A method, device, equipment and storage medium for vehicle trajectory prediction

Abstract

The present disclosure provides a vehicle trajectory prediction method, device, equipment and storage medium. The method comprises: searching for an associated lane which can be driven to reach from a current lane of a target vehicle, and constructing at least one candidate trajectory of the target vehicle; predicting a reference trajectory of the target vehicle by using historical driving information of the target vehicle; and determining a future driving trajectory of the target vehicle from each candidate trajectory by using the reference trajectory. The present disclosure can realize reasonable prediction of the future driving trajectory of the target vehicle by using the current lane and the historical driving information of the target vehicle in any road section scenario, without considering the size of road test data in any road section scenario, and ensures comprehensive applicability of the vehicle trajectory prediction in different road section scenarios. Moreover, by using the specific calculation process of the related driving data of the target vehicle, the vehicle trajectory prediction has certain explainability on the basis of ensuring the accuracy of the vehicle trajectory prediction.

Description

Technical Field

[0001] This disclosure relates to the field of data processing technology, specifically to a method, apparatus, device, and storage medium for predicting vehicle trajectories. Background Technology

[0002] In the field of autonomous driving, in order to ensure the safe driving of the master vehicle, it is usually necessary to accurately predict the driving trajectories of other vehicles around the master vehicle in order to generate reasonable safe driving decisions for the master vehicle.

[0003] Currently, trajectory prediction for other vehicles surrounding the primary vehicle is typically achieved using data-driven algorithms based on neural networks. First, for different road segment scenarios, a large amount of road test data is acquired as training data to train a trajectory prediction model suitable for that specific road segment scenario. Then, by using the trajectory prediction models specific to different road segment scenarios, trajectory prediction for any vehicle within that scenario can be achieved.

[0004] However, different trajectory prediction models need to be trained separately for different road segments, which limits the applicability of vehicle trajectory prediction. Moreover, in some specific and complex road segments (such as roundabouts), there is a problem of insufficient training data due to the limited amount of road test data, which greatly affects the accuracy of the trained trajectory prediction model in predicting vehicle trajectories. Summary of the Invention

[0005] This disclosure provides a method, apparatus, device, and storage medium for vehicle trajectory prediction, ensuring the comprehensive applicability of vehicle trajectory prediction in different road segment scenarios and improving the accuracy and interpretability of vehicle trajectory prediction.

[0006] According to one aspect of this disclosure, a method for predicting vehicle trajectories is provided, the method comprising:

[0007] Search for associated lanes that the target vehicle can reach from its current lane, and construct at least one candidate trajectory for the target vehicle;

[0008] Using the historical driving information of the target vehicle, predict the reference trajectory of the target vehicle;

[0009] Using the reference trajectory, the future driving trajectory of the target vehicle is determined from each of the candidate trajectories.

[0010] According to another aspect of this disclosure, a vehicle trajectory prediction apparatus is provided, the apparatus comprising:

[0011] The candidate trajectory construction module is used to search for associated lanes that the target vehicle can reach from its current lane and construct at least one candidate trajectory for the target vehicle.

[0012] The reference trajectory prediction module is used to predict the reference trajectory of the target vehicle using the target vehicle's historical driving information;

[0013] The future trajectory determination module is used to determine the future driving trajectory of the target vehicle from each of the candidate trajectories using the reference trajectory.

[0014] According to another aspect of this disclosure, an electronic device is provided, the electronic device comprising:

[0015] A processor and a memory, the memory for storing a computer program, the processor for calling and running the computer program stored in the memory to perform a vehicle trajectory prediction method provided in one or more embodiments of this disclosure.

[0016] According to another aspect of this disclosure, a computer-readable storage medium is provided for storing a computer program that causes a computer to perform a vehicle trajectory prediction method as provided in one or more embodiments of this disclosure.

[0017] The technical solution provided in this disclosure allows for the construction of candidate trajectories for any target vehicle traveling around the main vehicle by searching all associated lanes accessible from its current lane. Furthermore, by utilizing the target vehicle's historical driving information, a reference trajectory can be predicted. This reference trajectory can then be used to determine the target vehicle's future trajectory from among the candidate trajectories. Thus, in any road segment scenario, a reasonable prediction of the target vehicle's future trajectory can be achieved using its current lane and historical driving information, without considering the amount of road test data in any given road segment scenario, ensuring the comprehensive applicability of vehicle trajectory prediction across different road segment scenarios. Moreover, the specific calculation process for the target vehicle's relevant driving data ensures the accuracy of vehicle trajectory prediction while also providing a degree of interpretability. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this disclosure, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1This is a flowchart illustrating a method for predicting vehicle trajectories according to an embodiment of the present disclosure;

[0020] Figure 2 This is a flowchart illustrating the specific process of constructing candidate trajectories for a target vehicle according to an embodiment of this disclosure;

[0021] Figure 3 This is a schematic diagram illustrating the principle of the vehicle trajectory prediction process in an embodiment of this disclosure;

[0022] Figure 4 This is a flowchart illustrating the specific process of predicting the reference trajectory of a target vehicle according to an embodiment of this disclosure;

[0023] Figure 5 This is a flowchart illustrating the specific similarity matching process between candidate trajectories and reference trajectories of a target vehicle, as shown in an embodiment of this disclosure.

[0024] Figure 6 This is a schematic block diagram illustrating a vehicle trajectory prediction device according to an embodiment of the present disclosure;

[0025] Figure 7 This is a schematic block diagram of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0026] The technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0027] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0028] To address the limitations of trajectory prediction models trained separately for different road segment scenarios, and the inaccuracy of such models in certain complex road segment scenarios, this disclosure presents a non-neural network-based vehicle trajectory prediction scheme that does not require a large amount of road test data. For any target vehicle traveling around the main vehicle, candidate trajectories can be constructed by searching all associated lanes accessible from the target vehicle's current lane. Furthermore, a reference trajectory can be predicted using the target vehicle's historical driving information. Then, using this reference trajectory, the target vehicle's future trajectory can be determined from the candidate trajectories. Thus, in any road segment scenario, a reasonable prediction of the target vehicle's future trajectory can be achieved using the target vehicle's current lane and historical driving information, without considering the amount of road test data in any particular road segment scenario, ensuring the comprehensive applicability of vehicle trajectory prediction across different road segment scenarios. Moreover, the specific calculation process for the target vehicle's relevant driving data ensures the accuracy of vehicle trajectory prediction while also providing a degree of interpretability.

[0029] Figure 1 This is a flowchart illustrating a method for predicting vehicle trajectory according to an embodiment of this disclosure. This method can be executed by the vehicle trajectory prediction apparatus provided in this disclosure. The vehicle trajectory prediction apparatus can be implemented in any software and / or hardware manner. Exemplarily, the vehicle trajectory prediction apparatus can be applied to any electronic device, including but not limited to tablets, mobile phones (such as foldable phones, large-screen phones, etc.), wearable devices, in-vehicle devices, laptops, ultra-mobile personal computers (UMPCs), netbooks, personal digital assistants (PDAs), smart TVs, smart screens, high-definition TVs, 4K TVs, smart speakers, smart projectors, and other computing devices suitable for the field of autonomous driving. This disclosure does not limit the specific type of electronic device.

[0030] Specifically, such as Figure 1 As shown, the method may include the following steps:

[0031] S110, search for related lanes that the target vehicle can reach from its current lane, and construct at least one candidate trajectory for the target vehicle.

[0032] In this disclosure, the target vehicle refers to other vehicles traveling around the main vehicle during its journey on any road segment. Since the actual driving conditions of the target vehicles may pose a certain risk to the safe driving of the main vehicle, it is necessary to adjust the main vehicle's trajectory accordingly.

[0033] As an optional implementation of this disclosure, the target vehicle can be determined in the following way: based on the main vehicle lane and adjacent lanes where the main vehicle is located, the target vehicles around the main vehicle are determined.

[0034] In other words, during the driving of the main vehicle in any road segment scenario, environmental data of the surrounding driving environment can be collected periodically. Then, by performing location analysis on this environmental data, the main vehicle's current lane and the adjacent lanes to the left and right can be determined. Furthermore, by analyzing each of the three lanes—the main vehicle lane and the adjacent lanes—and the vehicles traveling around the main vehicle in each lane, the target vehicles around the main vehicle can be obtained.

[0035] For example, the target vehicle in this disclosure can be an adjacent vehicle traveling in the direction directly in front of and behind the main vehicle (i.e., the main vehicle lane), an adjacent vehicle traveling in the direction directly in front of and behind the main vehicle (i.e., the left adjacent lane), or an adjacent vehicle traveling in the direction directly in front of and behind the main vehicle (i.e., the right adjacent lane), etc.

[0036] According to one or more embodiments of this disclosure, considering that in any road segment scenario, vehicles travel within designated lanes, the actual travel trajectory of the target vehicle can be represented by analyzing the lanes it traverses. Furthermore, due to the unpredictability of vehicle movement, to ensure the accuracy of vehicle trajectory prediction, vehicle trajectories are typically predicted periodically for a short period in the future. Therefore, candidate trajectories that the target vehicle may travel in the short future period can be determined by searching the lanes that the target vehicle may traverse during that period.

[0037] Specifically, the first step is to determine the target vehicle's current lane. The target vehicle's trajectory prediction will reference its current position within its current lane as its starting point for a short future timeframe. Therefore, the target vehicle's current position within its current lane is further determined.

[0038] Then, based on the lane intersections, taking the target vehicle's current position within its current lane as the starting point, the system determines the lanes the target vehicle can reach within a short period of time when traveling from this starting point in different intersecting lanes. These lanes are then considered as associated lanes in this disclosure. Furthermore, the lanes the target vehicle travels through when traveling from this starting point to any associated lane can be sequentially combined into a travel trajectory, which serves as a candidate trajectory in this disclosure. Following this method, a candidate trajectory can be obtained for each associated lane.

[0039] It should be noted that the candidate trajectory in this disclosure can be the trajectory of the target vehicle when it leaves the current lane and enters another lane to continue driving, or it can be the trajectory of the target vehicle when it stays in the current lane and does not leave.

[0040] S120 uses the target vehicle's historical driving information to predict the target vehicle's reference trajectory.

[0041] In order to determine the actual future driving trajectory of the target vehicle from the various candidate trajectories, it is necessary to further combine the target vehicle's historical driving information to predict the target vehicle's future driving situation.

[0042] According to one or more embodiments of this disclosure, the historical driving information of the target vehicle may include, but is not limited to, the driving direction, driving position, and driving speed of the target vehicle during its historical driving process. Therefore, by analyzing the various historical driving information of the target vehicle, the historical driving state of the target vehicle can be determined. Then, taking into account that the target vehicle generally travels at a relatively constant speed over a short period of time, the driving trajectory of the target vehicle over a short period of time in the future can be estimated based on the historical driving state of the target vehicle, serving as the reference trajectory in this disclosure.

[0043] It should be noted that there is no fixed execution order between S110 and S120 in this disclosure. They can be executed simultaneously or sequentially, and this disclosure does not limit this.

[0044] S130 uses the reference trajectory to determine the future driving trajectory of the target vehicle from each candidate trajectory.

[0045] It should be understood that since the target vehicle's reference trajectory is estimated from its historical driving status, there may be a small deviation from the actual lane markings. Therefore, the target vehicle's reference trajectory can represent a general direction of travel for the target vehicle over a short period of time in the future.

[0046] Therefore, after obtaining the reference trajectory and each candidate trajectory of the target vehicle, the driving deviation between each candidate trajectory and the reference trajectory is determined. Then, from all the candidate trajectories, the one with the smallest driving deviation from the reference trajectory is selected as the future driving trajectory of the target vehicle, ensuring the accuracy of vehicle trajectory prediction.

[0047] Moreover, the specific calculation process using the relevant driving data of the target vehicle in this disclosure can achieve a reasonable prediction of the future driving trajectory of the target vehicle. Therefore, while ensuring the accuracy of the vehicle trajectory prediction, it also makes the vehicle trajectory prediction more interpretable.

[0048] It should be noted that the vehicle trajectory prediction scheme provided in this disclosure is applicable to any road segment scenario, ensuring the comprehensive applicability of vehicle trajectory prediction across different road segment scenarios. However, since the vehicle trajectory prediction scheme provided in this disclosure does not require a large amount of road test data, it is mainly applicable to some specific and complex road segment scenarios. The amount of road test data in such road segment scenarios is usually small, which cannot support accurate training of the trajectory prediction model, such as roundabout scenarios. For vehicle trajectory prediction in ordinary road segment scenarios, since the amount of road test data is large enough to meet the accurate training requirements of the trajectory prediction model, it can still be achieved by training a trajectory prediction model for this ordinary road segment scenario.

[0049] Furthermore, considering that in each candidate trajectory, the target vehicle either exits its current lane and enters another lane, or remains within its current lane without exiting, it indicates that there are identical segments in the initial part of each candidate trajectory. The target vehicle in the reference trajectory also starts from its current position within its current lane. Therefore, when the reference trajectory and the target vehicle's current lane are not significantly different, there may be some candidate trajectories with similar driving deviations from the reference trajectory, making it impossible to accurately predict the target vehicle's future trajectory.

[0050] In this disclosure, considering that in order to ensure the accuracy of vehicle trajectory prediction, vehicle trajectory prediction is usually performed periodically at short intervals, after obtaining each candidate trajectory of the target vehicle, the candidate trajectories are filtered according to the order in which the target vehicle leaves its current lane within each candidate trajectory, so as to determine the future driving trajectory of the target vehicle from the remaining candidate trajectories after filtering.

[0051] In other words, the candidate trajectories are ranked according to the order in which the target vehicle leaves its current lane within each candidate trajectory. If there is a candidate trajectory in which the target vehicle remains in its current lane without ever leaving, that candidate trajectory is ranked first.

[0052] Then, a filtering ratio can be preset. Based on the total number of candidate trajectories and the filtering ratio, some of the lower-ranked candidate trajectories can be filtered out. The remaining candidate trajectories after filtering are usually those where the target vehicle is always traveling in the current lane and those where the target vehicle will first leave the current lane and enter another lane to continue traveling.

[0053] Therefore, determining the future trajectory of the target vehicle from the remaining filtered candidate trajectories involves judging whether the target vehicle will continue to travel in its current lane or exit from one of the nearest intersections in the near future. Then, after a short interval, the vehicle trajectory prediction is performed again. Furthermore, if the target vehicle remains in its current lane, the previously filtered candidate trajectories are reused for further trajectory prediction, thereby improving the accuracy of the prediction.

[0054] The technical solution provided in this disclosure, for any target vehicle traveling around the main vehicle, can construct various candidate trajectories of the target vehicle by searching all related lanes that the target vehicle can reach from its current lane. Furthermore, by utilizing the target vehicle's historical driving information, a reference trajectory of the target vehicle can be predicted. Then, using this reference trajectory, the future driving trajectory of the target vehicle can be determined from the various candidate trajectories. Thus, in any road segment scenario, a reasonable prediction of the target vehicle's future driving trajectory can be achieved using the target vehicle's current lane and historical driving information, without considering the amount of road test data in any road segment scenario, ensuring the comprehensive applicability of vehicle trajectory prediction across different road segment scenarios. Moreover, through the specific calculation process of the target vehicle's relevant driving data, the vehicle trajectory prediction is made interpretable to a certain extent while ensuring accuracy.

[0055] According to one or more embodiments of this disclosure, in order to accurately describe each designated lane in a map, each lane is typically divided into multiple lane blocks of a fixed length. Each lane block is then assigned a unique identifier, serving as the unit for storing lanes in the map. Furthermore, by combining and storing the identifiers of each lane block within each lane in the map, the lane can be represented. Moreover, each lane block is configured with its own geographical location information, and a virtual lane baseline is defined for each lane block, with the length of the lane baseline representing the length of the lane block.

[0056] In this disclosure, the lane baseline can be a lane centerline virtually drawn at the center of the lane. Furthermore, the length of the lane baseline can be the same for different lane blocks.

[0057] As described above, this disclosure can utilize lane blocks within each lane to search for all associated lanes that a target vehicle can reach from its current lane. The following section combines... Figure 2 The paper provides a detailed explanation of the specific process of using lane blocks to search for related lanes that a target vehicle can reach from its current lane, thereby constructing at least one candidate trajectory for the target vehicle.

[0058] like Figure 2 As shown, the specific process of constructing the candidate trajectory of the target vehicle may include the following steps:

[0059] S210, determine the target vehicle's current lane block in its current lane and its adjacent lane block in the adjacent lane.

[0060] According to one or more embodiments of this disclosure, the lane currently occupied by the target vehicle is composed of multiple lane blocks with fixed position information. Therefore, by determining the current driving position of the target vehicle, the current lane block within the current lane can be determined. Then, the left and right lanes of the current lane within the same road segment are determined as adjacent lanes in this disclosure.

[0061] Then, based on the current lane block position information and the position information of each lane block in the adjacent lane, the adjacent lane block of the target vehicle in the adjacent lane is determined.

[0062] As an optional implementation of this disclosure, the current lane block in the current lane and the adjacent lane block in the adjacent lane of the target vehicle can be determined by performing the following steps: determining the current lane block of the target vehicle based on the current driving position of the target vehicle; determining the adjacent lane block in the adjacent lane from a preset lane block storage structure.

[0063] In other words, by determining the current driving position of the target vehicle, the difference between the current driving position and the position information of each lane block is judged to determine the current lane block where the target vehicle is located.

[0064] In this disclosure, when storing lanes in a map, a predefined lane block storage structure is used. This lane block storage structure can be a multi-branch tree structure. In this multi-branch tree structure, for any node represented by a lane block, its child nodes are any other nodes that allow vehicles to enter from that lane block. For example, the child node of any lane block can be the next lane block adjacent to the lane where the lane block is located, or the first lane block in another lane that can be directly entered after exiting the lane block.

[0065] As can be seen from the above, two adjacent lane blocks within two adjacent lanes will be at the same node depth in the lane block storage structure. Therefore, based on the current lane block's node position in the lane block storage structure, the lane blocks at adjacent node positions at the same node depth can be determined as the adjacent lane blocks in this disclosure.

[0066] S220, taking each lane block in the current lane block and adjacent lane blocks as the target lane block, and determining the corresponding candidate lane block based on the future reference driving distance of the target vehicle and the cumulative length of the lane baseline from the target lane block to any drivable associated lane block.

[0067] Considering that a road segment may consist of multiple lanes, and a target vehicle can change lanes within the same segment, the candidate trajectory of the target vehicle may also start from the adjacent lane of its current position. Therefore, when analyzing the candidate trajectory of the target vehicle, both the current lane block and adjacent lane blocks within the adjacent lane can be used as the starting point for the target vehicle to search for drivable associated lane blocks, thereby ensuring the comprehensiveness of the candidate trajectory.

[0068] According to one or more embodiments of this disclosure, each lane block in the current lane block and adjacent lane blocks is first used as a target lane block, and then the target lane block is used as the starting point for the target vehicle to search for drivable associated lane blocks.

[0069] To ensure the accuracy of vehicle trajectory prediction, this disclosure typically determines the associated lanes that the target vehicle can reach within a short period of time in the future. Therefore, the travel distance of the target vehicle within the short period of time in the future is first determined as the future reference travel distance in this disclosure.

[0070] Specifically, the target vehicle's current speed can be used as its average speed over a short future period, which is defined as the first future time period. Then, using the target vehicle's current speed, the distance traveled by the target vehicle within the first future time period can be calculated, serving as a future reference travel distance.

[0071] Then, taking the target lane block as the starting point for the target vehicle, we determine any associated lane block that the target vehicle can reach after starting from the target lane block. Here, each associated lane block can be a child node of the multi-branch tree structure rooted at the target lane block. Furthermore, for each associated lane block, we determine all the lane blocks the target vehicle passes through when traveling from the target lane block to any associated lane block, and calculate the cumulative length of the lane baselines for each of the passed lane blocks to represent the travel distance required for the target vehicle to travel from the target lane block to that associated lane block.

[0072] Furthermore, in order to accurately determine the candidate trajectories that the target vehicle may travel in the near future, the target vehicle must be able to travel for at least a short period of time in the future.

[0073] Therefore, if the cumulative length of the lane baseline from the target lane block to an associated lane block is greater than the future reference travel distance in this disclosure, it indicates that the target vehicle has already met the requirement of traveling for at least a short period of time in the future when it travels from the target lane block to the associated lane block, and thus the associated lane block is considered a candidate lane block. However, if the cumulative length of the lane baseline from the target lane block to an associated lane block is less than or equal to the future reference travel distance in this disclosure, it indicates that the target vehicle has not met the requirement of traveling for at least a short period of time in the future when it travels from the target lane block to the associated lane block, and thus the candidate trajectory that the target vehicle may travel in the short period of time in the future cannot be determined, and therefore the associated lane block is excluded. Following the above judgment principles, candidate lane blocks that can represent candidate trajectories can be selected from each associated lane block.

[0074] According to one or more embodiments of this disclosure, in order to ensure convenient determination of candidate lane blocks, this disclosure can determine each candidate lane block by performing the following steps through a loop:

[0075] S21, take the target lane block as the current lane block, and traverse the next lane block of the current lane block.

[0076] Since the target lane block is the starting point for the target vehicle's travel in the near future, it is necessary to start with the target lane block as the current lane block and then traverse the next lane block of the current lane block.

[0077] For example, when storing lane blocks in a map using a multi-branch tree structure, the next lane block of the current lane block can be the lane blocks represented by the child nodes of the current lane block as the parent node in the multi-branch tree structure.

[0078] S22, perform candidate judgment: determine whether the cumulative length of the lane baseline when traveling from the target lane block to the next lane block is greater than the future reference travel distance of the target vehicle. If so, the next lane block is taken as a candidate lane block; otherwise, the next lane block is taken as the new current lane block, and the new next lane block is traversed.

[0079] During the traversal of the next lane block, the cumulative length of the lane baseline as the target vehicle travels from the target lane block to each traversed next lane block is calculated. Then, it is determined whether the cumulative length of the lane baseline as the target vehicle travels from the target lane block to any next lane block is greater than the target vehicle's future reference travel distance. If so, it means that the target vehicle has met the requirement of traveling at least a short period of time from the target lane block to the next lane block, and the next lane block is selected as a candidate lane block. If not, it means that the target vehicle has not met the requirement of traveling at least a short period of time from the target lane block to the next lane block, and the next lane block is selected as the new current lane block, and the traversal of new next lane blocks continues.

[0080] S23, continue to perform candidate judgment to obtain at least one of the new candidate lane block and the new current lane block, until there is no new current lane block.

[0081] After obtaining a new next lane block, the above candidate judgment steps will continue. That is, it will determine whether the cumulative length of the lane baseline when the target vehicle travels from the target lane block to the new next lane block is greater than the future reference travel distance of the target vehicle. In this way, it will be analyzed whether the target vehicle has reached the requirement of traveling at least a short period of time in the future when it travels from the target lane block to the new next lane block, so as to obtain at least one of the new candidate lane block and the new current lane block.

[0082] As described above, for each new current lane block obtained after candidate evaluation, the process continues to iterate through the next lane block and repeat the candidate evaluation steps. This cycle continues until no new current lane blocks exist, at which point all candidate lane blocks are obtained, and the loop stops.

[0083] Taking the current lane block in a roundabout segment as the target lane block as an example, such as... Figure 3 As shown, the target vehicle is represented by a five-pointed star symbol "★", and each lane is divided into multiple lane blocks. The target lane block is initially set as the first lane block in the loop, and the loop iterates through each subsequent lane block. If the cumulative length of the lane baseline from the first lane block to a certain next lane block is greater than the target vehicle's future reference travel distance, then that next lane block is designated as a candidate lane block, and the traversal under that next lane block stops. If the cumulative length of the lane baseline from the first lane block to a certain next lane block is less than or equal to the target vehicle's future reference travel distance, then that next lane block is designated as the new current lane block, and the loop continues to generate new next lane blocks, repeating the above steps. This process is repeated until all candidate lane blocks reachable from the target vehicle's current lane block are obtained.

[0084] Furthermore, by using adjacent lane blocks as target lane blocks and performing the same steps as described above, we can obtain the candidate lane blocks that the target vehicle can reach by starting from the adjacent lane blocks in the adjacent lanes.

[0085] S230: Construct candidate trajectories for the target vehicle based on the lane baselines it travels from the target lane block to the candidate lane block.

[0086] After identifying all candidate lane blocks accessible from the target lane block, the lane blocks traversed by the target vehicle when traveling from the target lane block to any candidate lane block can be determined. Then, for each candidate lane block, connecting the lane baselines of the traversed lane blocks yields the candidate trajectory of the target vehicle traveling from the target lane block to the candidate lane block. By following these steps, multiple candidate trajectories for the target vehicle can be obtained.

[0087] like Figure 3 As shown, it consists of multiple " The line segments formed by the symbols in different lanes represent the candidate trajectories in this disclosure.

[0088] It should be noted that, in order to accurately record the lane blocks that the target vehicle passes through when it starts traveling from the target lane block, during the cyclic traversal of lane blocks, this disclosure uses a first linked list to store the lane sequence composed of the target lane block and the next lane block when traversing the next lane block of the current lane block, with the target lane block as the current lane block.

[0089] For example, if the target lane block is 'a', and the next lane blocks of the target lane block are b1, b2, and b3 respectively, then according to the format of the first linked list, there will be three lane sequences: a→b1, a→b2, and a→b3.

[0090] Correspondingly, if the cumulative length of the lane baseline when traveling from the target lane block to a certain next lane block is less than or equal to the future reference travel distance of the target vehicle, then the next lane block is taken as the new current lane block. When traversing the new next lane block, a new next lane block will be added to the lane sequence corresponding to the next lane block to update the lane sequence in the first linked list.

[0091] For example, if the first linked list stores three lane sequences a→b1, a→b2, and a→b3 respectively, after performing the above candidate judgment, the next lane blocks b1 and b2 are determined as candidate lane blocks, and the next lane block b3 is the new current lane block. Furthermore, the next lane blocks c1 and c2 of the new current lane block b3 are found. Then, the lane sequences a→b1 and a→b2 stored in the first linked list remain unchanged, while a→b3 is updated to a→b3→c1 and a→b3→c2, thereby updating the lane sequences in the first linked list.

[0092] In summary, during the cyclic traversal of the aforementioned lane blocks, the lane sequence in the first linked list is continuously updated. Therefore, in this disclosure, the cumulative length of the lane baseline from the target lane block to any drivable associated lane block is the total length of the lane baseline of the lane sequence in the first linked list with any drivable associated lane block as the terminal node.

[0093] As one or more embodiments of this disclosure, the lane baseline traversed when traveling from a target lane block to a candidate lane block can be determined by: determining a second linked list with the candidate lane block as the terminal node, and using the lane baseline of the lane sequence currently stored in the second linked list as the candidate trajectory of the target vehicle.

[0094] After completing the loop traversal of the lane blocks, the lane sequence in the first linked list is updated, and the terminal nodes of the first linked list are all candidate lane blocks. Therefore, the first linked list obtained after completing the loop traversal of the lane blocks is used as the second linked list in this disclosure. Then, the lane baseline of each lane sequence currently stored in the second linked list is determined as the candidate trajectory of the target vehicle.

[0095] The technical solution provided in this disclosure quickly determines the corresponding candidate lane blocks by comparing the future reference driving distance of the target vehicle with the cumulative length of the lane baseline from the target lane block to any drivable associated lane block. It then constructs the corresponding candidate trajectories by using the lane baselines traversed when driving from the target lane block to each candidate lane block, thereby improving the convenience, effectiveness, and accuracy of candidate trajectory construction.

[0096] According to one or more embodiments of this disclosure, considering that the driving trajectory of the target vehicle over a short period of time will be related to the yaw rate and driving speed of the target vehicle, the prediction of the reference trajectory of the target vehicle can be made by estimating the front wheel steering angle change rate of the target vehicle based on the historical driving information of the target vehicle; and by predicting the reference trajectory of the target vehicle based on the front wheel steering angle change rate and the current driving information of the target vehicle.

[0097] In other words, firstly, based on the target vehicle's historical driving information, the steering changes of the target vehicle during its historical driving process are analyzed to determine the front wheel steering angle change rate of the target vehicle. Then, based on the front wheel steering angle information in the target vehicle's current driving information, the front wheel steering angle change rate of the target vehicle is used to predict the specific front wheel steering changes of the target vehicle during a short period of time in the future, thereby determining the target vehicle's reference trajectory.

[0098] The following is combined with Figure 4 The specific prediction process for the reference trajectory of the target vehicle is explained in detail.

[0099] like Figure 4 As shown, the specific prediction process for the reference trajectory of the target vehicle may include the following steps:

[0100] S410 inputs the historical driving information of the target vehicle into the pre-built kinematic model to determine the historical front wheel steering angle of the target vehicle.

[0101] The kinematic model is a network model trained using a large number of vehicle trajectories and various specific driving parameters, capable of analyzing vehicle driving conditions. The target vehicle's historical driving information can include its historical speed, yaw rate, heading, historical position, and historical driving trajectory within a historical time period. For example... Figure 3 As shown, it consists of a large number of " The line segment composed of symbols represents the historical driving trajectory of the target vehicle in this disclosure.

[0102] After obtaining various historical driving information of the target vehicle during a historical period, this information can be directly input into a pre-built kinematic model. Before this, zero-phase filtering can be used to filter the historical driving information of the target vehicle before inputting it into the kinematic model to eliminate noise present in the historical driving information. Then, by analyzing the driving state of the historical driving information through the kinematic model, the historical front wheel steering angle of the target vehicle during the historical driving process can be determined.

[0103] S420 estimates the rate of change of the front wheel steering angle of the target vehicle based on the historical front wheel steering angle of the target vehicle.

[0104] By using the least squares algorithm to process each historical front wheel steering angle of the target vehicle during its historical driving process, the rate of change of the target vehicle's front wheel steering angle can be estimated. Prior to this, zero-phase filtering can also be used to filter the historical front wheel steering angles of the target vehicle to eliminate noise.

[0105] S430, based on the rate of change of the front wheel steering angle of the target vehicle, predicts the future front wheel steering angle of the target vehicle in the second future time period.

[0106] After obtaining the rate of change of the front wheel steering angle of the target vehicle, the current front wheel steering angle of the target vehicle is used as the initial value. Then, by adjusting this initial value according to the rate of change of the front wheel steering angle of the target vehicle over a future time period, the future front wheel steering angle of the target vehicle in the second future time period can be estimated.

[0107] The second future time period and the first future time period used for the candidate trajectory of the target vehicle can be the same or different, but the difference between the two is not significant.

[0108] S440 inputs the target vehicle's current driving information and future front wheel steering angle into a pre-built kinematic model to predict the target vehicle's reference trajectory.

[0109] By inputting the target vehicle's current driving information and future front wheel steering angle into a pre-built kinematic model, the driving state of the target vehicle in the second future time period can be analyzed through the kinematic model to obtain the target vehicle's reference trajectory.

[0110] like Figure 3 As shown, the segment consisting of multiple "+" symbols represents the reference trajectory in this disclosure.

[0111] The technical solution provided in this disclosure analyzes the future front wheel angle of the target vehicle in the second future time period by estimating the rate of change of the front wheel angle. Then, it uses a pre-constructed kinematic model to analyze the driving state of the target vehicle based on its current driving information and future front wheel angle, thereby obtaining the reference trajectory of the target vehicle and ensuring the authenticity and validity of the reference trajectory.

[0112] According to one or more embodiments of this disclosure, after obtaining multiple candidate trajectories and a reference trajectory for a target vehicle, the candidate trajectory with the best match can be used as the future driving trajectory of the target vehicle by performing similarity matching between each candidate trajectory and the reference trajectory.

[0113] In other words, various similarity algorithms, such as Euclidean distance, Pearson correlation coefficient, and cosine similarity, can be used to calculate the trajectory similarity between each candidate trajectory and the reference trajectory, thereby determining the driving deviation between each candidate trajectory and the reference trajectory. The candidate trajectory with the optimal match has the smallest driving deviation from the reference trajectory; therefore, the candidate trajectory with the optimal match is taken as the future driving trajectory of the target vehicle. The following section combines... Figure 5 Taking Euclidean distance as an example, this paper provides a detailed explanation of the specific similarity matching process between the candidate trajectory and the reference trajectory of the target vehicle.

[0114] like Figure 5 As shown, the specific similarity matching process between the candidate trajectory and the reference trajectory of the target vehicle may include the following steps:

[0115] S510: Resample each candidate trajectory using the current speed of the target vehicle to obtain a candidate trajectory that is aligned with the time axis of the reference trajectory.

[0116] Since each candidate trajectory of the target vehicle is mainly composed of lane baselines of multiple lane blocks, it does not carry any travel time information of the target vehicle. However, the reference trajectory is estimated from the historical travel information of the target vehicle, and is obtained by combining the estimated travel positions of the target vehicle at various future times. Moreover, the similarity matching between the candidate trajectory and the reference trajectory mainly determines the difference between two specific positions of the target vehicle on the candidate trajectory and the reference trajectory at the same time.

[0117] Therefore, before performing similarity matching on each candidate trajectory and the reference trajectory, it is necessary to resample each candidate trajectory at the corresponding time step.

[0118] In this disclosure, the current speed of the target vehicle can be used as the average speed of the target vehicle over a short period of time in the future. Then, for each candidate trajectory, the current speed is used to calculate the distance traveled by the target vehicle at each moment within the time axis used by the reference trajectory, and the sampling points corresponding to the distance traveled at each moment are determined from the candidate trajectory, thereby obtaining a candidate trajectory that is resampled and aligned with the time axis of the reference trajectory.

[0119] S520 performs Euclidean weighted calculations on each resampled candidate trajectory and the reference trajectory to determine the future driving trajectory of the target vehicle.

[0120] The Euclidean weighted formula in this disclosure is:

[0121] Where d is the Euclidean weighted distance between the candidate trajectory and the reference trajectory, used to represent the similarity between the two trajectories. The smaller d is, the higher the similarity between the candidate trajectory and the reference trajectory; while the larger d is, the lower the similarity between the candidate trajectory and the reference trajectory.

[0122] Furthermore, T represents the time axis length after aligning the candidate trajectory and the reference trajectory, and w t The distance weight is the distance at any travel time t within the time axis. This refers to the position information of the target vehicle in the candidate trajectory at any given travel time t. This refers to the position information of the target vehicle on the reference trajectory at any given travel time t.

[0123] Using the above formula, the Euclidean weighted distance between each resampled candidate trajectory and the reference trajectory can be calculated. At this point, the candidate trajectory with the smallest Euclidean weighted distance has the highest similarity to the reference trajectory; therefore, the candidate trajectory with the smallest Euclidean weighted distance is taken as the future driving trajectory of the target vehicle.

[0124] It should be noted that the reliability of the predicted trajectory position in the reference trajectory decreases with increasing time points, thus having a smaller impact on trajectory similarity matching. Therefore, to ensure the accuracy of similarity matching between candidate trajectories and reference trajectories, the weight of the same candidate trajectory at different time points is set to decrease with increasing time points.

[0125] The technical solution provided in this disclosure first resamples each candidate trajectory to align the time axis between the candidate trajectory and the reference trajectory, and then uses Euclidean weighted distance analysis to analyze the similarity between each candidate trajectory and the reference trajectory, thereby obtaining the future driving trajectory of the target vehicle and improving the accuracy of similarity matching between the candidate trajectory and the reference trajectory.

[0126] Figure 6 This is a schematic block diagram illustrating a vehicle trajectory prediction device according to an embodiment of this disclosure. Figure 6 As shown, the device 600 may include:

[0127] The candidate trajectory construction module 610 is used to search for associated lanes that the target vehicle can reach from its current lane, and construct at least one candidate trajectory for the target vehicle.

[0128] The reference trajectory prediction module 620 is used to predict the reference trajectory of the target vehicle using the historical driving information of the target vehicle.

[0129] The future trajectory determination module 630 is used to determine the future driving trajectory of the target vehicle from each of the candidate trajectories using the reference trajectory.

[0130] According to one or more embodiments of this disclosure, the candidate trajectory construction module 610 may include: a first lane block unit, a second lane block unit, and a candidate trajectory determination unit.

[0131] The first lane block unit is used to determine the current lane block of the target vehicle in its current lane and the adjacent lane block in the adjacent lane.

[0132] The second lane block unit is used to take each of the current lane block and the adjacent lane blocks as the target lane block, and determine the corresponding candidate lane block based on the future reference driving distance of the target vehicle and the cumulative length of the lane baseline from the target lane block to any drivable associated lane block.

[0133] The candidate trajectory determination unit is used to construct a candidate trajectory of the target vehicle based on the lane baseline it passes through when traveling from the target lane block to the candidate lane block.

[0134] According to one or more embodiments of this disclosure, the second lane block unit may be specifically used for:

[0135] Take the target lane block as the current lane block and traverse the next lane block of the current lane block;

[0136] Candidate determination: Determine whether the cumulative length of the lane baseline when traveling from the target lane block to the next lane block is greater than the future reference travel distance of the target vehicle. If so, the next lane block is selected as the candidate lane block. Otherwise, the next lane block is selected as the new current lane block, and the new next lane block is traversed.

[0137] Continue performing the candidate determination to obtain at least one of the new candidate lane block and the new current lane block, until no new current lane block exists.

[0138] According to one or more embodiments of this disclosure, the vehicle trajectory prediction device 600 may further include: a linked list storage module. Wherein,

[0139] A linked list storage module is used to store a lane sequence consisting of the target lane block and the next lane block using a first linked list.

[0140] Correspondingly, the vehicle trajectory prediction device 600 may also include a linked list update module. Wherein,

[0141] A linked list update module is used to add the new next lane block within the lane sequence to update the lane sequence in the first linked list.

[0142] According to one or more embodiments of this disclosure, the cumulative length of the lane baseline from the target lane block to any drivable associated lane block is the total length of the lane baseline of the lane sequence in the first linked list with the any drivable associated lane block as the terminal node.

[0143] According to one or more embodiments of this disclosure, the candidate trajectory determination unit may be specifically used for:

[0144] A second linked list is determined with the candidate lane block as the terminal node, and the lane baseline of the lane sequence currently stored in the second linked list is used as the candidate trajectory of the target vehicle.

[0145] According to one or more embodiments of this disclosure, the first lane block unit may be specifically used for:

[0146] Based on the current driving position of the target vehicle, determine the current lane block where the target vehicle is located;

[0147] From the preset lane block storage structure, determine the adjacent lane blocks of the current lane block in the adjacent lane.

[0148] According to one or more embodiments of this disclosure, the lane block storage structure is a multi-branch tree structure.

[0149] According to one or more embodiments of this disclosure, the vehicle trajectory prediction apparatus 600 may further include: a reference distance determination module. Wherein,

[0150] The reference distance determination module is used to calculate the travel distance of the target vehicle in a first future time period using the current speed of the target vehicle, and use it as the future reference travel distance.

[0151] According to one or more embodiments of this disclosure, the reference trajectory prediction module 620 may include: a rotation angle change rate estimation unit and a reference trajectory prediction unit. Wherein,

[0152] Angle change rate prediction unit is used to predict the front wheel steering angle change rate of the target vehicle based on the target vehicle's historical driving information;

[0153] The reference trajectory prediction unit is used to predict the reference trajectory of the target vehicle based on the front wheel steering angle change rate and the current driving information of the target vehicle.

[0154] According to one or more embodiments of this disclosure, the angle change rate prediction unit may be specifically used for:

[0155] The historical driving information of the target vehicle is input into a pre-constructed kinematic model to determine the historical front wheel steering angle of the target vehicle.

[0156] Based on the historical front wheel steering angle of the target vehicle, the rate of change of the front wheel steering angle of the target vehicle is estimated.

[0157] According to one or more embodiments of this disclosure, the reference trajectory prediction unit may be specifically used for:

[0158] Based on the rate of change of the front wheel steering angle of the target vehicle, the future front wheel steering angle of the target vehicle in the second future time period is estimated;

[0159] The current driving information of the target vehicle and the future front wheel steering angle are input into a pre-built kinematic model to predict the reference trajectory of the target vehicle.

[0160] According to one or more embodiments of this disclosure, the future trajectory determination module 630 may be specifically used for:

[0161] The candidate trajectories and the reference trajectories are matched for similarity, and the candidate trajectory with the best match is taken as the future driving trajectory of the target vehicle.

[0162] According to one or more embodiments of this disclosure, the future trajectory determination module 630 may be specifically used for:

[0163] The candidate trajectories are resampled using the current speed of the target vehicle to obtain candidate trajectories that are aligned with the time axis of the reference trajectory.

[0164] The resampled candidate trajectories are compared with the reference trajectory using Euclidean weighted calculation to determine the future driving trajectory of the target vehicle.

[0165] According to one or more embodiments of this disclosure, the weight of the same candidate trajectory at different time points decreases as the time point increases.

[0166] According to one or more embodiments of this disclosure, the vehicle trajectory prediction apparatus 600 may further include: a candidate trajectory filtering module. Wherein,

[0167] The candidate trajectory filtering module is used to filter each candidate trajectory according to the exit order of the target vehicle when it leaves the current lane in each candidate trajectory, so as to determine the future driving trajectory of the target vehicle from the remaining candidate trajectories after filtering.

[0168] According to one or more embodiments of this disclosure, the vehicle trajectory prediction apparatus 600 may further include: a target vehicle determination module. Wherein,

[0169] The target vehicle determination module is used to determine the target vehicles around the main vehicle based on the main vehicle lane and adjacent lanes where the main vehicle is located.

[0170] In this embodiment, for any target vehicle traveling around the main vehicle, candidate trajectories for the target vehicle can be constructed by searching all associated lanes accessible from its current lane. Furthermore, a reference trajectory for the target vehicle can be predicted using its historical driving information. Then, using this reference trajectory, the future driving trajectory of the target vehicle can be determined from the candidate trajectories. Thus, in any road segment scenario, a reasonable prediction of the target vehicle's future driving trajectory can be achieved using its current lane and historical driving information, without considering the amount of road test data in any particular road segment scenario, ensuring the comprehensive applicability of vehicle trajectory prediction across different road segment scenarios. Moreover, through the specific calculation process of the target vehicle's relevant driving data, the accuracy of vehicle trajectory prediction is guaranteed, while also ensuring a certain degree of interpretability.

[0171] It should be understood that the device embodiments and method embodiments can correspond to each other, and similar descriptions can be referred to the method embodiments. To avoid repetition, further details will not be provided here. Specifically, Figure 6The apparatus 600 shown can execute any of the method embodiments provided in this disclosure, and the foregoing and other operations and / or functions of each module in the apparatus 600 are respectively for implementing the corresponding processes in the various methods of the embodiments of this disclosure. For the sake of brevity, they will not be described in detail here.

[0172] The apparatus 600 of this disclosure embodiment has been described above from the perspective of functional modules in conjunction with the accompanying drawings. It should be understood that this functional module can be implemented in hardware, in software instructions, or in a combination of hardware and software modules. Specifically, the steps of the method embodiments in this disclosure can be completed by integrated logic circuits in the processor's hardware and / or by software instructions. The steps of the method disclosed in this disclosure embodiment can be directly manifested as execution by a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. Optionally, the software module can reside in a mature storage medium in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, etc. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps in the above method embodiments.

[0173] Figure 7 This is a schematic block diagram of an electronic device provided in an embodiment of this disclosure.

[0174] like Figure 7 As shown, the electronic device 700 may include:

[0175] The system includes a memory 710 for storing computer programs and a processor 720 for transferring program code to the processor 720. In other words, the processor 720 can retrieve and run the computer program from the memory 710 to implement the methods described in this embodiment.

[0176] For example, the processor 720 can be used to execute the above-described method embodiments according to instructions in the computer program.

[0177] In some embodiments of this disclosure, the processor 720 may include, but is not limited to:

[0178] General-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0179] In some embodiments of this disclosure, the memory 710 includes, but is not limited to:

[0180] Volatile memory and / or non-volatile memory. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), and Direct Rambus RAM (DR RAM).

[0181] In some embodiments of this disclosure, the computer program may be divided into one or more modules, which are stored in the memory 710 and executed by the processor 720 to perform the method provided in this disclosure. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the electronic device.

[0182] like Figure 7 As shown, the electronic device may also include:

[0183] Transceiver 730, which can be connected to processor 720 or memory 710.

[0184] The processor 720 can control the transceiver 730 to communicate with other devices; specifically, it can send information or data to other devices or receive information or data sent by other devices. The transceiver 730 may include a transmitter and a receiver. The transceiver 730 may further include antennas, and the number of antennas may be one or more.

[0185] It should be understood that the various components in the electronic device are connected through a bus system, which includes a data bus, a power bus, a control bus, and a status signal bus.

[0186] This disclosure also provides a computer storage medium storing a computer program thereon, which, when executed by a computer, enables the computer to perform the methods of the above-described method embodiments. Alternatively, this disclosure also provides a computer program product containing instructions that, when executed by a computer, cause the computer to perform the methods of the above-described method embodiments.

[0187] When implemented using software, it can be implemented wholly or partially as a computer program product. This computer program product includes one or more computer instructions. When these computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this disclosure are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., digital video disc (DVD)), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0188] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this disclosure.

[0189] In the embodiments provided in this disclosure, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0190] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. For example, the functional modules in the various embodiments of this disclosure may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0191] The above are merely specific embodiments of this disclosure, but the scope of protection of this disclosure is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this disclosure should be included within the scope of protection of this disclosure. Therefore, the scope of protection of this disclosure should be determined by the scope of the claims.

Claims

1. A method for predicting vehicle trajectories, characterized in that, include: Search for associated lanes that the target vehicle can reach from its current lane, and construct at least one candidate trajectory for the target vehicle; Using the historical driving information of the target vehicle, predict the reference trajectory of the target vehicle; Using the reference trajectory, the future driving trajectory of the target vehicle is determined from each of the candidate trajectories; The step of searching for related lanes that the target vehicle can reach from its current lane, and constructing at least one candidate trajectory for the target vehicle, includes: Determine the target vehicle's current lane block within its current lane and its adjacent lane blocks within adjacent lanes; Each lane block in the current lane block and the adjacent lane blocks is taken as the target lane block. Based on the future reference driving distance of the target vehicle and the cumulative length of the lane baseline from the target lane block to any drivable associated lane block, the corresponding candidate lane block is determined. Based on the lane baselines traversed when traveling from the target lane block to the candidate lane block, a candidate trajectory of the target vehicle is constructed. The step of determining the corresponding candidate lane block based on the future reference driving distance of the target vehicle and the cumulative length of the lane baseline from the target lane block to any drivable associated lane block includes: Take the target lane block as the current lane block and traverse the next lane block of the current lane block; Candidate determination: Determine whether the cumulative length of the lane baseline when traveling from the target lane block to the next lane block is greater than the future reference travel distance of the target vehicle. If yes, then the next lane block is taken as the candidate lane block. If no, then the next lane block is taken as the new current lane block, and the new next lane block is traversed. Continue performing the candidate determination to obtain at least one of the new candidate lane block and the new current lane block, until no new current lane block exists.

2. The method according to claim 1, characterized in that, When using the target lane block as the current lane block and traversing the next lane block of the current lane block, the method further includes: A first linked list is used to store the lane sequence consisting of the target lane block and the next lane block; Correspondingly, when traversing the new next lane block as the new current lane block, the method also includes: The new next lane block is added within the lane sequence to update the lane sequence in the first linked list.

3. The method according to claim 2, characterized in that, The cumulative length of the lane baseline from the target lane block to any drivable associated lane block is the total length of the lane baseline of the lane sequence in the first linked list with the any drivable associated lane block as the terminal node.

4. The method according to claim 2, characterized in that, The step of constructing a candidate trajectory for the target vehicle based on the lane baseline it traverses when traveling from the target lane block to the candidate lane block includes: A second linked list is determined with the candidate lane block as the terminal node, and the lane baseline of the lane sequence currently stored in the second linked list is used as the candidate trajectory of the target vehicle.

5. The method according to claim 1, characterized in that, Determining the target vehicle's current lane block within its current lane and its adjacent lane blocks in adjacent lanes includes: Based on the current driving position of the target vehicle, determine the current lane block where the target vehicle is located; From the preset lane block storage structure, determine the adjacent lane blocks of the current lane block in the adjacent lane.

6. The method according to claim 5, characterized in that, The lane block storage structure is a multi-branch tree structure.

7. The method according to claim 1, characterized in that, The method further includes: Using the current speed of the target vehicle, the distance traveled by the target vehicle in the first future time period is calculated and used as the future reference travel distance.

8. The method according to claim 1, characterized in that, The step of predicting the reference trajectory of the target vehicle using its historical driving information includes: Based on the historical driving information of the target vehicle, estimate the rate of change of the front wheel steering angle of the target vehicle; Based on the front wheel steering angle change rate and the current driving information of the target vehicle, the reference trajectory of the target vehicle is predicted.

9. The method according to claim 8, characterized in that, The step of estimating the front wheel steering angle change rate of the target vehicle based on its historical driving information includes: The historical driving information of the target vehicle is input into a pre-constructed kinematic model to determine the historical front wheel steering angle of the target vehicle. Based on the historical front wheel steering angle of the target vehicle, the rate of change of the front wheel steering angle of the target vehicle is estimated.

10. The method according to claim 8, characterized in that, The step of predicting the reference trajectory of the target vehicle based on the front wheel steering angle change rate and the current driving information of the target vehicle includes: Based on the rate of change of the front wheel steering angle of the target vehicle, the future front wheel steering angle of the target vehicle in the second future time period is estimated; The current driving information of the target vehicle and the future front wheel steering angle are input into a pre-built kinematic model to predict the reference trajectory of the target vehicle.

11. The method according to claim 1, characterized in that, Determining the future driving trajectory of the target vehicle from each of the candidate trajectories using the reference trajectory includes: The candidate trajectories and the reference trajectories are matched for similarity, and the candidate trajectory with the best match is taken as the future driving trajectory of the target vehicle.

12. The method according to claim 11, characterized in that, The similarity matching of each candidate trajectory and the reference trajectory includes: The candidate trajectories are resampled using the current speed of the target vehicle to obtain candidate trajectories that are aligned with the time axis of the reference trajectory. The resampled candidate trajectories are compared with the reference trajectory using Euclidean weighted calculation to determine the future driving trajectory of the target vehicle.

13. The method according to claim 12, characterized in that, The weight of the same candidate trajectory at different time points decreases as the time point increases.

14. The method according to any one of claims 1-13, characterized in that, After searching for associated lanes that the target vehicle can reach from its current lane and constructing at least one candidate trajectory for the target vehicle, the method further includes: Based on the exit order of the target vehicle when it leaves its current lane within each of the candidate trajectories, the candidate trajectories are filtered to determine the future driving trajectory of the target vehicle from the remaining candidate trajectories after filtering.

15. The method according to any one of claims 1-13, characterized in that, The method further includes: Based on the main vehicle's lane and adjacent lanes, target vehicles around the main vehicle are identified.

16. A device for predicting vehicle trajectory, characterized in that, include: The candidate trajectory construction module is used to search for associated lanes that the target vehicle can reach from its current lane and construct at least one candidate trajectory for the target vehicle. The reference trajectory prediction module is used to predict the reference trajectory of the target vehicle using the target vehicle's historical driving information; A future trajectory determination module is used to determine the future driving trajectory of the target vehicle from each of the candidate trajectories using the reference trajectory; The candidate trajectory construction module is further used for: Determine the target vehicle's current lane block within its current lane and its adjacent lane blocks within adjacent lanes; Each lane block in the current lane block and the adjacent lane blocks is taken as the target lane block. Based on the future reference driving distance of the target vehicle and the cumulative length of the lane baseline from the target lane block to any drivable associated lane block, the corresponding candidate lane block is determined. Based on the lane baselines traversed when traveling from the target lane block to the candidate lane block, a candidate trajectory of the target vehicle is constructed. The step of determining the corresponding candidate lane block based on the future reference driving distance of the target vehicle and the cumulative length of the lane baseline from the target lane block to any drivable associated lane block includes: Take the target lane block as the current lane block and traverse the next lane block of the current lane block; Candidate determination: Determine whether the cumulative length of the lane baseline when traveling from the target lane block to the next lane block is greater than the future reference travel distance of the target vehicle. If yes, then the next lane block is taken as the candidate lane block. If no, then the next lane block is taken as the new current lane block, and the new next lane block is traversed. Continue performing the candidate determination to obtain at least one of the new candidate lane block and the new current lane block, until no new current lane block exists.

17. An electronic device, characterized in that, include: A processor and a memory, the memory for storing a computer program, the processor for calling and running the computer program stored in the memory to perform the vehicle trajectory prediction method according to any one of claims 1-15.

18. A computer-readable storage medium, characterized in that, Used to store a computer program that causes a computer to perform the vehicle trajectory prediction method as described in any one of claims 1-15.

Images