Trajectory prediction model construction method and device, and readable storage medium

Abstract

The application provides a trajectory prediction model construction method and device and a readable storage medium, relates to the field of data processing, and can improve the accuracy of the predicted trajectory. The trajectory prediction model construction method comprises the following steps: acquiring first backup trajectory data of a plurality of users in a first preset time period, each user in the plurality of users corresponding to one or more pieces of first backup trajectory data; determining a road section ID sequence corresponding to each piece of first backup trajectory data of the plurality of users according to a preconfigured corresponding relationship between trajectory data geographical position information and road section ID; respectively constructing an individual trajectory mode tree corresponding to each user in the plurality of users according to information of the road section ID sequence corresponding to one or more pieces of first backup trajectory data corresponding to each user in the plurality of users; and constructing a global trajectory mode tree according to information of the road section ID sequence corresponding to all first backup trajectory data corresponding to the plurality of users.

Description

Technical Field

[0001] This application relates to the field of data processing, and in particular to a method for constructing a trajectory prediction model, a trajectory prediction method, an apparatus, and a readable storage medium. Background Technology

[0002] In real life, it is often necessary to predict users' movement trajectories for purposes such as intercepting vehicles or clearing congested roads.

[0003] Current methods for predicting user trajectories typically involve gridding the user's original trajectory on a map during the model building phase, grouping users into groups, and constructing a corresponding model for each group. During the prediction phase, the model corresponding to the target user's group is selected to predict the trajectory. However, the spatial information expressed by the gridded trajectory is regional information, which is not accurate enough. Furthermore, for the target individual, the prediction results output by the grouped model are often not precise enough.

[0004] Therefore, the user trajectories predicted by the current solutions are often not accurate enough, which leads to the slow progress of related work. Summary of the Invention

[0005] This application provides a trajectory prediction model construction method, trajectory prediction method, apparatus, and readable storage medium to solve the problem that the predicted trajectory is not accurate enough.

[0006] To achieve the above objectives, the embodiments of this application adopt the following technical solutions:

[0007] Firstly, a method for constructing a trajectory prediction model is provided. This method includes: acquiring first backup trajectory data of multiple users within a first preset time period, where each user corresponds to one or more segments of first backup trajectory data, wherein each segment of first backup trajectory data is represented by trajectory data geographic location information; determining a road segment ID sequence corresponding to each segment of first backup trajectory data of the multiple users based on a pre-configured correspondence between trajectory data geographic location information and road segment IDs, wherein each road segment ID sequence includes one or more road segment IDs, and each road segment ID corresponds to a trajectory point in the first backup trajectory data; constructing an individual trajectory pattern tree corresponding to each user based on the information of the road segment ID sequence corresponding to one or more segments of first backup trajectory data of each user; and constructing a global trajectory pattern tree based on the information of the road segment ID sequence corresponding to all first backup trajectory data of the multiple users; wherein each road segment ID corresponds to a tree node in the individual trajectory pattern tree and / or the global trajectory pattern tree, and the individual trajectory pattern tree and / or the global trajectory pattern tree are used for trajectory prediction.

[0008] Based on the trajectory prediction model construction method provided in this application, spatial information of trajectory data can be represented by road segment IDs, which can more accurately express spatial information compared to map gridding methods. Furthermore, on the one hand, the individual trajectory pattern tree constructed based on individual data is more accurate for individual predictions; on the other hand, the global trajectory pattern tree constructed based on all data contains more comprehensive and complete trajectory patterns.

[0009] In one possible implementation, obtaining first backup trajectory data of multiple users within a first preset time period includes: obtaining original trajectory data of multiple users within the first preset time period; and segmenting the original trajectory data of each user based on the time information and / or spatial information corresponding to the original trajectory data of each user to obtain one or more segments of first backup trajectory data corresponding to each user.

[0010] Based on this scheme, the original trajectory data can be segmented to obtain backup trajectory data for constructing the trajectory pattern tree, which makes it easier to construct the trajectory pattern tree and makes the constructed trajectory pattern tree more accurate.

[0011] In one possible implementation, an individual trajectory pattern tree corresponding to each of the multiple users is constructed based on information of one or more road segment ID sequences corresponding to each user. This includes: constructing an individual trajectory pattern tree corresponding to each of the multiple users based on frequent pattern information and time pattern information corresponding to one or more road segment ID sequences corresponding to each user.

[0012] Based on this scheme, a trajectory pattern tree can be constructed based on a variety of information, so that when the trajectory pattern tree generates a predicted trajectory, more comprehensive information can be considered, making the generated predicted trajectory more accurate.

[0013] In one possible implementation, a global trajectory pattern tree is constructed based on the information of all road segment ID sequences corresponding to multiple users, including: constructing a global trajectory pattern tree based on the frequent pattern information and time pattern information corresponding to one or more road segment ID sequences corresponding to each of the multiple users.

[0014] Based on this scheme, a trajectory pattern tree can be constructed based on multiple different types of information, so that when the trajectory pattern tree generates a predicted trajectory, it can consider more comprehensive information and make the generated predicted trajectory more accurate.

[0015] In one possible implementation, the method further includes: acquiring second backup trajectory data of multiple users within a second preset time period, wherein each user corresponds to one or more segments of second backup trajectory data, and each segment of second backup trajectory data of the multiple users is characterized by trajectory data geographic location information; determining a sequence of road segment IDs corresponding to each segment of second backup trajectory data of the multiple users based on a pre-configured correspondence between trajectory data geographic location information and road segment IDs, wherein each road segment ID corresponds to a trajectory point in the second backup trajectory data; updating the individual trajectory pattern tree corresponding to each user based on the information of the road segment ID sequence corresponding to one or more segments of second backup trajectory data of the multiple users; and updating the global trajectory pattern tree based on the information of the road segment ID sequence corresponding to all second backup trajectory data of the multiple users.

[0016] This solution can continuously update the individual trajectory pattern tree and / or the global trajectory pattern tree, enabling incremental trajectory pattern mining and maintaining the integrity and effectiveness of the trajectory patterns.

[0017] In one possible implementation, the method further includes: acquiring backup trajectory data of the target user within a third preset time period; dividing the backup trajectory data of the target user into target trajectory data and adjacent trajectory data according to a preset duration; wherein, the target trajectory data is the trajectory data from the first trajectory point to the second trajectory point in the backup trajectory data of the target user, the first trajectory point is the last trajectory point in the backup trajectory data of the target user, and the time distance from the second trajectory point to the first trajectory point is the preset duration; the adjacent trajectory data is the backup trajectory data of the target user other than the target trajectory data; the target trajectory data is used for trajectory prediction of the target user; determining the road segment ID sequence corresponding to the adjacent trajectory data; and updating the individual trajectory pattern tree corresponding to the target user according to the road segment ID sequence corresponding to the adjacent trajectory data.

[0018] This solution can continuously update the individual trajectory pattern tree, enabling incremental trajectory pattern mining and maintaining the integrity and effectiveness of the trajectory patterns.

[0019] Secondly, a trajectory prediction method is provided. This method uses an individual trajectory pattern tree or a global trajectory pattern tree as described in the first aspect above for trajectory prediction. The method includes: obtaining the user identifier of the target user and the target trajectory data of the target user; determining the road segment ID sequence corresponding to the target trajectory data and the individual trajectory pattern tree corresponding to the target user based on the user identifier of the target user; if there is a trajectory pattern in the individual trajectory pattern tree corresponding to the target user whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to a preset threshold, determining and outputting the first trajectory prediction result of the target user based on the trajectory pattern whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to the preset threshold; or, if there is no trajectory pattern in the individual trajectory pattern tree corresponding to the target user whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to the preset threshold, determining and outputting the second trajectory prediction result of the target user based on the trajectory pattern with the highest matching degree with the road segment ID sequence corresponding to the target trajectory data in the global trajectory pattern tree.

[0020] Based on the trajectory prediction method provided in this application, when performing trajectory prediction, both the individual trajectory pattern tree and the global trajectory pattern tree can be considered. When there is a trajectory pattern in the individual trajectory pattern tree that is close to the user's historical action trajectory, trajectory prediction is performed by the individual trajectory pattern tree. When there is no trajectory pattern in the individual trajectory pattern tree that is close to the user's historical action trajectory, trajectory prediction is performed by the global trajectory pattern tree, making the prediction results more accurate.

[0021] In one possible implementation, acquiring the target trajectory data of the target user includes: acquiring the target user's backup trajectory data within a third preset time period; dividing the target user's backup trajectory data into target trajectory data and adjacent trajectory data according to a preset duration; wherein, the target trajectory data is the trajectory data from the first trajectory point to the second trajectory point in the target user's backup trajectory data, the first trajectory point is the last trajectory point in the target user's backup trajectory data, and the time distance from the second trajectory point to the first trajectory point is the preset duration; the adjacent trajectory data is the backup trajectory data of the target user other than the target trajectory data.

[0022] Based on this solution, the user's most recent backup trajectory data can be obtained as target trajectory data according to a preset time period, so that when the trajectory pattern tree performs trajectory prediction, the generated predicted trajectory is closer to the user's most recent historical action trajectory.

[0023] In one possible implementation, the method further includes: determining the road segment ID sequence corresponding to the adjacent trajectory data; and updating the individual trajectory pattern tree corresponding to the target user based on the road segment ID sequence corresponding to the adjacent trajectory data.

[0024] Based on this scheme, the updated trajectory pattern tree can be made closer to the user's recent historical action trajectory, and the prediction results determined based on the target trajectory can be more accurate.

[0025] In one possible implementation, determining and outputting the first trajectory prediction result for the target user based on trajectory patterns whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to a preset threshold includes: determining the next-hop target tree node of the first starting tree node based on the support degree of trajectory patterns between the first starting tree node and the next-hop tree node of the first starting tree node in one or more trajectory patterns with the last tree node of the first trajectory pattern as the first starting tree node; the first trajectory pattern is one of the trajectory patterns in the individual trajectory pattern tree corresponding to the target user whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to a preset threshold; determining the next-hop target tree node of the second starting tree node based on the support degree of trajectory patterns between the second starting tree node and the next-hop tree node of the second starting tree node in one or more trajectory patterns with the next-hop target tree node of the first starting tree node as the second starting tree node; and outputting the first trajectory prediction result for the target user after obtaining the first trajectory prediction result for the target user.

[0026] Based on this scheme, the next target tree node to be reached can be determined from the next hop node of the current node according to the support between the current node and the next hop node, thereby determining the prediction result step by step.

[0027] In one possible implementation, determining and outputting the second trajectory prediction result for the target user based on the trajectory pattern with the highest matching degree to the road segment ID sequence corresponding to the target trajectory data in the global trajectory pattern tree includes: determining the next-hop target tree node in the global trajectory pattern tree based on the support of trajectory patterns between the third starting tree node and its next-hop tree node in one or more trajectory patterns with the last tree node of the second trajectory pattern as the third starting tree node, and the support of trajectory patterns between the third starting tree node and its next-hop tree node in the individual trajectory pattern tree corresponding to the target user. The second trajectory pattern is the trajectory pattern with the highest matching degree with the road segment ID sequence corresponding to the target trajectory data in the global trajectory pattern tree. Based on the support of trajectory patterns between the fourth starting tree node and the next-hop tree node of the fourth starting tree node in one or more trajectory patterns with the next-hop target tree node of the third starting tree node as the fourth starting tree node, and the support of trajectory patterns between the fourth starting tree node and the next-hop tree node of the fourth starting tree node in the individual trajectory pattern tree corresponding to the target user, the next-hop target tree node of the fourth starting tree node is determined. The second trajectory prediction result of the target user is then output after obtaining the second trajectory prediction result of the target user.

[0028] Based on this scheme, the target tree node to be reached next can be determined from the next hop node of the current node according to the support between the current node and the next hop node, thereby determining the prediction result step by step.

[0029] In one possible implementation, the support is determined based on the frequent pattern information and time pattern information of the trajectory pattern.

[0030] Based on this scheme, the support of trajectory patterns can be determined according to multiple different types of information, thereby making the trajectory predicted based on the support of trajectory patterns more accurate.

[0031] Thirdly, this application provides a trajectory prediction model construction apparatus. The trajectory prediction model construction apparatus includes modules for performing the method described in the first aspect or any possible design of the first aspect.

[0032] Fourthly, this application provides a trajectory prediction apparatus. The trajectory prediction apparatus includes modules for performing the methods described in the second aspect or any possible design of the second aspect.

[0033] Fifthly, this application provides a trajectory prediction model building apparatus, which includes a memory and a processor. The memory and processor are coupled. The memory stores computer program code, which includes computer instructions. When the processor executes the computer instructions, the trajectory prediction model building apparatus performs the trajectory prediction model building method as described in the first aspect and any possible design thereof.

[0034] In a sixth aspect, this application provides a trajectory prediction apparatus, which includes a memory and a processor. The memory and processor are coupled. The memory stores computer program code, which includes computer instructions. When the processor executes the computer instructions, the trajectory prediction apparatus performs the trajectory prediction method as described in the second aspect and any possible design thereof.

[0035] In a seventh aspect, this application provides a chip system applied to a trajectory prediction model building apparatus; the chip system includes one or more interface circuits and one or more processors. The interface circuits and the processors are interconnected via lines; the interface circuits are used to receive signals from the memory of the trajectory prediction model building apparatus and send the signals to the processors, the signals including computer instructions stored in the memory. When the processor executes the computer instructions, the trajectory prediction model building apparatus performs the trajectory prediction model building method as described in the first aspect and any possible design of the first aspect.

[0036] Eighthly, this application provides a chip system applied to a trajectory prediction device; the chip system includes one or more interface circuits and one or more processors. The interface circuits and the processors are interconnected via lines; the interface circuits are used to receive signals from the memory of the trajectory prediction device and send the signals to the processors, the signals including computer instructions stored in the memory. When the processor executes the computer instructions, the trajectory prediction device performs the trajectory prediction method as described in the second aspect and any possible design of the application.

[0037] Ninthly, this application provides a computer-readable storage medium including computer instructions that, when executed on a trajectory prediction model building apparatus, cause the trajectory prediction model building apparatus to perform the trajectory prediction model building method as described in the first aspect and any possible design thereof.

[0038] In a tenth aspect, this application provides a computer-readable storage medium including computer instructions that, when executed on a trajectory prediction apparatus, cause the trajectory prediction model building apparatus to perform the trajectory prediction method as described in the second aspect and any possible design thereof.

[0039] Eleventhly, this application provides a computer program product including computer instructions that, when executed on a trajectory prediction model building apparatus, cause the trajectory prediction model building apparatus to perform the trajectory prediction model building method as described in the first aspect and any possible design thereof.

[0040] In a twelfth aspect, this application provides a computer program product comprising computer instructions that, when executed on a trajectory prediction device, cause the trajectory prediction device to perform the trajectory prediction method as described in the second aspect and any possible design thereof.

[0041] For a detailed description of aspects three through twelfth and their various implementations in this application, please refer to the detailed descriptions in aspects one and two and their various implementations; and for a detailed analysis of the beneficial effects of aspects three through twelfth and their various implementations in aspects one and two and their various implementations, please refer to the beneficial effect analyses in aspects one and two and their various implementations, which will not be repeated here.

[0042] These or other aspects of this application will become more readily apparent in the following description. Attached Figure Description

[0043] Figure 1This is a schematic diagram of the structure of a trajectory prediction system provided in an embodiment of this application;

[0044] Figure 2 A schematic diagram illustrating a trajectory prediction model construction method provided in an embodiment of this application;

[0045] Figure 3 A schematic diagram illustrating a trajectory prediction method provided in an embodiment of this application;

[0046] Figure 4 A hardware structure diagram of a computing device provided in an embodiment of this application;

[0047] Figure 5 A flowchart illustrating a trajectory prediction model construction method provided in this application embodiment;

[0048] Figure 6 This application provides an embodiment of a method for constructing an individual trajectory pattern tree. Figure 1 ;

[0049] Figure 7 This application provides an embodiment of a method for constructing an individual trajectory pattern tree. Figure 2 ;

[0050] Figure 8 This application provides an embodiment of a method for constructing an individual trajectory pattern tree. Figure 3 ;

[0051] Figure 9 This application provides an embodiment of a method for constructing a global trajectory pattern tree. Figure 1 ;

[0052] Figure 10 This application provides an embodiment of a method for constructing a global trajectory pattern tree. Figure 2 ;

[0053] Figure 11 A schematic diagram illustrating an updated individual trajectory pattern tree provided in an embodiment of this application;

[0054] Figure 12 A flowchart illustrating a trajectory prediction method provided in this application embodiment;

[0055] Figure 13 A schematic diagram of a trajectory prediction model construction device provided in an embodiment of this application;

[0056] Figure 14 This is a schematic diagram of the structure of a trajectory prediction device provided in an embodiment of this application. Detailed Implementation

[0057] The technical solutions of the embodiments of this application will be described below with reference to the accompanying drawings. In the description of this application, unless otherwise stated, " / " indicates that the objects before and after are in an "or" relationship. For example, A / B can represent A or B. "And / or" in this application is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone, where A and B can be singular or plural. Furthermore, in the description of this application, unless otherwise stated, "multiple" refers to two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple. Furthermore, to facilitate a clear description of the technical solutions in the embodiments of this application, the terms "first" and "second" are used in the embodiments of this application to distinguish identical or similar items with substantially the same function and effect. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and that "first" and "second" are not necessarily different. Meanwhile, in the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is being used as an example, illustration, or description. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of terms such as "exemplary" or "for example" is intended to present related concepts in a concrete manner for ease of understanding.

[0058] Furthermore, the network architecture and business scenarios described in the embodiments of this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided in the embodiments of this application. As those skilled in the art will know, with the evolution of network architecture and the emergence of new business scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0059] The embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0060] The trajectory prediction model construction method or trajectory prediction method provided in the embodiments of this application can be applied to trajectory prediction systems. Figure 1 One possible structure of this trajectory prediction system is shown. For example... Figure 1 As shown, the trajectory prediction system 10 provided in this application embodiment may include: a server 11, a client 12, and multiple acquisition devices 13.

[0061] Among them, such as Figure 2 As shown, when the trajectory prediction system applies the trajectory prediction model construction method, the acquisition device 13 can be used to collect user trajectory data. Multiple acquisition devices 13 can be respectively set in different users' mobile phones, or different users' vehicles, or other devices that can obtain user location data. The client 12 is a terminal device that the user can use. The client 12 can be mobile or fixed in location. Optionally, the acquisition device 13 can send the collected user trajectory data to the server 11 through the client 12. Alternatively, the acquisition device 13 can also directly send the collected user trajectory data to the server 11. The server 11 can construct an individual trajectory pattern tree and a global trajectory pattern tree based on the acquired user trajectory data. In one possible implementation, a certain acquisition device 13 can be set in a certain client 12, that is, some clients 12 can simultaneously have the function of obtaining user location data. This application embodiment does not specifically limit this.

[0062] like Figure 3 As shown, when the trajectory prediction system applies the trajectory prediction model construction method, the acquisition device 13 can send the collected trajectory data of the target user to the server 11 either through the client 12 or directly. The user can operate the client 12 to select the target user whose trajectory they want to predict. The client 12 can respond to the user's operation by generating a trajectory prediction request for the target user and sending it to the server 11 to request trajectory prediction. The server 11 can respond to the trajectory prediction request by determining the trajectory prediction result based on the individual trajectory pattern tree or the global trajectory pattern tree and the acquired target user's trajectory data. Optionally, the server 11 can output the trajectory prediction result to the client 12, and the client 12 can provide the trajectory prediction result to the user. In one possible implementation, the server 11 or the device for implementing the functions of the server 11 can be integrated into the client 12; in other words, in some cases, the client 12 can simultaneously have the function of responding to user operations and outputting predicted trajectories.

[0063] In this embodiment of the application, the client 12 can be any electronic device including a display module, such as a personal computer (PC), a laptop computer, a mobile device, a tablet computer, a laptop computer, etc. Figures 1-3 The client 12 in the example is a PC, which is merely illustrative. This application does not limit the specific form of the client 12.

[0064] The basic hardware structures of the server 11 and client 12 are similar, both including Figure 4 The components included in the computing device shown. The following are examples... Figure 4 Taking the computing device shown as an example, the hardware structure of server 11 and client 12 will be introduced.

[0065] like Figure 4 As shown, the computing device may include a processor 41, a memory 42, a communication interface 43, and a bus 44. The processor 41, the memory 42, and the communication interface 43 can be connected via the bus 44.

[0066] Processor 41 is the control center of the computing device. It can be a single processor or a collective term for multiple processing elements. For example, processor 41 can be a general-purpose central processing unit (CPU) or other general-purpose processors. Among them, the general-purpose processor can be a microprocessor or any conventional processor.

[0067] As one embodiment, processor 41 may include one or more CPUs, for example Figure 4 CPU 0 and CPU 1 are shown in the diagram.

[0068] The memory 42 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto.

[0069] In one possible implementation, the memory 42 can exist independently of the processor 41. The memory 42 can be connected to the processor 41 via a bus 44 and is used to store instructions or program code. When the processor 41 calls and executes the instructions or program code stored in the memory 42, it can implement the trajectory prediction method provided in the embodiments of this application.

[0070] In this embodiment, the software programs stored in the memory 42 are different for server 11 and client 12, and the functions implemented by server 11 and client 12 are different. The functions performed by each device will be described with reference to the following flowchart.

[0071] In another possible implementation, the memory 42 can also be integrated with the processor 41.

[0072] Communication interface 43 is used for connecting the computing device to other devices via a communication network, which may be Ethernet, radio access network (RAN), wireless local area network (WLAN), etc. Communication interface 43 may include a receiving unit for receiving data and a transmitting unit for sending data.

[0073] Bus 44 can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. This bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 4 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0074] It should be pointed out that, Figure 4 The structure shown does not constitute a limitation on the computing device, except Figure 4 In addition to the components shown, the computing device may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0075] The trajectory prediction model construction method or trajectory prediction method provided in this application embodiment is executed by a trajectory prediction generation device (hereinafter referred to as a trajectory prediction device). This trajectory prediction device can be the aforementioned server 11, the CPU within the aforementioned server 11, or the control module within the aforementioned server 11 used to generate the predicted trajectory. Of course, the trajectory prediction device can also be the aforementioned client 12, the CPU within the aforementioned client 12, or the control module within the aforementioned client 12 used to generate the predicted trajectory. This application embodiment does not limit the specific form of the trajectory prediction device.

[0076] The trajectory prediction model construction method and trajectory prediction method provided in the embodiments of this application are described below with reference to the accompanying drawings.

[0077] The trajectory prediction model construction method provided in this application embodiment can also be called the trajectory pattern mining stage, and the trajectory prediction method can also be called the trajectory prediction stage. The trajectory mining stage is used to construct a model (or trajectory pattern tree) for generating predicted trajectories. The trajectory prediction stage is used to generate predicted trajectories corresponding to the target trajectory based on the target user's target trajectory, using the constructed trajectory pattern tree.

[0078] In this embodiment of the application, the predicted trajectory corresponding to the target trajectory is the user's action trajectory after the target trajectory predicted by the trajectory pattern tree.

[0079] The following sections will elaborate on the trajectory pattern mining stage and the trajectory prediction stage.

[0080] In this embodiment of the application, the trajectory mining stage mainly includes the following steps:

[0081] S501, the trajectory prediction device acquires the first backup trajectory data of multiple users within a first preset time period. Each user corresponds to one or more segments of the first backup trajectory data. Each segment of the first backup trajectory data of the multiple users is represented by the geographical location information of the trajectory data.

[0082] S502. The trajectory prediction device determines the sequence of road segment IDs corresponding to each segment of first backup trajectory data of multiple users based on the pre-configured correspondence between the geographical location information of trajectory data and the road segment ID. Each road segment ID sequence includes one or more road segment IDs, and each road segment ID corresponds to a trajectory point in the first backup trajectory data.

[0083] S503. The trajectory prediction device constructs an individual trajectory pattern tree corresponding to each of the multiple users based on the information of the road segment ID sequence corresponding to one or more segments of first backup trajectory data for each user; and constructs a global trajectory pattern tree based on the information of the road segment ID sequence corresponding to all first backup trajectory data for the multiple users. Each road segment ID corresponds to a tree node in the individual trajectory pattern tree and / or the global trajectory pattern tree, which are used for trajectory prediction.

[0084] It should be noted that, in the embodiments of this application, the trajectory prediction device, when executing the method embodiments shown in S501-S503 above, can also be referred to as a trajectory prediction model construction device.

[0085] It should be noted that the names of various information, parameters, models or devices in the embodiments of this application are only example names, and may be other names in the specific implementation. The embodiments of this application do not make specific limitations on this.

[0086] In S501, in this embodiment of the application, the user's backup trajectory data can be used to construct a trajectory pattern tree for predicting the user's movement trajectory. The backup trajectory data acquired by the trajectory prediction device for constructing the initial trajectory pattern tree can be referred to as the first backup trajectory data. Optionally, the user's backup trajectory data can also be used to update the trajectory pattern tree; the backup trajectory data used to update the trajectory pattern tree can be referred to as the second backup trajectory data, which will be described in detail below.

[0087] In this embodiment, the user's backup trajectory data is represented by trajectory points. Each trajectory point includes geographical location information, such as latitude and longitude coordinates, and the time information at which the trajectory point was collected. Therefore, the user's current geographical location can be determined based on the trajectory points. It is understood that when multiple trajectory points are arranged chronologically according to their corresponding time information, these multiple trajectory points can be used to represent the user's (historical) movement trajectory within a corresponding time period (from the time corresponding to the first trajectory point to the time corresponding to the last trajectory point).

[0088] Optionally, in this embodiment of the application, the trajectory points can be global positioning system (GPS) data points.

[0089] In this embodiment, a user can correspond to one or more segments of backup trajectory data. A user's backup trajectory data can be understood as trajectory points of one or more users arranged in chronological order, which can be used to represent the user's historical action trajectory within the corresponding time period. Therefore, in order to predict the action trajectories of multiple users, the trajectory prediction device can acquire one or more segments of first backup trajectory data corresponding to multiple users within a first preset time period, and construct a trajectory pattern tree (an individual trajectory pattern tree corresponding to each user and a global trajectory pattern tree corresponding to all users) based on the acquired first backup trajectory data of multiple users to generate the user's predicted trajectory according to the user's historical action trajectory. The first preset time period can be determined according to business requirements.

[0090] Optionally, in this embodiment, the user's backup trajectory data can be obtained by processing the user's original trajectory data. The original trajectory data consists of trajectory points.

[0091] In this embodiment of the application, the trajectory prediction device can assign an identification information to the original trajectory data of each user. This identification information can also be called user identification (ID). The target identification is used to characterize the user to which the original trajectory data belongs.

[0092] In one possible implementation, in this embodiment of the application, the trajectory prediction device can obtain raw trajectory data of multiple users within a preset time period from a data acquisition platform used to query users' raw trajectory data, according to business needs. For example, GPS data of vehicles in city A within one month can be obtained from the data acquisition platform as raw trajectory data. Optionally, the data acquisition platform can run on a client or a server, and this embodiment of the application does not limit this.

[0093] Optionally, in this embodiment of the application, the trajectory prediction device can be connected to a data acquisition platform to obtain the user's original trajectory data from the data acquisition platform in real time.

[0094] In another possible implementation, the trajectory prediction device can be connected to the acquisition device to directly obtain the user's raw trajectory data collected by the acquisition device. The acquisition device can send the raw trajectory data to the trajectory prediction device after each time it collects the user's raw trajectory data, or it can periodically send all the raw trajectory data acquired within a preset time period to the trajectory prediction device. This application embodiment does not impose any limitations on this.

[0095] Furthermore, the trajectory prediction device can store the acquired raw trajectory data and / or the backup trajectory data obtained after processing the raw trajectory in the database of the trajectory prediction device for subsequent use.

[0096] The following section details how the trajectory prediction device processes the raw trajectory data to obtain backup trajectory data.

[0097] Optionally, in this embodiment, after the trajectory prediction device obtains the user's original trajectory data within a preset time period, it can segment the user's original trajectory data according to business needs, dividing the user's original trajectory data within the preset time period into one or more sub-trajectory data segments. The user's original trajectory data within the preset time period can be understood as a long segment of trajectory data within the preset time period. The preset time period can be determined according to business needs.

[0098] In this embodiment of the application, after the trajectory prediction device obtains one or more segments of sub-trajectory data based on the user's original trajectory data within a preset time period, each segment of sub-trajectory data can be used as a backup trajectory data segment for the user within the preset time period. Alternatively, the trajectory prediction device can further process each segment of sub-trajectory data to obtain a backup trajectory data segment for the user within the preset time period.

[0099] In one possible implementation, assuming a first preset time period and multiple users, the trajectory prediction device, after acquiring the original trajectory data of multiple users within the first preset time period, can segment the original trajectory data of each user based on the time and / or spatial information of the original trajectory data (the trajectory points constituting the original trajectory data), obtaining one or more segments of backup trajectory data corresponding to each user. The time information may include pre-configured time period information, and the spatial information may include pre-configured geographic nodes and / or dwell point information. For example, assuming the pre-configured time period is one day (time information), the original trajectory data is segmented into sub-trajectory data corresponding to each day based on the time information. Assuming the pre-configured geographic node is city A, the original trajectory data representing the geographic location of city A is segmented as sub-trajectory data, or the original trajectory data representing the geographic location other than A is segmented as sub-trajectory data. Assuming the pre-configured dwell point is passenger station B, duplicate data representing users dwelling at passenger station B in the original trajectory data is cleaned up. Here, "staying" refers to remaining at a specific point of residence within a certain timeframe. For example, if all of User 1's original trajectory data for a day corresponds to Bus Station B, then it is considered that User 1 stayed at Bus Station B on that day, and User 1's original trajectory data for that day can be cleared to a single trajectory point as sub-trajectory data.

[0100] Based on this scheme, the original trajectory data can be segmented to obtain backup trajectory data for constructing the trajectory pattern tree, which makes it easier to construct the trajectory pattern tree and makes the constructed trajectory pattern tree more accurate.

[0101] Optionally, in this embodiment, the trajectory prediction device can denoise the original trajectory data and remove or repair the original trajectory data with poor quality. Here, poor-quality trajectory data can be understood as unreasonable trajectory data. The trajectory prediction device can analyze and determine the poor-quality trajectory data based on a pre-configured algorithm or model, and then remove or repair it. For example, within a day, if the trajectory data at 12:00 corresponds to city A, the trajectory data at 12:05 corresponds to city B, and the trajectory data at 12:10 corresponds to city A, then the trajectory data at 12:05 can be considered of poor quality and can be removed.

[0102] Optionally, in this embodiment, the original trajectory data for denoising can be segmented sub-trajectory data. In other words, the trajectory prediction device can segment the original trajectory data and then denoise the resulting sub-trajectory data. The trajectory prediction device can use each segmented sub-trajectory data as a backup trajectory data segment.

[0103] In this embodiment, the trajectory prediction device processes the acquired raw trajectory data to obtain multiple segments of backup trajectory data that can be used to construct a trajectory pattern tree. The trajectory prediction device can assign an identification information, referred to as a trajectory ID, to each backup trajectory segment.

[0104] It is understandable that, since the backup trajectory data is obtained by processing the original trajectory data, the trajectory prediction device can assign a corresponding user ID of the original trajectory data to each segment of backup trajectory data, and can determine the corresponding user based on the user ID of the backup trajectory data.

[0105] For S502, in this embodiment of the application, in order to construct a trajectory pattern tree, after the trajectory prediction device obtains the first backup trajectory data of multiple users within a first preset time period, it can perform map matching on each segment of backup trajectory data, converting the spatial information represented by latitude and longitude coordinates in the backup trajectory data into road segment ID information on the map. The road segment ID is used to represent the corresponding road segment on the map. In other words, the trajectory prediction device can convert each trajectory point in the first backup trajectory data into a corresponding road segment ID. The mapping relationship between latitude and longitude coordinates and road segment IDs can be pre-configured.

[0106] It is understood that in this embodiment, because trajectory points correspond to time information, the converted road segment IDs also correspond to the same time information. Therefore, according to the chronological order, each segment of backup trajectory data can be converted into a sequence of road segment IDs. The road segment ID sequence includes one or more road segment IDs converted from trajectory points, and the chronological order of the road segment IDs in this sequence can be used to represent the temporal order. In other words, a road segment ID sequence can be used to represent a user's historical route over a period of time.

[0107] For example, a spare trajectory data for user 1 includes trajectory point 1 corresponding to 12:00, trajectory point 2 corresponding to 14:00, and trajectory point 3 corresponding to 16:00. After map matching, assuming that the road segment ID corresponding to trajectory point 1 is road segment A, the road segment ID corresponding to trajectory point 2 is road segment B, and the road segment ID corresponding to trajectory point 3 is road segment C, the road segment ID sequence corresponding to this spare trajectory data for user 1 can be {road segment A, road segment B, road segment C}. This road segment ID sequence can be used to represent user 1's journey from road segment A to road segment B to road segment C, where road segment A corresponds to 12:00, road segment B corresponds to 14:00, and road segment C corresponds to 16:00.

[0108] It is understood that, in the embodiments of this application, since each road segment ID sequence is converted from each corresponding backup trajectory data, each road segment ID sequence can be assigned the same identification information as each corresponding backup trajectory data, such as user ID and / or trajectory ID.

[0109] In this embodiment, after assigning identification information to each road segment ID sequence of multiple users within a preset time period, the trajectory prediction device can group and store the road segment ID sequences according to the identification information for later retrieval. For example, if a road segment ID sequence is assigned user ID: A and trajectory ID: 1, the trajectory prediction device can determine that the user ID corresponding to the road segment ID sequence is A and the trajectory ID is 1 based on the identification information of the road segment ID sequence. Assuming that user ID A represents user A and trajectory ID 1 represents the first backup trajectory in user A's backup trajectory, the trajectory prediction device can determine that the road segment ID sequence corresponds to user A and is the road segment ID sequence corresponding to user A's first backup trajectory based on the identification information of the road segment ID sequence.

[0110] For S503, in this embodiment of the application, after the trajectory prediction device determines the road segment ID sequence corresponding to one or more segments of first backup trajectory data for each user among multiple users within a first preset time period, it can, based on the determined road segment ID sequence, mine the trajectory pattern of an individual and construct an individual trajectory pattern tree (also known as an individual trajectory prediction model), and mine the trajectory pattern of the whole and construct a global trajectory pattern tree (also known as a global trajectory prediction model).

[0111] In this embodiment, the road segment IDs in the road segment ID sequence are applied to the nodes (or tree nodes) used to construct the trajectory pattern tree. The trajectory prediction device can construct the nodes in the trajectory pattern tree based on the road segment IDs in the road segment sequence. The node information includes the corresponding road segment ID information, and the road segment ID corresponding to a node can be determined based on the node information in the trajectory pattern tree. Furthermore, two different nodes can be connected, and the "edge" (or "route") formed after the nodes are connected, along with the attribute information of the "edge," can be called a trajectory pattern. It can be understood that a trajectory pattern can be used to represent a route from one road segment represented by a road segment ID to another road segment represented by a road segment ID. The trajectory pattern tree can be understood as a tree-like model composed of trajectory patterns.

[0112] The following uses an individual trajectory pattern tree as an example to illustrate how to construct a trajectory pattern tree in the embodiments of this application.

[0113] In one possible implementation, the trajectory prediction device can acquire all road segment ID sequences corresponding to the same user and use them to construct an individual trajectory pattern tree for that user. Since user IDs are used to represent different users, the trajectory prediction device can determine the user corresponding to a road segment ID sequence based on the user IDs in the sequence. The device can use each different road segment ID in the sequence as a node in the individual trajectory pattern tree and connect adjacent nodes to obtain the trajectory pattern. Based on the temporal order of the road segment IDs in the sequence, the device can determine the direction of the connections between different nodes, i.e., determine the temporal order of the trajectory patterns.

[0114] For example, the road segment ID sequence 1 is {road segment A, road segment B, road segment C}, where road segment A corresponds to 12:00, road segment B corresponds to 14:00, and road segment C corresponds to 16:00. Figure 6 As shown, when the trajectory prediction device constructs a trajectory pattern in the individual trajectory pattern tree based on this road segment ID sequence, it uses road segment A, road segment B, and road segment C as nodes. Based on the time information of road segment A, road segment B, and road segment C, it can connect the node corresponding to road segment A (which can be referred to as node A) to the node corresponding to road segment B (which can be referred to as node B), with the connection direction being from node A to node B. It can also connect node B to the node corresponding to road segment C (which can be referred to as node C), with the connection direction being from node B to node C.

[0115] In this embodiment of the application, when the trajectory prediction device constructs a trajectory pattern tree based on the road segment ID sequence, the starting road segment ID (the first road segment ID in the road segment ID sequence) corresponds to the starting point of the trajectory pattern. If different road segment ID sequences have the same starting road segment ID, then different road segment ID sequences can correspond to different branch routes corresponding to the same starting point. If different road segment ID sequences have different starting road segment IDs, then different road segment ID sequences can correspond to different routes corresponding to different starting points. For example, as shown... Figure 7The image shows the individual trajectory pattern tree corresponding to user 1. Assume that the road segment ID sequence corresponding to user 1 includes road segment ID sequence 1, road segment ID sequence 2, and road segment ID sequence 3. Road segment ID sequence 1 is {road segment A, road segment B, road segment C}, and road segment ID sequence 2 is {road segment A, road segment B, road segment D}. Then, the routes (or trajectory patterns) corresponding to road segment ID sequence 1 and road segment ID sequence 2 can both start from the node corresponding to road segment A (hereinafter referred to as node A). Road segment ID sequence 1 corresponds to a branch route (or branch trajectory pattern) from node A to the node corresponding to road segment B (hereinafter referred to as node B) to the node corresponding to road segment C (hereinafter referred to as node C), with the connecting direction being from node A to node B to node C. Road segment ID sequence 2 corresponds to a branch route from node A to node B to the node corresponding to road segment D (hereinafter referred to as node D), with the connecting direction being from node A to node B to node D. If the road segment ID sequence 3 is {road segment E, road segment F, road segment G}, then the trajectory pattern corresponding to road segment ID sequence 3 is from the node corresponding to road segment E (hereinafter referred to as node E) to the node corresponding to road segment F (hereinafter referred to as node F) and from F to the node corresponding to road segment G (hereinafter referred to as node G), with the connection direction being from node E to node F to node G.

[0116] It should be noted that, in the embodiments of this application, for road segment ID sequences with different starting road segment IDs, if there are identical road segment IDs in the road segment ID sequences, when constructing the trajectory pattern, the identical road segment IDs in the road segment ID sequences with different starting road segment IDs correspond to different nodes in the trajectory patterns with different starting points. For example, suppose user 1 corresponds to road segment ID sequence 1 {road segment A, road segment B, road segment C} and road segment ID sequence 2 {road segment B, road segment C, road segment A}. Because the starting road segment IDs of road segment ID sequence 1 and road segment ID sequence 2 are different, therefore, as Figure 8 As shown, when the trajectory prediction device constructs the individual trajectory pattern tree corresponding to user 1 based on road segment ID sequence 1 and road segment ID sequence 2, in the trajectory pattern corresponding to road segment ID sequence 1, the node corresponding to road segment A can be node A1, the node corresponding to road segment B can be node B1, and the node corresponding to road segment C can be node C1, with the connection direction being node A1 to node B1 to node C1. However, in the trajectory pattern corresponding to road segment ID sequence 2, the node corresponding to road segment B can be node B2, the node corresponding to road segment C can be node C2, and the node corresponding to road segment A can be node A2, with the connection direction being node B2 to node C2 to node A2. However, nodes A1 and A2 both correspond to road segment A, nodes B1 and B2 both correspond to road segment B, and nodes C1 and C2 both correspond to road segment C.

[0117] In this embodiment, after the trajectory prediction device traverses all road segment ID sequences corresponding to all first backup trajectory data for the same user (same user ID) (hereinafter referred to as the road segment ID sequence corresponding to the user), it can construct the individual trajectory pattern tree corresponding to that user. Similarly, after the trajectory prediction device traverses the road segment ID sequence corresponding to each of the multiple users, it can construct the individual trajectory pattern tree corresponding to each of the multiple users. For example, if there are n users, the trajectory prediction device can construct n individual trajectory pattern trees.

[0118] In this embodiment of the application, for constructing a global trajectory pattern tree, you can refer to the above description of how to construct an individual trajectory pattern tree. The difference is that the trajectory prediction device traverses all road segment ID sequences corresponding to multiple users to construct a global trajectory pattern tree.

[0119] The following example illustrates how to construct a global trajectory pattern tree. Assume the trajectory prediction device obtains the road segment ID sequences corresponding to user 1, user 2, and user 3. User 1 corresponds to road segment ID sequence 1 {road segment A, road segment B, road segment C} and road segment ID sequence 2 {road segment A, road segment B, road segment D, road segment E}; user 2 corresponds to road segment ID sequence 3 {road segment B, road segment C, road segment D, road segment E}; and user 3 corresponds to road segment ID sequence 4 {road segment B, road segment A, road segment C}. The global trajectory pattern tree constructed by the trajectory prediction device based on the road segment ID sequences corresponding to users 1, 2, and 3 is as follows: Figure 9 As shown, the trajectory patterns corresponding to road segment ID sequence 1 and road segment ID sequence 2 both start from node A1 corresponding to road segment A. Road segment ID sequence 1 corresponds to the branch trajectory pattern from node A1 to node B1 corresponding to road segment B, and then to node C1 corresponding to road segment C, with the connection direction being node A1 to node B1 to node C1. Road segment ID sequence 2 corresponds to the branch route from node A1 to node B1 to node D1 corresponding to road segment D, and then to node E1 corresponding to road segment E, with the connection direction being node A1 to node B1 to node D1 to node E1. The trajectory patterns corresponding to ID sequence 3 and road segment ID sequence 4 both start from node B2 corresponding to road segment B. For road segment ID sequence 3, the branch trajectory pattern from node B2 to node C2 corresponding to road segment C, to node D2 corresponding to road segment D, and to node E2 corresponding to road segment E has the connection direction from node B2 to node C2 to node D2 to node E2. For road segment ID sequence 4, the branch trajectory pattern from node B2 to node A2 corresponding to road segment A, and to node C3 corresponding to road segment C has the connection direction from node B2 to node A2 to node C3.

[0120] Optionally, the trajectory prediction device can determine frequent pattern information and time pattern information based on the information of the road segment IDs included in one or more road segment ID sequences corresponding to each user among multiple users, and record the determined frequent pattern information and time pattern information as attribute information of the edge (trajectory pattern) connecting two nodes to construct an individual trajectory pattern tree corresponding to each user, and / or construct a global trajectory pattern tree.

[0121] Based on this scheme, a trajectory pattern tree can be constructed based on multiple different types of information, so that when the trajectory pattern tree generates a predicted trajectory, it can consider more comprehensive information and make the generated predicted trajectory more accurate.

[0122] The frequent pattern information refers to the number of transitions between two identical nodes in a trajectory pattern originating from the same point in time. For example, if a trajectory pattern originating from node A occurs 3 times, then the frequent pattern information for the edge connecting node A and node B in that trajectory pattern is 3. It should be noted that for trajectory patterns originating from different points in time, even if multiple transitions occur between the same two road segment IDs, the corresponding frequent pattern information will not be increased. For example, in the above example... Figure 9 As shown in the example of constructing a global trajectory pattern tree based on road segment ID sequences 1, 2, 3, and 4, the trajectory patterns corresponding to road segment ID sequences 1 and 2 share the same starting point: node A. Furthermore, the {road segment A, road segment B} in both road segment ID sequences 1 and 2 can be converted into a trajectory pattern from node A1 to node B1. Therefore, the frequent pattern information of the edge connecting node A1 and node B1 in both road segment ID sequences 1 and 2 is 2. However, the starting points of the trajectory patterns corresponding to road segment ID sequences 1 and 3 are different. Therefore, even though {road segment B, road segment C} appears in both road segment ID sequences 1 and 3, the frequent pattern information of the edge connecting node B1 and node C1 in the trajectory pattern corresponding to road segment ID sequences 1 is not 2, but 1. Similarly, the frequent pattern information of the edge connecting node B2 and node C2 in the trajectory pattern corresponding to road segment ID sequences 3 is 1. And so on, in the above... Figure 9 As shown, in the global trajectory pattern tree constructed based on road segment ID sequence 1, road segment ID sequence 2, road segment ID sequence 3, and road segment ID sequence 4, the frequent pattern information of each trajectory pattern is as follows: Figure 10 As shown.

[0123] Time pattern information refers to the time information of the transfer between nodes. For example, in the trajectory pattern of the same starting point, if the time of arrival from node A to node B is 18:00, then the time pattern information of the edge connecting node A to node B is 18:00.

[0124] Furthermore, the trajectory prediction device can also determine sequence pattern information and / or start-end point pattern information, and record the determined sequence pattern information and / or start-end point pattern information as attribute information of the trajectory pattern to construct an individual trajectory pattern tree corresponding to each user, and / or construct a global trajectory pattern tree.

[0125] The sequence pattern information refers to the node order information before the current node transitions. For example, from node A to node B, where the current node is node B, node A can be reached from node C or node D. Therefore, the sequence pattern information of the edge connecting node A to node B could be from node C to node A to node B, or from node D to node A to node B. The sequence pattern information can take the starting point corresponding to the trajectory pattern as its starting point and continue to the current node. For example, if the starting point of the trajectory pattern corresponding to node A to node B is node E, node E can go to node C or node D, and both node C and node D go to node A, then the sequence pattern information of the edge connecting node A to node B could be from node E to node C to node A to node B, or from node E to node D to node A to node B.

[0126] The start-end point pattern information refers to the node information of the starting point and the node information of the ending point of the trajectory pattern to which the current node belongs. For example, for the trajectory pattern corresponding to node A to node B, the current node is B, the starting point of this trajectory pattern is node D, and the ending point is node C. Then the start-end point information is node C and node D.

[0127] Based on the trajectory prediction model construction method provided in this application, spatial information of trajectory data can be represented by road segment IDs, which can more accurately express spatial information compared to map gridding methods. Furthermore, on the one hand, the individual trajectory pattern tree constructed based on individual data is more accurate for individual predictions; on the other hand, the global trajectory pattern tree constructed based on all data contains more comprehensive and complete trajectory patterns.

[0128] Optionally, in this embodiment of the application, after the trajectory prediction device completes the construction of the individual trajectory pattern tree, it can assign a user identifier (e.g., the same user ID as the user ID of the road segment ID sequence used to construct the individual trajectory pattern tree) to the individual trajectory pattern tree based on the user ID of the road segment ID sequence used to construct the individual trajectory pattern tree. The user identifier is used to represent the user corresponding to the individual trajectory pattern tree.

[0129] Optionally, in this embodiment of the application, after the trajectory prediction device completes the construction of the global trajectory pattern tree, it can assign an identification information to the global trajectory pattern tree to characterize the trajectory pattern tree as a global trajectory pattern tree.

[0130] Optionally, in this embodiment, after constructing the individual trajectory pattern tree and the global trajectory pattern tree, the trajectory prediction device can update the individual trajectory pattern tree and / or the global trajectory pattern tree. In one possible implementation, the trajectory prediction device can acquire second backup trajectory data from multiple users within a second preset time period, where each user corresponds to one or more segments of second backup trajectory data. It then determines the road segment ID sequence corresponding to each segment of second backup trajectory data and updates the corresponding individual trajectory pattern tree and / or global trajectory pattern tree based on the determined road segment ID sequence information. The specific update process can be referred to the above description of constructing the individual trajectory pattern tree and the global trajectory pattern tree, and will not be repeated here. The second preset time period can be determined according to business requirements. Optionally, the second preset time period can be a preset time cycle; in other words, the trajectory prediction device can update the individual trajectory pattern tree and / or the global trajectory pattern tree periodically.

[0131] For example, assuming the above Figure 7 The individual trajectory pattern tree for user 1 shown is the initial individual trajectory pattern tree. The preset second time period is 1 day. At 12:00 noon each day, the trajectory prediction device can determine the road segment ID sequence corresponding to user 1 for that day based on the backup trajectory data acquired that day and update the individual trajectory pattern tree. If the road segment ID sequence corresponding to user 1 for that day is {road segment E, road segment H}, and in the trajectory pattern corresponding to this road segment ID sequence, the node corresponding to road segment H is node H, then for... Figure 7 After the initial individual trajectory pattern tree of User 1 is updated, the individual trajectory pattern tree of User 1 is as follows: Figure 11 As shown.

[0132] This solution can continuously update the individual trajectory pattern tree and / or the global trajectory pattern tree, enabling incremental trajectory pattern mining and maintaining the integrity and effectiveness of the trajectory patterns.

[0133] It should be noted that, in the embodiments of this application, the multiple users corresponding to the second backup trajectory data used to update the trajectory pattern tree may be the same as or different from the multiple users corresponding to the first backup trajectory data used to construct the initial trajectory pattern tree. The embodiments of this application do not impose any restrictions on this.

[0134] If the multiple users corresponding to the second backup trajectory data are the same as the multiple users corresponding to the first backup trajectory data used to construct the initial trajectory pattern tree, the trajectory prediction device can update the individual trajectory pattern tree corresponding to each of these multiple users based on the information of the road segment ID sequence corresponding to one or more segments of the second backup trajectory data for each of these multiple users.

[0135] Optionally, if the user corresponding to the second backup trajectory data acquired by the trajectory prediction device is a new user compared to the user corresponding to the first backup trajectory data, the trajectory prediction device can construct a new individual trajectory pattern tree corresponding to the new user based on the second backup trajectory data of the new user.

[0136] After the trajectory prediction device has completed the construction of individual trajectory pattern trees and global trajectory pattern trees, it can enter the trajectory prediction stage.

[0137] like Figure 12 As shown in the embodiments of this application, the trajectory prediction stage mainly includes the following steps:

[0138] S1201, The trajectory prediction device acquires the user identifier of the target user and the target trajectory data of the target user.

[0139] S1202. The trajectory prediction device determines the road segment ID sequence corresponding to the target trajectory data and the individual trajectory pattern tree corresponding to the target user based on the user identifier of the target user.

[0140] S1203. If there is a trajectory pattern in the individual trajectory pattern tree corresponding to the target user whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to a preset threshold, the trajectory prediction device determines and outputs the first trajectory prediction result of the target user based on the trajectory pattern whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to the preset threshold.

[0141] Alternatively, if there is no trajectory pattern in the individual trajectory pattern tree corresponding to the target user whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to a preset threshold, the trajectory prediction device determines and outputs the second trajectory prediction result of the target user based on the trajectory pattern with the highest matching degree with the road segment ID sequence corresponding to the target trajectory data in the global trajectory pattern tree.

[0142] For S1201, the trajectory prediction device can respond to the user's submitted request for trajectory prediction of the target user, determine the user identifier of the target user, and obtain the target trajectory data of the target user based on the user identifier. The user identifier of the target user can be the user ID described above.

[0143] Optionally, in this embodiment of the application, the trajectory prediction device or the client connected to the trajectory prediction device can respond to the user's operation, such as selecting a target user on the front-end interface and generating a trajectory prediction request. After obtaining the trajectory prediction request, the trajectory prediction device can determine the user identifier of the target user selected by the user, and then obtain the target trajectory data of the target user based on the user identifier of the target user.

[0144] In one possible implementation, the trajectory prediction device can acquire the target user's target trajectory data through the following steps.

[0145] The trajectory prediction device acquires backup trajectory data of the target user within a third preset time period.

[0146] The trajectory prediction device divides the target user's backup trajectory data into target trajectory data and adjacent trajectory data according to a preset time period. The target trajectory data is the trajectory data from the first trajectory point to the second trajectory point in the target user's backup trajectory data. The first trajectory point is the last trajectory point in the target user's backup trajectory data, and the time distance from the second trajectory point to the first trajectory point is the preset time period. The adjacent trajectory data is the backup trajectory data of the target user other than the target trajectory data.

[0147] Understandably, in order to predict a user's next movement trajectory based on the target user's target trajectory, the obtained target trajectory should be the most recent movement trajectory in time. Therefore, the backup trajectory data of the target user within the third preset time period acquired by the trajectory prediction device should be the most recent backup trajectory data of the target user in time that the trajectory prediction device can obtain. The duration of the third preset time period can be configured according to business needs. For example, the backup trajectory data of the target user within the third preset time period can be the backup trajectory data of the target user within the most recent 24 hours.

[0148] Optionally, the trajectory prediction device can query a database used to store backup trajectory data of the user to obtain the original trajectory data of the target user within a third preset time period.

[0149] Optionally, the trajectory prediction device can process the acquired original trajectory data of the target user within a third preset time period to obtain backup trajectory data of the target user within the third preset time period. In one possible implementation, the trajectory prediction device can query a data acquisition platform to obtain the original trajectory data of the target user within the third preset time period; alternatively, the trajectory prediction device can query a database used to store the user's original trajectory data to obtain the original trajectory data of the target user within the third preset time period. The specific processing procedure can be referred to the description of S501 above, and will not be repeated here.

[0150] After acquiring the backup trajectory data of the target user within a third preset time period, the trajectory prediction device can, based on the preset duration and the time information of the backup trajectory data, divide the data backward from the last trajectory point (first trajectory point) until a second trajectory point that meets the preset time distance. The divided trajectory point data is then used as the target trajectory data for the target user, and the remaining trajectory point data is used as nearby trajectory data. For example, if the preset duration is 1 hour and the third preset time period is the most recent 24 hours, the trajectory prediction device can use the backup trajectory data from the most recent hour as the target trajectory data, and the backup trajectory data from the remaining 23 hours as nearby trajectory data.

[0151] Based on this solution, the user's most recent backup trajectory data can be obtained as target trajectory data according to a preset time period, so that when the trajectory pattern tree performs trajectory prediction, the generated predicted trajectory is closer to the user's most recent historical action trajectory.

[0152] Optionally, after acquiring the target user's nearby trajectory data, the trajectory prediction device can update the target user's individual trajectory pattern tree based on this data. Specifically, the trajectory prediction device can determine the road segment ID sequence corresponding to the nearby trajectory data and update the target user's individual trajectory pattern tree accordingly. The specific update process can be found in the description of S503 above and will not be repeated here. This solution allows the target user's individual trajectory pattern tree to more closely resemble the target user's most recent movement trajectory, making the predicted trajectory data generated by the target user's individual trajectory pattern tree more accurate.

[0153] Optionally, the trajectory prediction device can update the global trajectory pattern tree based on the target user's nearby trajectory data. For details, please refer to the above description of how the trajectory prediction device updates the target user's individual trajectory pattern tree based on the target user's nearby trajectory data; it will not be repeated here.

[0154] It should be noted that, in this embodiment, after the trajectory prediction device acquires nearby trajectory data, it can update the trajectory pattern tree based on the nearby trajectory data before determining the prediction result based on the target trajectory. This update makes the trajectory pattern tree closer to the user's most recent historical movement trajectory, resulting in a more accurate prediction result determined based on the target trajectory. Alternatively, the trajectory pattern tree can be updated based on the nearby trajectory data after determining the prediction result based on the target trajectory, thus continuously updating the trajectory pattern tree. This embodiment does not impose any limitations on this approach.

[0155] Regarding S1202, in this embodiment of the application, the target trajectory data of the target user obtained by the trajectory prediction device can be referred to as a segment of target trajectory data of the target user. The trajectory prediction device can determine a corresponding road segment ID sequence based on the segment of target trajectory data. The specific determination process can be referred to the above description of S502, and will not be repeated here.

[0156] The trajectory prediction device can determine the individual trajectory pattern tree corresponding to the target user based on the user identifier of the target user.

[0157] Optionally, the trajectory prediction device can assign a user ID to the road segment ID sequence corresponding to the target trajectory data based on the user identifier of the determined target user.

[0158] For S1203, in this embodiment of the application, after the trajectory prediction device determines the road segment ID sequence corresponding to the target trajectory data, it matches the road segment ID sequence with all trajectory patterns in the individual trajectory pattern tree corresponding to the target user. If there is a trajectory pattern in the individual trajectory pattern tree corresponding to the target user whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to a preset threshold, then a first trajectory prediction result is generated from the individual trajectory pattern tree corresponding to the target user. If there is no trajectory pattern in the individual trajectory pattern tree corresponding to the target user whose matching degree with the input road segment ID sequence is greater than or equal to a pre-configured threshold, then the individual trajectory pattern tree is abandoned, and the road segment ID sequence corresponding to the target trajectory data is matched with all trajectory patterns in the global trajectory pattern tree to determine the trajectory pattern with the highest matching degree and generate a second trajectory prediction result.

[0159] Optionally, if there is no corresponding individual trajectory pattern tree for the target user, the trajectory prediction device directly matches the road segment ID sequence corresponding to the target trajectory data with the global trajectory pattern tree and generates a predicted trajectory.

[0160] In this embodiment, the matching degree between the trajectory pattern and the road segment ID sequence is determined based on the road segment IDs included in the road segment ID sequence and the nodes included in the trajectory pattern. The trajectory prediction device can pre-configure an algorithm or model to calculate the matching degree between the trajectory pattern and the road segment ID sequence.

[0161] For example, the road segment ID sequence corresponding to the target trajectory data is {road segment A, road segment B, road segment C}. The trajectory pattern with the highest matching degree in the individual trajectory pattern tree corresponding to the target user is from node A corresponding to road segment A to node B corresponding to road segment B. The calculated matching degree is 66.7%, but it does not meet the pre-configured threshold of 80%. Therefore, the road segment ID sequence is matched with the global trajectory pattern tree again.

[0162] In this embodiment of the application, for a trajectory pattern in the individual trajectory pattern tree corresponding to the target user where the matching degree of the road segment ID sequence corresponding to the target trajectory data is greater than or equal to a preset threshold, the trajectory prediction device determines and outputs the first trajectory prediction result in the following possible way:

[0163] The trajectory prediction device determines the next-hop target tree node of the first starting tree node based on the support between the first starting tree node and the next-hop tree node of the first starting tree node in one or more trajectory patterns in the individual trajectory pattern tree corresponding to the target user, with the last tree node of the first trajectory pattern as the first starting tree node. The first trajectory pattern is one of the trajectory patterns in the individual trajectory pattern tree corresponding to the target user whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to a preset threshold.

[0164] Optionally, the trajectory prediction device can determine the next-hop tree node in the trajectory pattern with the highest support between the first starting tree node and the next-hop tree node of the first starting tree node as the next-hop target tree node.

[0165] Optionally, the first trajectory mode can be the trajectory mode with the highest matching degree among the trajectory modes that have a matching degree with the road segment ID sequence corresponding to the target trajectory data that is greater than or equal to a preset threshold.

[0166] Then, the trajectory prediction device determines the next-hop target tree node of the second starting tree node based on the support between the second starting tree node and the next-hop target tree node of the second starting tree node in one or more trajectory patterns in the individual trajectory pattern tree corresponding to the target user, with the next-hop target tree node of the first starting tree node as the second starting tree node, until the first trajectory prediction result of the target user is obtained and then outputs the first trajectory prediction result of the target user.

[0167] Optionally, the trajectory prediction device can determine the next-hop tree node in the trajectory pattern with the highest support between the second starting tree node and the next-hop tree node of the second starting tree node as the next-hop target tree node.

[0168] For example, the road segment ID sequence corresponding to the target trajectory data is {road segment A, road segment B, road segment C}. In the individual trajectory pattern tree corresponding to the target user, among the trajectory patterns with a matching degree that meets a preset threshold, the trajectory pattern with the highest matching degree is the path from node A corresponding to road segment A to node C corresponding to road segment B. In this trajectory pattern, node B's next-hop node is either node D corresponding to road segment D or node E corresponding to road segment E. The support of the trajectory pattern from node B to node D and the support of the trajectory pattern from node B to node E are calculated. If the support of the trajectory pattern from node B to node D is higher, then the trajectory pattern from node B to node D is determined as the next predicted trajectory after node A to node B. Then, the support of the trajectory patterns corresponding to node C and all nodes corresponding to the next hop of node C in this trajectory pattern is calculated until the final predicted trajectory is generated, and the final predicted trajectory is output as the first prediction result.

[0169] In one possible implementation, in this embodiment of the application, the trajectory prediction device can calculate the support of the trajectory pattern based on the frequent pattern information and time pattern information corresponding to the trajectory pattern. Further, the trajectory prediction device can also calculate the support of the trajectory pattern based on sequence pattern information and / or start-end point pattern information. For details, please refer to the above introduction on constructing an individual trajectory pattern tree. Based on this scheme, the support of the trajectory pattern can be determined based on multiple different types of information, thereby making the trajectory predicted based on the support of the trajectory pattern more accurate.

[0170] Optionally, if there is no trajectory pattern in the individual trajectory pattern tree corresponding to the target user whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to a preset threshold, the trajectory prediction device can determine the predicted trajectory based solely on the global trajectory pattern tree. For details, please refer to the above introduction on generating predicted trajectories based on individual trajectory pattern trees.

[0171] Optionally, if there is no trajectory pattern in the individual trajectory pattern tree corresponding to the target user whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to a preset threshold, the trajectory prediction device can combine the individual trajectory pattern tree corresponding to the target user with the trajectory pattern tree when determining the predicted trajectory based on the global trajectory pattern tree.

[0172] In one possible implementation, when the trajectory prediction device determines the predicted trajectory based on the global trajectory pattern tree, if it encounters a branch route and needs to select a route, it can calculate the support of the branch trajectory pattern in the global trajectory pattern tree and simultaneously calculate the support of the same branch trajectory pattern in the individual trajectory pattern tree corresponding to the target user. This allows it to calculate the mixed support of the branch trajectory patterns and determine the branch trajectory pattern with higher mixed support as the predicted trajectory based on the mixed support of the branch trajectory patterns.

[0173] In some cases, route selection is required when the difference in support between different branch trajectory patterns corresponding to the same node is less than a certain threshold, or when there is no restriction on the difference in support between different branch trajectory patterns corresponding to the same node, and route selection is performed for all of them.

[0174] In this embodiment, the mixed support of a trajectory pattern can be the result of a weighted calculation of the support of that trajectory pattern in the global trajectory pattern tree and the support of that trajectory pattern in the individual trajectory pattern tree corresponding to the target user. The weights of the support of the trajectory pattern in the global trajectory pattern tree and the support of the trajectory pattern in the individual trajectory pattern tree can be pre-configured.

[0175] Specifically, the trajectory prediction device can determine the next-hop target tree node of the third starting tree node in the global trajectory pattern tree based on the support between the third starting tree node and the next-hop tree node of the third starting tree node in one or more trajectory patterns with the last tree node of the second trajectory pattern as the third starting tree node, and the support between the third starting tree node and the next-hop tree node of the third starting tree node in the individual trajectory pattern tree corresponding to the target user. The second trajectory pattern is the trajectory pattern with the highest matching degree with the road segment ID sequence corresponding to the target trajectory data in the global trajectory pattern tree.

[0176] Optionally, the trajectory prediction device can determine the next-hop target tree node as the next-hop target tree node in the trajectory pattern with the highest mixed support between the third starting tree node and the next-hop tree node of the third starting tree node in the global trajectory pattern tree.

[0177] Then, the trajectory prediction device determines the next-hop target tree node of the fourth starting tree node based on the support of trajectory patterns between the fourth starting tree node and the next-hop tree node of the fourth starting tree node in one or more trajectory patterns in the global trajectory pattern tree with the next-hop target tree node of the third starting tree node as the fourth starting tree node, and the support of trajectory patterns between the fourth starting tree node and the next-hop tree node of the fourth starting tree node in the individual trajectory pattern tree corresponding to the target user, until the second trajectory prediction result of the target user is obtained and then outputs the second trajectory prediction result of the target user.

[0178] Optionally, the trajectory prediction device can determine the next-hop target tree node as the trajectory pattern with the highest mixed support between the fourth starting tree node and the next-hop tree node of the fourth starting tree node in the global trajectory pattern tree.

[0179] For example, suppose the road segment ID sequence corresponding to the target trajectory data is {road segment A, road segment B, road segment C}. The trajectory pattern with the highest matching degree matched in the global trajectory pattern tree is from node A corresponding to road segment A to node B corresponding to road segment B to node C corresponding to road segment C. Then node C is the last tree node of the second trajectory pattern. The next hop tree nodes of node C are node D corresponding to road segment D and node E corresponding to road segment E. In other words, there are different branch trajectory patterns from node C to node D and from node C to node E, and the calculated support is 60% for both. Then, in the individual trajectory pattern tree corresponding to the target user, the support from the node corresponding to road segment C to the node corresponding to road segment D, and from the node corresponding to road segment C to the node corresponding to road segment E are calculated. In this individual trajectory pattern tree, the support from node C to node D of road segment C is 70%, and the support from node C to node E of road segment E is 50%. Then, according to a 1:1 weight, the mixed support of the trajectory pattern from node C to node D in the global trajectory pattern tree is calculated to be higher than the mixed support of the trajectory pattern from node C to node E. Node C to node D is used as the predicted trajectory after node A to node B to node C, and so on, until the final predicted trajectory is generated and the final predicted trajectory is output as the second prediction result.

[0180] In this embodiment of the application, when generating a predicted trajectory, the individual trajectory pattern tree or the global trajectory pattern tree can complete the current trajectory prediction and generate the final predicted trajectory when the trajectory generation termination condition is met.

[0181] Optionally, in this embodiment, the trajectory generation termination condition may include the generated predicted trajectory having a duration that meets a preset duration, or determining that the generated predicted trajectory has reached the end of the trajectory pattern and has no next-hop node. The duration of the generated predicted trajectory can be determined based on the time pattern information of the trajectory pattern corresponding to the predicted trajectory.

[0182] Based on the trajectory prediction method provided in this application, when performing trajectory prediction, both the individual trajectory pattern tree and the global trajectory pattern tree can be considered. When there is a trajectory pattern in the individual trajectory pattern tree that is close to the user's historical action trajectory, trajectory prediction is performed by the individual trajectory pattern tree. When there is no trajectory pattern in the individual trajectory pattern tree that is close to the user's historical action trajectory, trajectory prediction is performed by the global trajectory pattern tree, making the prediction results more accurate.

[0183] Optionally, after obtaining the prediction results, the trajectory prediction device can output the prediction results to the client so that the client can provide the prediction results to the user, for example, by displaying the prediction results on the front-end interface, or by generating the prediction results in file form and providing them to the user.

[0184] The foregoing mainly describes the solutions provided by the embodiments of this application from a methodological perspective. To achieve the above functions, it includes corresponding hardware structures and / or software modules for executing each function. Those skilled in the art should readily recognize that, based on the units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0185] like Figure 13 The diagram shown is a structural schematic of a trajectory prediction model construction device provided in an embodiment of this application. The trajectory prediction model construction device 130 is used to execute... Figure 5 The trajectory prediction model construction method shown is illustrated. The trajectory prediction model construction apparatus 120 may include an acquisition module 131, a matching module 132, and a construction module 133. Of course, the trajectory prediction model construction apparatus provided in this application embodiment includes, but is not limited to, the above-described modules.

[0186] The acquisition module 131 is used to acquire first backup trajectory data of multiple users within a first preset time period. Each user corresponds to one or more segments of first backup trajectory data, and each segment of first backup trajectory data of multiple users is represented by trajectory data geographic location information. The matching module 132 is used to determine the road segment ID sequence corresponding to each segment of first backup trajectory data of multiple users according to the pre-configured correspondence between trajectory data geographic location information and road segment ID. Each road segment ID sequence includes one or more road segment ID information, and each road segment ID information corresponds to a trajectory point in the first backup trajectory data. The construction module 123 is used to construct an individual trajectory pattern tree corresponding to each user based on the information of the road segment ID sequence corresponding to one or more segments of first backup trajectory data of each user; and to construct a global trajectory pattern tree based on the information of the road segment ID sequence corresponding to all first backup trajectory data of multiple users. Each road segment ID information corresponds to a tree node in the individual trajectory pattern tree and / or the global trajectory pattern tree, and the individual trajectory pattern tree and / or the global trajectory pattern tree are used for trajectory prediction.

[0187] In one possible implementation, the acquisition module 131 is specifically used to: acquire the original trajectory data of multiple users within a first preset time period; and, based on the time information and / or spatial information corresponding to the original trajectory data of each user among the multiple users, segment the original trajectory data of each user among the multiple users to acquire one or more segments of first backup trajectory data corresponding to each user among the multiple users.

[0188] In another possible implementation, the construction module 133 is specifically used to: construct individual trajectory pattern trees corresponding to each user among the multiple users based on the frequent pattern information and time pattern information corresponding to one or more road segment ID sequences corresponding to each user among the multiple users; and construct a global trajectory pattern tree based on the frequent pattern information and time pattern information corresponding to one or more road segment ID sequences corresponding to each user among the multiple users.

[0189] In another possible implementation, the acquisition module 131 is further configured to acquire second backup trajectory data of multiple users within a second preset time period, wherein each user corresponds to one or more segments of second backup trajectory data, and each segment of second backup trajectory data of multiple users is represented by trajectory data geographic location information; the matching module 132 is further configured to determine the road segment ID sequence corresponding to each segment of second backup trajectory data of multiple users according to the pre-configured correspondence between trajectory data geographic location information and road segment ID, wherein each road segment ID corresponds to a trajectory point in the second backup trajectory data; the construction module 133 is further configured to update the individual trajectory pattern tree corresponding to each user of multiple users according to the information of the road segment ID sequence corresponding to one or more segments of second backup trajectory data of each user; and update the global trajectory pattern tree according to the information of the road segment ID sequence corresponding to all second backup trajectory data of multiple users.

[0190] In another possible implementation, the acquisition module 131 is further configured to acquire the backup trajectory data of the target user within a third preset time period; according to a preset duration, the backup trajectory data of the target user is divided into target trajectory data and adjacent trajectory data; wherein, the target trajectory data is the trajectory data from the first trajectory point to the second trajectory point in the target user's backup trajectory data, the first trajectory point is the last trajectory point in the target user's backup trajectory data, and the time distance from the second trajectory point to the first trajectory point is the preset duration; the adjacent trajectory data is the backup trajectory data of the target user excluding the target trajectory data; the target trajectory data is used for trajectory prediction of the target user. The matching module 132 is further configured to determine the road segment ID sequence corresponding to the adjacent trajectory data. The construction module 133 is further configured to update the individual trajectory pattern tree corresponding to the target user according to the road segment ID sequence corresponding to the adjacent trajectory data.

[0191] Of course, the trajectory prediction model construction device 130 provided in this application embodiment includes, but is not limited to, the above-described modules.

[0192] In actual implementation, the acquisition module 131, the matching module 132, and the construction module 133 can be derived by... Figure 4 The processor 41 shown calls the program code in memory 42 to implement this. For the specific execution process, please refer to... Figure 5 The description of the trajectory prediction model construction method shown is omitted here.

[0193] like Figure 14 The diagram shown is a structural schematic of a trajectory prediction device provided in an embodiment of this application. The trajectory prediction device 140 is used to perform... Figure 11 The trajectory prediction method is shown. The trajectory prediction device 140 may include an acquisition module 141 and a determination module 142. Of course, the trajectory prediction device provided in this application embodiment includes, but is not limited to, the above-described modules.

[0194] The acquisition module 141 is used to acquire the user identifier of the target user and the target trajectory data of the target user. The determination module 142 is used to determine the road segment ID sequence corresponding to the target trajectory data and the individual trajectory pattern tree corresponding to the target user based on the user identifier of the target user. If there is a trajectory pattern in the individual trajectory pattern tree corresponding to the target user whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to a preset threshold, the determination module 142 is also used to determine and output the first trajectory prediction result of the target user based on the trajectory pattern with the matching degree with the road segment ID sequence corresponding to the target trajectory data that is greater than or equal to the preset threshold; or, if there is no trajectory pattern in the individual trajectory pattern tree corresponding to the target user whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to the preset threshold, the determination module 142 is also used to determine and output the second trajectory prediction result of the target user based on the trajectory pattern with the highest matching degree with the road segment ID sequence corresponding to the target trajectory data in the global trajectory pattern tree.

[0195] In one possible implementation, the acquisition module 141 is specifically used to: acquire backup trajectory data of the target user within a third preset time period; divide the backup trajectory data of the target user into target trajectory data and adjacent trajectory data according to a preset duration; wherein, the target trajectory data is the trajectory data from the first trajectory point to the second trajectory point in the backup trajectory data of the target user, the first trajectory point is the last trajectory point in the backup trajectory data of the target user, and the time distance from the second trajectory point to the first trajectory point is the preset duration; the adjacent trajectory data is the backup trajectory data of the target user other than the target trajectory data.

[0196] In another possible implementation, the determining module 142 is also used to determine the road segment ID sequence corresponding to the adjacent trajectory data; and update the individual trajectory pattern tree corresponding to the target user based on the road segment ID sequence corresponding to the adjacent trajectory data.

[0197] In another possible implementation, the determining module 142 is specifically used to: determine the next-hop target tree node of the first starting tree node based on the support of trajectory patterns between the first starting tree node and the next-hop tree node of the first starting tree node in one or more trajectory patterns with the last tree node of the first trajectory pattern as the first starting tree node; the first trajectory pattern is one of the trajectory patterns in the individual trajectory pattern tree corresponding to the target user whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to a preset threshold; determine the next-hop target tree node of the second starting tree node based on the support of trajectory patterns between the second starting tree node and the next-hop tree node of the second starting tree node in one or more trajectory patterns with the next-hop target tree node of the first starting tree node as the second starting tree node; and output the first trajectory prediction result of the target user after obtaining the first trajectory prediction result of the target user.

[0198] In another possible implementation, the determining module 142 is specifically used to: determine the next-hop target tree node of the third starting tree node in the global trajectory pattern tree based on the support of trajectory patterns between the third starting tree node and the next-hop tree node of the third starting tree node in one or more trajectory patterns with the last tree node of the second trajectory pattern as the third starting tree node, and the support of trajectory patterns between the third starting tree node and the next-hop tree node of the third starting tree node in the individual trajectory pattern tree corresponding to the target user, wherein the second trajectory pattern is the trajectory pattern with the highest matching degree with the road segment ID sequence corresponding to the target trajectory data in the global trajectory pattern tree; determine the next-hop target tree node of the fourth starting tree node based on the support of trajectory patterns between the fourth starting tree node and the next-hop tree node of the fourth starting tree node in one or more trajectory patterns with the next-hop target tree node of the third starting tree node as the fourth starting tree node, and the support of trajectory patterns between the fourth starting tree node and the next-hop tree node of the fourth starting tree node in the individual trajectory pattern tree corresponding to the target user, until the second trajectory prediction result of the target user is obtained and then output the second trajectory prediction result of the target user.

[0199] In another possible implementation, the support is determined based on the frequent pattern information and time pattern information of the trajectory pattern.

[0200] Of course, the trajectory prediction device 140 provided in this application embodiment includes, but is not limited to, the modules described above.

[0201] In actual implementation, the acquisition module 141 and the determination module 142 can be generated by... Figure 4 The processor 41 shown calls the program code in memory 42 to implement this. For the specific execution process, please refer to... Figure 11 The description of the trajectory prediction method shown is omitted here.

[0202] Another embodiment of this application provides a computer-readable storage medium storing computer instructions that, when executed on a trajectory prediction model building device, cause the trajectory prediction model building device to perform each step of the method flow shown in the above method embodiment.

[0203] Another embodiment of this application provides a computer-readable storage medium storing computer instructions that, when executed on a trajectory prediction device, cause the trajectory prediction device to perform each step of the method flow shown in the above method embodiment.

[0204] Another embodiment of this application provides a chip system applied to a trajectory prediction model building apparatus. The chip system includes one or more interface circuits and one or more processors. The interface circuits and processors are interconnected via lines. The interface circuits are used to receive signals from the memory of the trajectory prediction model building apparatus and send the signals to the processors, the signals including computer instructions stored in the memory. When the processor executes the computer instructions, the trajectory prediction model building apparatus performs each step of the method flow shown in the above method embodiment.

[0205] Another embodiment of this application provides a chip system applied to a trajectory prediction device. The chip system includes one or more interface circuits and one or more processors. The interface circuits and processors are interconnected via lines. The interface circuits are used to receive signals from the memory of the trajectory prediction device and send the signals to the processors, the signals including computer instructions stored in the memory. When the processor executes the computer instructions, the trajectory prediction device performs each step of the trajectory prediction device in the method flow shown in the above method embodiment.

[0206] In another embodiment of this application, a computer program product is also provided, which includes instructions that, when executed on a trajectory prediction device, cause the trajectory prediction device to perform the various steps of the trajectory prediction device in the method flow shown in the above method embodiment.

[0207] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented using software programs, implementation can be, in whole or in part, in the form of a computer program product. This computer program product includes one or more computer instructions. When these computer instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is 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, 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 containing one or more servers, data centers, etc., that can be integrated with the medium. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state disks, SSDs).

[0208] The above description is merely a specific embodiment of this application. Any variations or substitutions conceived by those skilled in the art based on the specific embodiments provided in this application should be covered within the protection scope of this application.

Claims

1. A method for constructing a trajectory prediction model, characterized in that, The method includes: Acquire first backup trajectory data of multiple users within a first preset time period, wherein each user corresponds to one or more segments of first backup trajectory data, and each segment of first backup trajectory data of the multiple users is represented by trajectory data geographic location information. Based on the pre-configured correspondence between the geographic location information of the trajectory data and the road segment identifier ID, a road segment ID sequence corresponding to each segment of the first backup trajectory data of the multiple users is determined, wherein each road segment ID sequence includes one or more road segment ID information, and each road segment ID information corresponds to a trajectory point in the first backup trajectory data; Based on the information of the road segment ID sequence corresponding to one or more segments of first backup trajectory data for each of the multiple users, an individual trajectory pattern tree corresponding to each of the multiple users is constructed; and based on the information of the road segment ID sequence corresponding to all the first backup trajectory data for the multiple users, a global trajectory pattern tree is constructed; the trajectory pattern tree is a tree-like model composed of trajectory patterns, wherein the trajectory pattern is used to represent the route from a road segment represented by one road segment ID to a road segment represented by another road segment ID; each road segment ID information corresponds to a tree node of the individual trajectory pattern tree and / or the global trajectory pattern tree, and the connection direction of the tree nodes is determined based on the time order of the road segment ID information in the road segment ID sequence; The individual trajectory pattern tree and / or the global trajectory pattern tree are used to determine the trajectory pattern corresponding to the target trajectory data based on the matching degree between the target user's target trajectory data and themselves, and to predict the trajectory of the target user based on the trajectory pattern corresponding to the target trajectory data.

2. The method according to claim 1, characterized in that, The step of obtaining the first backup trajectory data of multiple users within a first preset time period includes: Obtain the original trajectory data of multiple users within the first preset time period; Based on the time and / or spatial information corresponding to the original trajectory data of each of the multiple users, the original trajectory data of each of the multiple users is segmented to obtain one or more segments of first backup trajectory data corresponding to each of the multiple users.

3. The method according to claim 1 or 2, characterized in that, The step of constructing an individual trajectory pattern tree corresponding to each of the multiple users based on information from one or more road segment ID sequences corresponding to each user includes: Based on the frequent pattern information and time pattern information corresponding to one or more road segment ID sequences for each of the multiple users, an individual trajectory pattern tree corresponding to each of the multiple users is constructed.

4. The method according to claim 1 or 2, characterized in that, The step of constructing a global trajectory pattern tree based on the information of all road segment ID sequences corresponding to the multiple users includes: A global trajectory pattern tree is constructed based on the frequent pattern information and time pattern information corresponding to one or more road segment ID sequences for each of the multiple users.

5. The method according to claim 1 or 2, characterized in that, The method further includes: Acquire the second backup trajectory data of the multiple users within a second preset time period. Each of the multiple users corresponds to one or more segments of the second backup trajectory data. Each segment of the second backup trajectory data of the multiple users is represented by the geographical location information of the trajectory data. Based on the pre-configured correspondence between the geographic location information of the trajectory data and the road segment ID, a sequence of road segment IDs corresponding to each segment of the second backup trajectory data of the multiple users is determined, wherein each road segment ID corresponds to a trajectory point in the second backup trajectory data; Based on the information of the road segment ID sequence corresponding to one or more segments of second backup trajectory data for each of the multiple users, the individual trajectory pattern tree corresponding to each of the multiple users is updated respectively; and based on the information of the road segment ID sequence corresponding to all the second backup trajectory data of the multiple users, the global trajectory pattern tree is updated.

6. The method according to claim 1 or 2, characterized in that, The method further includes: Obtain backup trajectory data of the target user within the third preset time period; According to a preset duration, the backup trajectory data of the target user is divided into target trajectory data and adjacent trajectory data; wherein, the target trajectory data is the trajectory data from the first trajectory point to the second trajectory point in the target user's backup trajectory data, the first trajectory point is the last trajectory point in the target user's backup trajectory data, and the time distance from the second trajectory point to the first trajectory point is the preset duration; the adjacent trajectory data is the backup trajectory data of the target user other than the target trajectory data; the target trajectory data is used for trajectory prediction of the target user; Determine the road segment ID sequence corresponding to the adjacent trajectory data; Update the individual trajectory pattern tree corresponding to the target user based on the road segment ID sequence corresponding to the adjacent trajectory data.

7. A trajectory prediction method, characterized in that, The trajectory prediction method uses the individual trajectory pattern tree or the global trajectory pattern tree as described in any one of claims 1-6 to perform trajectory prediction; the trajectory prediction method includes: Obtain the user identifier of the target user and the target trajectory data of the target user; Based on the user identifier of the target user, determine the road segment ID sequence corresponding to the target trajectory data and the individual trajectory pattern tree corresponding to the target user; If there is a trajectory pattern in the individual trajectory pattern tree corresponding to the target user whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to a preset threshold, the first trajectory prediction result of the target user is determined and output based on the trajectory pattern whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to the preset threshold. Alternatively, if there is no trajectory pattern in the individual trajectory pattern tree corresponding to the target user whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to a preset threshold, the second trajectory prediction result of the target user is determined and output according to the trajectory pattern with the highest matching degree with the road segment ID sequence corresponding to the target trajectory data in the global trajectory pattern tree.

8. The method according to claim 7, characterized in that, The acquisition of the target user's target trajectory data includes: Obtain backup trajectory data of the target user within a third preset time period; According to a preset duration, the backup trajectory data of the target user is divided into target trajectory data and adjacent trajectory data; wherein, the target trajectory data is the trajectory data from the first trajectory point to the second trajectory point in the backup trajectory data of the target user, the first trajectory point is the last trajectory point in the backup trajectory data of the target user, and the time distance from the second trajectory point to the first trajectory point is the preset duration; the adjacent trajectory data is the backup trajectory data of the target user other than the target trajectory data.

9. The method according to claim 8, characterized in that, The method further includes: Determine the road segment ID sequence corresponding to the adjacent trajectory data; Update the individual trajectory pattern tree corresponding to the target user based on the road segment ID sequence corresponding to the adjacent trajectory data.

10. The method according to claim 7 or 8, characterized in that, The step of determining and outputting the first trajectory prediction result of the target user based on trajectory patterns whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to a preset threshold includes: Based on the support of trajectory patterns between the first starting tree node and the next hop tree node of the first starting tree node in one or more trajectory patterns with the last tree node of the first trajectory pattern as the first starting tree node, the next hop target tree node of the first starting tree node is determined. The first trajectory pattern is one of the trajectory patterns in the individual trajectory pattern tree corresponding to the target user whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to a preset threshold. Based on the support of trajectory patterns between the second starting tree node and its next-hop tree node in one or more trajectory patterns where the next-hop target tree node of the first starting tree node is the second starting tree node, the next-hop target tree node of the second starting tree node is determined, and the first trajectory prediction result of the target user is output after obtaining the first trajectory prediction result of the target user.

11. The method according to claim 7 or 8, characterized in that, The step of determining and outputting the second trajectory prediction result for the target user based on the trajectory pattern with the highest matching degree to the road segment ID sequence corresponding to the target trajectory data in the global trajectory pattern tree includes: Based on the support of trajectory patterns between the third starting tree node and its next-hop tree node in one or more trajectory patterns with the last tree node of the second trajectory pattern as the third starting tree node, and the support of trajectory patterns between the third starting tree node and its next-hop tree node in the individual trajectory pattern tree corresponding to the target user, the next-hop target tree node of the third starting tree node in the global trajectory pattern tree is determined, and the second trajectory pattern is the trajectory pattern with the highest matching degree with the road segment ID sequence corresponding to the target trajectory data in the global trajectory pattern tree; Based on the support of trajectory patterns between the fourth starting tree node and its next-hop tree node in one or more trajectory patterns where the next-hop target tree node of the third starting tree node is the fourth starting tree node, and the support of trajectory patterns between the fourth starting tree node and its next-hop tree node in the individual trajectory pattern tree corresponding to the target user, the next-hop target tree node of the fourth starting tree node is determined, and the second trajectory prediction result of the target user is output after obtaining the second trajectory prediction result of the target user.

12. The method according to claim 10, characterized in that, The support is determined based on the frequent pattern information and time pattern information of the trajectory pattern.

13. A trajectory prediction model construction device, characterized in that, The device includes: an acquisition module, a matching module, and a construction module; The acquisition module is used to acquire first backup trajectory data of multiple users within a first preset time period. Each user in the multiple users corresponds to one or more segments of first backup trajectory data. Each segment of first backup trajectory data of the multiple users is represented by trajectory data geographical location information. The matching module is used to determine the road segment ID sequence corresponding to each segment of the first backup trajectory data of the multiple users according to the pre-configured correspondence between the geographical location information of trajectory data and the road segment ID. Each road segment ID sequence includes one or more road segment ID information, and each road segment ID information corresponds to a trajectory point in the first backup trajectory data. The construction module is used to construct an individual trajectory pattern tree corresponding to each of the multiple users based on the information of the road segment ID sequence corresponding to one or more segments of first backup trajectory data for each user; and to construct a global trajectory pattern tree based on the information of the road segment ID sequence corresponding to all the first backup trajectory data for the multiple users. The trajectory pattern tree is a tree-like model composed of trajectory patterns, which represent the route from one road segment ID to another road segment ID. Each road segment ID corresponds to a tree node in the individual trajectory pattern tree and / or the global trajectory pattern tree, and the direction of the connection between the tree nodes is determined based on the time order of the road segment ID information in the road segment ID sequence. The individual trajectory pattern tree and / or the global trajectory pattern tree are used to determine the trajectory pattern corresponding to the target trajectory data based on the matching degree between the target user's target trajectory data and itself, and to predict the trajectory of the target user based on the trajectory pattern corresponding to the target trajectory data.

14. The apparatus according to claim 13, characterized in that, The acquisition module is specifically used to: acquire the original trajectory data of multiple users within the first preset time period; and, based on the time information and / or spatial information corresponding to the original trajectory data of each user among the multiple users, segment the original trajectory data of each user among the multiple users to acquire one or more segments of first backup trajectory data corresponding to each user among the multiple users. The construction module is specifically used to: construct an individual trajectory pattern tree corresponding to each of the multiple users based on the frequent pattern information and time pattern information corresponding to one or more road segment ID sequences corresponding to each user; A global trajectory pattern tree is constructed based on the frequent pattern information and time pattern information corresponding to one or more road segment ID sequences for each of the multiple users. The acquisition module is further configured to acquire second backup trajectory data of the multiple users within a second preset time period, wherein each user corresponds to one or more segments of second backup trajectory data, and each segment of second backup trajectory data of the multiple users is characterized by trajectory data geographic location information. The matching module is further configured to determine the sequence of road segment IDs corresponding to each segment of the second backup trajectory data of the multiple users based on the correspondence between the pre-configured trajectory data geographic location information and road segment IDs, wherein each road segment ID corresponds to a trajectory point in the second backup trajectory data. The construction module is further configured to update the individual trajectory pattern tree corresponding to each of the multiple users based on the information of the road segment ID sequence corresponding to one or more segments of second backup trajectory data for each user; and to update the global trajectory pattern tree based on the information of the road segment ID sequence corresponding to all the second backup trajectory data of the multiple users. The acquisition module is further configured to acquire backup trajectory data of the target user within a third preset time period; and to divide the backup trajectory data of the target user into target trajectory data and adjacent trajectory data according to a preset duration; wherein, the target trajectory data is the trajectory data from the first trajectory point to the second trajectory point in the backup trajectory data of the target user, the first trajectory point is the last trajectory point in the backup trajectory data of the target user, and the time distance from the second trajectory point to the first trajectory point is the preset duration; the adjacent trajectory data is the backup trajectory data of the target user other than the target trajectory data; the target trajectory data is used for trajectory prediction of the target user; The matching module is also used to determine the road segment ID sequence corresponding to the adjacent trajectory data; The construction module is also used to update the individual trajectory pattern tree corresponding to the target user based on the road segment ID sequence corresponding to the adjacent trajectory data.

15. A trajectory prediction device, characterized in that, The trajectory prediction device uses an individual trajectory pattern tree or a global trajectory pattern tree as described in any one of claims 1-6 to perform trajectory prediction; the trajectory prediction device includes: an acquisition module and a determination module; The acquisition module is used to acquire the user identifier of the target user and the target trajectory data of the target user; The determining module is used to determine the road segment ID sequence corresponding to the target trajectory data and the individual trajectory pattern tree corresponding to the target user based on the user identifier of the target user; If there is a trajectory pattern in the individual trajectory pattern tree corresponding to the target user whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to a preset threshold, the determining module is further used to determine and output the first trajectory prediction result of the target user based on the trajectory pattern whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to the preset threshold. Alternatively, if there is no trajectory pattern in the individual trajectory pattern tree corresponding to the target user that has a matching degree greater than or equal to a preset threshold with the road segment ID sequence corresponding to the target trajectory data, the determining module is further configured to determine and output the second trajectory prediction result of the target user based on the trajectory pattern in the global trajectory pattern tree that has the highest matching degree with the road segment ID sequence corresponding to the target trajectory data.

16. The apparatus according to claim 15, characterized in that, The acquisition module is specifically used for: acquiring backup trajectory data of the target user within a third preset time period; dividing the backup trajectory data of the target user into target trajectory data and adjacent trajectory data according to a preset duration; wherein, the target trajectory data is the trajectory data from the first trajectory point to the second trajectory point in the backup trajectory data of the target user, the first trajectory point is the last trajectory point in the backup trajectory data of the target user, and the time distance from the second trajectory point to the first trajectory point is the preset duration; the adjacent trajectory data is the backup trajectory data of the target user other than the target trajectory data. The determining module is further configured to determine the road segment ID sequence corresponding to the adjacent trajectory data; and update the individual trajectory pattern tree corresponding to the target user based on the road segment ID sequence corresponding to the adjacent trajectory data. The determining module is specifically configured to: determine the next-hop target tree node of the first starting tree node based on the support of trajectory patterns between the first starting tree node and the next-hop tree node of the first starting tree node in one or more trajectory patterns with the last tree node of the first trajectory pattern as the first starting tree node, wherein the first trajectory pattern is one of the trajectory patterns in the individual trajectory pattern tree corresponding to the target user whose matching degree with the road segment ID sequence corresponding to the target trajectory data is greater than or equal to a preset threshold; determine the next-hop target tree node of the second starting tree node based on the support of trajectory patterns between the second starting tree node and the next-hop tree node of the second starting tree node in one or more trajectory patterns with the next-hop target tree node of the first starting tree node as the second starting tree node, until the first trajectory prediction result of the target user is obtained, and then output the first trajectory prediction result of the target user; The determining module is specifically used to: determine the next-hop target tree node of the third starting tree node in the global trajectory pattern tree based on the support of trajectory patterns between the third starting tree node and the next-hop tree node of the third starting tree node in one or more trajectory patterns with the last tree node of the second trajectory pattern as the third starting tree node, and the support of trajectory patterns between the third starting tree node and the next-hop tree node of the third starting tree node in the individual trajectory pattern tree corresponding to the target user, wherein the second trajectory pattern is the trajectory pattern with the highest matching degree with the road segment ID sequence corresponding to the target trajectory data in the global trajectory pattern tree; determine the next-hop target tree node of the fourth starting tree node based on the support of trajectory patterns between the fourth starting tree node and the next-hop tree node of the fourth starting tree node in one or more trajectory patterns with the next-hop target tree node of the third starting tree node as the fourth starting tree node, and the support of trajectory patterns between the fourth starting tree node and the next-hop tree node of the fourth starting tree node in the individual trajectory pattern tree corresponding to the target user, until the second trajectory prediction result of the target user is obtained, and then output the second trajectory prediction result of the target user; The support is determined based on the frequent pattern information and time pattern information of the trajectory pattern.

17. A trajectory prediction model construction device, characterized in that, The trajectory prediction model building apparatus includes a memory and a processor; the memory and the processor are coupled; the memory is used to store computer program code, the computer program code including computer instructions; when the processor executes the computer instructions, the trajectory prediction model building apparatus performs the trajectory prediction model building method as described in any one of claims 1-6.

18. A trajectory prediction device, characterized in that, The trajectory prediction device includes a memory and a processor; the memory and the processor are coupled; the memory is used to store computer program code, the computer program code including computer instructions; when the processor executes the computer instructions, the trajectory prediction device performs the trajectory prediction method as described in any one of claims 7-12.

19. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes computer instructions that, when executed on the trajectory prediction device, cause the trajectory prediction device to perform the trajectory prediction model construction method as described in any one of claims 1-6.

20. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes computer instructions that, when executed on the trajectory prediction device, cause the trajectory prediction device to perform the trajectory prediction method as described in any one of claims 7-12.

Images