverse behavior samples in a pre-constructed driving paradigm behavior space in response to detecting an oncoming vehicle, wherein the behavior samples include longitudinal acceleration and lateral offset parameters of the vehicle;

[0104] The deduction unit 504 is configured to, after each traversal of a behavior sample, perform trajectory deduction on the ego vehicle and the oncoming vehicle based on the behavior sample and current motion parameters of the ego vehicle and the oncoming vehicle, respectively, to obtain a trajectory pair of the ego vehicle and the oncoming vehicle;

[0105] A first evaluation unit 506 is configured to perform quantitative evaluation on the plurality of trajectory pairs obtained after the traversal to determine a target trajectory pair suitable for the vehicle;

[0106] The first control unit 508 is used to control the ego vehicle to travel according to the ego vehicle deduced trajectory in the target trajectory.

[0107] In some embodiments, before traversing the behavior samples in the pre-constructed driving paradigm behavior space, the apparatus further comprises:

[0108] An acquisition unit 510 is configured to acquire a vehicle real driving data set, wherein each data item in the vehicle real driving data set includes longitudinal acceleration and lateral offset parameters of the vehicle meeting;

[0109] The screening unit 512 is configured to screen the data in the real vehicle driving data set, retain only the representative data of different driving modes, and construct a driving paradigm behavior space based on the representative data of different driving modes.

[0110] In some embodiments, after each traversal to a behavior sample, trajectory deduction is performed on the ego vehicle and the oncoming vehicle respectively based on the behavior sample and the current motion parameters of the ego vehicle and the oncoming vehicle to obtain a trajectory pair of the ego vehicle and the oncoming vehicle, including:

[0111] Each time a behavior sample is traversed, the trajectory of the ego vehicle and the oncoming vehicle is deduced within a corresponding time period after the current moment based on the behavior sample and the current motion parameters of the ego vehicle and the oncoming vehicle.

[0112] The behavior samples in the driving paradigm behavior space are retraversed. When another behavior sample is traversed, the above trajectory deduction process is repeated until the deduced trajectory pair is sufficient for the vehicle and the oncoming vehicle to complete the meeting.

[0113] In some embodiments, the apparatus further comprises:

[0114] An estimating unit 514 is configured to estimate the time required for the vehicle and the oncoming vehicle to complete the meeting, so as to determine that at least a preset number of deduction processes need to be performed based on the corresponding time of one deduction and the time required to complete the meeting;

[0115] The first determining unit 516 is configured to determine that the deduced trajectory pair is sufficient for the ego vehicle and the oncoming vehicle to complete the meeting if the deduced trajectory pair has completed a preset number of deduction processes.

[0116] In some embodiments, after controlling the ego vehicle to travel according to the ego vehicle deduced trajectory in the target trajectory pair, the apparatus further comprises:

[0117] A prediction unit 518 is configured to predict an ideal position of the oncoming vehicle based on a trajectory of the oncoming vehicle aligned with the target trajectory if the oncoming vehicle and the oncoming vehicle have not yet completed the meeting process;

[0118] A second determination unit 520 is configured to detect a deviation between the actual position of the oncoming vehicle and the ideal position. If the deviation exceeds a preset deviation threshold, the behavior samples in the pre-constructed driving paradigm behavior space are re-traversed to re-determine a target trajectory pair that is suitable for the ego vehicle.

[0119] The second control unit 522 is configured to continue controlling the ego vehicle to travel according to the ego vehicle deduced trajectory in the target trajectory if the deviation does not exceed a preset deviation threshold.

[0120] In some embodiments, controlling the ego vehicle to travel according to the ego vehicle deduced trajectory in the target trajectory pair includes: controlling the ego vehicle to travel according to the ego vehicle deduced trajectory in the target trajectory pair, and accumulating a number of decision steps;

[0121] After controlling the ego vehicle to travel according to the ego vehicle deduced trajectory in the target trajectory pair, the method further includes:

[0122] If the meeting process between the vehicle and the oncoming vehicle has not been completed, the accumulated number of decision steps is counted;

[0123] If the accumulated number of decision steps reaches a preset step threshold, the behavior samples in the pre-constructed driving paradigm behavior space are traversed again to determine a target trajectory pair that is adapted to the vehicle;

[0124] If the accumulated number of decision steps does not reach the preset step threshold, the ego vehicle continues to be controlled to travel according to the ego vehicle deduced trajectory in the target trajectory pair, and the number of decision steps is accumulated once.

[0125] In some embodiments, before performing quantitative evaluation on the plurality of trajectory pairs obtained after the traversal to determine a target trajectory pair suitable for the vehicle, the apparatus further comprises:

[0126] A second evaluation unit 524 is configured to perform collision risk evaluation on the plurality of trajectory pairs obtained after the traversal according to a preset risk evaluation rule;

[0127] The retaining unit 526 is configured to retain only a number of trajectory pairs with qualified collision risk assessment results according to the collision risk assessment result.

[0128] The implementation process of the functions and effects of each unit in the above-mentioned device is specifically described in the implementation process of the corresponding steps in the above-mentioned method, and will not be repeated here.

[0129] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to the partial description of the method embodiments. The device embodiments described above are only schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the disclosed solution. A person of ordinary skill in the art can understand and implement it without paying any creative work.

[0130] The systems, devices, modules, or units described in the above embodiments may be implemented by computer chips or entities, or by products having certain functions. A typical implementation device is a computer, which may be in the form of a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email transceiver, game console, tablet computer, wearable device, or any combination of these devices.

[0131] In a typical configuration, a computer includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0132] Memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0133] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can be implemented using any method or technology to store information. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic disk storage, quantum memory, graphene-based storage media or other magnetic storage devices, or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media such as modulated data signals and carrier waves.

[0134] The user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this disclosure are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions, and corresponding operation entrances shall be provided for users to choose to authorize or refuse.

[0135] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0136] The foregoing description describes specific embodiments of the present disclosure. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different from that described in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order shown or the sequential order to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0137] The terms used in one or more embodiments of the present disclosure are for the purpose of describing specific embodiments only and are not intended to limit one or more embodiments of the present disclosure. The singular forms "a," "the," and "the" used in one or more embodiments of the present disclosure and the appended claims are also intended to include plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0138] It should be understood that although the terms first, second, third, etc. may be used to describe various information in one or more embodiments of the present disclosure, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from each other. For example, without departing from the scope of one or more embodiments of the present disclosure, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0139] The above description is merely a preferred embodiment of one or more embodiments of the present disclosure and is not intended to limit one or more embodiments of the present disclosure. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of one or more embodiments of the present disclosure shall be included in the scope of protection of one or more embodiments of the present disclosure.

Claims

1. A method for predicting a vehicle trajectory, characterized in that: The method comprises: Acquire a vehicle real driving data set, wherein each data item in the vehicle real driving data set includes longitudinal acceleration and lateral offset parameters of the vehicle meeting; Filtering the data in the real driving dataset of the vehicle to retain only representative data of different driving modes, and constructing a driving paradigm behavior space based on the representative data of different driving modes; In response to detecting an oncoming vehicle, traversing behavior samples in a pre-constructed driving paradigm behavior space, wherein the behavior samples include longitudinal acceleration and lateral offset parameters of the vehicle; Each time a behavior sample is traversed, the trajectory of the ego vehicle and the oncoming vehicle are deduced based on the behavior sample and the current motion parameters of the ego vehicle and the oncoming vehicle, respectively, to obtain a trajectory pair of the ego vehicle and the oncoming vehicle; Performing quantitative evaluation on the plurality of trajectory pairs obtained after the traversal to determine a target trajectory pair suitable for the vehicle; The ego vehicle is controlled to travel according to the ego vehicle deduced trajectory in the target trajectory.

2. The method according to claim 1, characterized in that After traversing a behavior sample each time, the trajectory of the ego vehicle and the oncoming vehicle are deduced based on the behavior sample and the current motion parameters of the ego vehicle and the oncoming vehicle, respectively, to obtain a trajectory pair of the ego vehicle and the oncoming vehicle, including: Each time a behavior sample is traversed, the trajectory of the ego vehicle and the oncoming vehicle is deduced within a corresponding time period after the current moment based on the behavior sample and the current motion parameters of the ego vehicle and the oncoming vehicle. The behavior samples in the driving paradigm behavior space are retraversed. When another behavior sample is traversed, the above trajectory deduction process is repeated until the deduced trajectory pair is sufficient for the vehicle and the oncoming vehicle to complete the meeting.

3. The method according to claim 2, characterized in that The method further comprises: estimating the time required for the vehicle and the oncoming vehicle to complete the meeting, and determining that at least a preset number of deduction processes need to be performed based on the corresponding time of one deduction and the time required to complete the meeting; If the deduced trajectory pair has completed a preset number of deduction processes, it is determined that the deduced trajectory pair is sufficient for the ego vehicle and the oncoming vehicle to complete the meeting.

4. The method according to claim 1, wherein After controlling the ego vehicle to travel according to the ego vehicle deduced trajectory in the target trajectory pair, the method further includes: If the meeting process between the ego vehicle and the oncoming vehicle has not yet been completed, predicting the ideal position of the oncoming vehicle based on the trajectory of the oncoming vehicle aligned with the target trajectory; Detecting the deviation between the actual position of the oncoming vehicle and the ideal position, and if the deviation exceeds a preset deviation threshold, retraversing the behavior samples in the pre-constructed driving paradigm behavior space to re-determine a target trajectory pair that is suitable for the ego vehicle; If the deviation does not exceed the preset deviation threshold, the ego vehicle continues to be controlled to travel according to the ego vehicle deduced trajectory in the target trajectory.

5. The method according to claim 1 or 4, characterized in that Controlling the ego vehicle to travel according to the ego vehicle deduced trajectory in the target trajectory pair, comprising: controlling the ego vehicle to travel according to the ego vehicle deduced trajectory in the target trajectory pair, and accumulating a number of decision steps; After controlling the ego vehicle to travel according to the ego vehicle deduced trajectory in the target trajectory pair, the method further includes: If the meeting process between the vehicle and the oncoming vehicle has not been completed, the accumulated number of decision steps is counted; If the accumulated number of decision steps reaches a preset step threshold, the behavior samples in the pre-constructed driving paradigm behavior space are traversed again to determine a target trajectory pair that is adapted to the vehicle; If the accumulated number of decision steps does not reach the preset step threshold, the ego vehicle continues to be controlled to travel according to the ego vehicle deduced trajectory in the target trajectory pair, and the number of decision steps is accumulated once.

6. The method according to claim 1, wherein Before quantitatively evaluating the plurality of trajectory pairs obtained after the traversal to determine a target trajectory pair suitable for the vehicle, the method further includes: According to the preset risk assessment rules, collision risk assessment is performed on several trajectory pairs obtained after the traversal; According to the collision risk assessment result, only a number of trajectory pairs with qualified collision risk assessment results are retained.

7. A vehicle trajectory prediction device, characterized in that: The device comprises: an acquisition unit, configured to acquire a vehicle real driving data set, wherein each data item in the vehicle real driving data set includes longitudinal acceleration and lateral offset parameters of the vehicle meeting; a screening unit, configured to screen the data in the real vehicle driving data set, retain only representative data of different driving modes, and construct a driving paradigm behavior space based on the representative data of different driving modes; A traversal unit, configured to traverse behavior samples in a pre-constructed driving paradigm behavior space in response to detecting an oncoming vehicle, wherein the behavior samples include longitudinal acceleration and lateral offset parameters of the vehicle; a deduction unit configured to, after traversing a behavior sample each time, perform trajectory deduction on the ego vehicle and the oncoming vehicle based on the behavior sample and current motion parameters of the ego vehicle and the oncoming vehicle, respectively, to obtain a trajectory pair of the ego vehicle and the oncoming vehicle; an evaluation unit, configured to perform quantitative evaluation on the plurality of trajectory pairs obtained after the traversal to determine a target trajectory pair suitable for the ego vehicle; A control unit is used to control the vehicle to travel according to the deduced trajectory of the vehicle in the target trajectory.

8. A vehicle, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor implements the method according to any one of claims 1 to 6 by running the executable instructions.

9. A computer program product, characterized in that The method comprises a computer program / instruction, which implements the steps of the method according to any one of claims 1 to 6 when the computer program / instruction is executed by a processor.

Images