Positioning method and device based on signaling, electronic equipment, medium and program product

By constructing spatiotemporal coding vectors and a preset trajectory reconstruction model, and utilizing Points of Interest (POIs) and time-related signaling in the signaling sequence, the problem of low accuracy in user location positioning was solved, and accurate prediction of user location and activity trajectory was achieved.

CN119031473APending Publication Date: 2024-11-26CHINA MOBILE INFORMATION TECHNOLOGY CO LTD +1
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411303554.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-18
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

In existing technologies, user location positioning based on mobile phone signaling data suffers from low accuracy, failing to accurately determine the user's actual location and activity trajectory.

Method used

By acquiring the user's signaling sequence, constructing a spatiotemporal coding vector, using a preset trajectory reconstruction model, and combining POIs and time-related signaling in the signaling sequence, the user's actual target location is mapped, thereby improving positioning accuracy.

Benefits of technology

It achieves precise location tracking of users, enabling more accurate prediction of user activity trajectories around the signaling time, thus improving the accuracy of user location tracking.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119031473A_ABST
    Figure CN119031473A_ABST
Patent Text Reader

Abstract

The invention discloses a signaling-based positioning method and device, electronic equipment, a medium and a program product, and relates to the technical field of position and big data, a signaling sequence of a first signaling of a user to be positioned is obtained, the signaling sequence comprises a plurality of sub-signaling which are the first signaling and the second signaling respectively, and the first signaling and the second signaling are different from each other. And a time associated signaling following a first signaling time of the first signaling; according to the obtained POI point of interest of each sub-signaling and the obtained POI point of interest of each sub-signaling, respectively determining a space-time coding vector of each sub-signaling; and inputting a space-time coding sequence constructed by each space-time coding vector into a preset trajectory reconstruction model to obtain a target actual position of the to-be-positioned user in the first signaling mapped by the space-time coding sequence. The technical problem of low user position positioning accuracy is solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The embodiments of the present application relate to the field of location and big data technology, and in particular to a positioning method, device, electronic device, medium and program product based on signaling. Background Art

[0002] Mobile phone signaling data records the connection interaction relationship and timestamp information between the mobile terminal carried by the user and the communication base station, and also includes the latitude and longitude position of the base station. Through mobile phone signaling data, the user's spatiotemporal trajectory can be extracted to determine the user's location. For example, the location of the base station to which the mobile terminal is connected can be directly used as the user's location.

[0003] However, since each base station usually covers a certain spatial service range, directly using the location of the base station to which the mobile terminal is connected as the user's location will result in a certain spatial error, which will lead to inaccurate user location positioning.

[0004] The above contents are only used to assist in understanding the technical solutions of the embodiments of the present application, and do not constitute an admission that the above contents are prior art. Summary of the invention

[0005] The main purpose of the embodiments of the present application is to provide a signaling-based positioning method, device, electronic device, medium and program product, aiming to solve the technical problem of low user position positioning accuracy.

[0006] To achieve the above purpose, the present application provides a positioning method, including:

[0007] Acquire a signaling sequence of a first signaling of a user to be located, wherein the signaling sequence includes a plurality of sub-signals, which are respectively the first signaling and a time-associated signaling that is time-continuous with the first signaling of the first signaling;

[0008] According to the obtained POI (Point of Information) of each sub-signal

[0009] The interest point and each of the sub-signals respectively determine a respective spatiotemporal coding vector of each of the sub-signals;

[0010] The space-time coding sequence constructed by each of the space-time coding vectors is input into a preset trajectory reconstruction model to obtain the target actual position of the user to be located in the first signaling mapped by the space-time coding sequence.

[0011] In one embodiment, the step of determining the spatiotemporal coding vector of each sub-signaling according to the acquired POI points of interest of each sub-signaling and each sub-signaling comprises:

[0012] For each of the sub-signals, obtaining each POI point of interest of the sub-base station in the sub-signal;

[0013] Determine a spatial coding vector of the sub-signaling according to the sub-signaling and the corresponding POI points of interest;

[0014] Determining a time coding vector of the sub-signaling according to the sub-signaling time of the sub-signaling;

[0015] The spatial coding vector and the temporal coding vector of the same sub-signaling are concatenated to obtain the spatiotemporal coding vector of the sub-signaling.

[0016] In one embodiment, the step of determining the spatial coding vector of the sub-signaling according to the sub-signaling and the corresponding POI points of interest includes:

[0017] Performing a nonlinear change on the sub-signaling to obtain a signaling vector, and performing a nonlinear change on each of the POI points of interest to obtain an interest vector corresponding to each of the POI points of interest;

[0018] Determine the signaling query parameters of the signaling vector according to a preset attention algorithm, and determine the interest key and interest value corresponding to each of the interest vectors;

[0019] Calculating the attention weights between the signaling query parameters and each of the interest keys;

[0020] According to the attention weights of the interest keys, weighted summation is performed on the interest values ​​of the interest vectors where the interest keys are located, to obtain the interest point convergence score of the sub-signaling;

[0021] The longitude vector and the latitude vector of the sub-signaling are determined from the signaling vector, and the longitude vector, the latitude vector and the interest point convergence score are concatenated to obtain a spatial coding vector.

[0022] In one embodiment, the sub-signaling time includes a start time and an end time;

[0023] The step of determining the time coding vector of the sub-signaling according to the sub-signaling time of the sub-signaling comprises:

[0024] Calculate the duration of the sub-signaling according to the start time and the end time;

[0025] Performing nonlinear change on the duration to obtain a time difference vector;

[0026] Performing nonlinear transformation on the start time and the end time respectively to obtain a first time semantic vector of the start time and a second time semantic vector of the end time;

[0027] The time difference vector, the first time semantic vector and the second time semantic vector are concatenated to obtain a time coding vector.

[0028] In one embodiment, the method comprises:

[0029] Acquire a training sample set, wherein the training sample set includes a plurality of training samples, each of the training samples includes a training signaling sequence, a training POI and an actual location tag, and each training sub-signaling included in the training signaling sequence is a signaling to be trained, and a training associated signaling that is time-continuous with the training signaling of the signaling to be trained;

[0030] Determining, according to each of the training sub-signals and the respective training POIs, respective training spatiotemporal coding vectors of each of the training sub-signals;

[0031] Inputting the spatiotemporal coding sequence constructed by each of the training spatiotemporal coding vectors into a preset model to be trained to obtain a first prediction result, wherein the first prediction result includes a predicted position of the signaling to be trained;

[0032] Calculating a prediction loss according to the first prediction result and the actual position label;

[0033] The model to be trained is optimized according to the prediction loss to obtain a preset trajectory reconstruction model that has been trained.

[0034] In one embodiment, the step of obtaining a training sample set includes:

[0035] Acquire each signaling to be trained, each precise positioning data and each training POI to be screened, wherein the training POI to be screened includes the POI position, and the signaling to be trained includes the training base station position;

[0036] For each of the signalings to be trained, in each of the training POIs to be screened, a target POI position whose first spatial distance between the POI position and the training base station position is less than a preset distance threshold is screened out, and the signaling to be trained is associated with the training POI of the target POI position;

[0037] According to each of the signalings to be trained and each of the precise positioning data, the actual location tag corresponding to each of the signalings to be trained and the training-related signaling corresponding to each of the signalings to be trained are determined.

[0038] In one embodiment, the training-associated signaling includes first-associated signaling and second-associated signaling;

[0039] The step of determining, according to each of the signaling to be trained and each of the precise positioning data, the actual location tag corresponding to each of the signaling to be trained and the training associated signaling corresponding to each of the signaling to be trained comprises:

[0040] Determine the training signaling of each user and the precise positioning data of each user belonging to the same user identifier according to the respective user identifiers in the signaling to be trained and the respective user identifiers in the precise positioning data;

[0041] Determine a target positioning mapping relationship in each of the user training signaling and each of the user precise positioning data, wherein the target positioning mapping relationship is a target user training signaling and a target user precise positioning data that are temporally and spatially associated;

[0042] Using the target user precise positioning data of the target user training signaling in the target positioning mapping relationship as the actual position label of the target user training signaling;

[0043] For any target positioning mapping relationship, according to the user signaling time of the target user training signaling in the target positioning mapping relationship, among each user training signaling corresponding to the user identifier of the target user training signaling, filter out the first associated signaling whose training signaling time is ranked before the user signaling time and the number thereof is a preset number, and filter out the second associated signaling whose training signaling time is ranked after the user signaling time and the number thereof is a preset number.

[0044] In one embodiment, the step of determining the target positioning mapping relationship in each of the user training signaling and each of the user precise positioning data includes:

[0045] Determine the respective user signaling time from the respective user training signaling, and determine the respective sampling time from the respective user precise positioning data, wherein the user signaling time includes the user signaling start time and the user signaling end time;

[0046] Associating the initial user precise positioning data and the initial user training signaling at the sampling time within the user signaling time to obtain an initial positioning mapping relationship;

[0047] Respectively calculating the distance between the initial user training signaling and the initial user precise positioning data in each of the initial positioning mapping relationships to obtain a second spatial distance corresponding to each of the initial positioning mapping relationships;

[0048] A target positioning mapping relationship whose second spatial distance is less than a preset distance threshold is selected from each of the initial positioning mapping relationships.

[0049] In one embodiment, the step of optimizing the model to be trained according to the prediction loss to obtain a preset trajectory reconstruction model that has been trained includes:

[0050] Optimizing the model to be trained according to the prediction loss until the prediction loss is less than a preset training error threshold, thereby obtaining an initial trajectory reconstruction model;

[0051] Verifying the initial trajectory reconstruction model according to the obtained verification set to obtain a verification position output by the initial trajectory reconstruction model;

[0052] If the verification position does not meet the preset optimal effect conditions, the hyperparameters of the model to be trained are updated, and the model to be trained is retrained until the verification position meets the preset optimal effect conditions to obtain a preset trajectory reconstruction model.

[0053] In addition, to achieve the above-mentioned purpose, an embodiment of the present application provides a positioning device, the device comprising:

[0054] An acquisition module 10 is used to acquire a signaling sequence of a first signaling of a user to be located, wherein the signaling sequence includes a plurality of sub-signals, which are respectively the first signaling and a time-associated signaling that is time-continuous with the first signaling of the first signaling;

[0055] The coding module 20 is used to determine the spatiotemporal coding vector of each sub-signaling according to the acquired POI points of interest of each sub-signaling and each sub-signaling;

[0056] The positioning prediction module 30 is used to input the space-time coding sequence constructed by each of the space-time coding vectors into a preset trajectory reconstruction model to obtain the target actual position of the user to be located in the first signaling mapped by the space-time coding sequence.

[0057] In addition, to achieve the above-mentioned purpose, an embodiment of the present application also provides an electronic device, which includes: a memory, a processor, and a program of the positioning method stored in the memory and executable on the processor. When the program of the positioning method is executed by the processor, the steps of the positioning method as described above can be implemented.

[0058] In addition, to achieve the above-mentioned purpose, an embodiment of the present application also provides a computer-readable storage medium, on which a program for implementing the positioning method is stored. When the program of the positioning method is executed by a processor, the steps of the positioning method as described above are implemented.

[0059] In addition, to achieve the above-mentioned purpose, an embodiment of the present application also provides a computer program product, including a computer program, and when the computer program is executed by a processor, the steps of the positioning method as described above are implemented.

[0060] One or more technical solutions proposed in the embodiments of the present application have at least the following technical effects: the embodiments of the present application obtain a signaling sequence of a first signaling of a user to be located, wherein the multiple sub-signals in the first signaling sequence are respectively the first signaling and a time-associated signaling that is time-continuous with the first signaling of the first signaling, and then determine the respective spatiotemporal coding vectors of the sub-signals according to the respective POI points of interest of the obtained sub-signals and the respective sub-signals, and the spatiotemporal coding vectors can reflect the distribution of the surrounding POI points of interest corresponding to the sub-signals, so that the spatiotemporal coding sequence obtained after the splicing of the spatiotemporal coding vectors can comprehensively reflect the distribution of all the sub-signals. The distribution of POI points of interest around each signaling, that is, the space-time coding sequence can reflect the distribution of surrounding POI points of interest corresponding to the first signaling, and at the same time can reflect the distribution of surrounding POI points of interest corresponding to the time-related signaling, and because the time-related signaling is continuous with the first signaling time of the first signaling, the space-time coding sequence can also reflect the activity trajectory of the user to be located near the first signaling time of the first signaling, and then through the space-time coding sequence and the preset trajectory reconstruction model, the preset trajectory reconstruction model is used to predict the target actual position of the user to be located mapped by the space-time coding sequence in the first signaling.

[0061] Furthermore, the distribution of POI points of interest around the first signaling reflected by the space-time coding sequence, the distribution of surrounding POI points of interest corresponding to the time-related signaling, and the activity trajectory of the user to be located near the first signaling time of the first signaling are realized, and the actual target position of the user to be located corresponding to the first signaling is accurately predicted, instead of directly taking the base station position in the first signaling as the actual position of the user to be located corresponding to the first signaling, thereby improving the accuracy of signaling-based user positioning. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The drawings herein are incorporated into and constitute a part of the specification, illustrate embodiments consistent with the embodiments of the present application, and together with the description are used to explain the principles of the embodiments of the present application.

[0063] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0064] Figure 1 A schematic diagram of a flow chart of an embodiment of the positioning method of the present application;

[0065] Figure 2 A schematic diagram of a flow chart of another embodiment of the positioning method according to an embodiment of the present application;

[0066] Figure 3 This is a flow chart of an example of a positioning method in an embodiment of the present application;

[0067] Figure 4 A schematic diagram of another exemplary process of the positioning method in an embodiment of the present application;

[0068] Figure 5 A schematic diagram of another exemplary process of the positioning method in an embodiment of the present application;

[0069] Figure 6 This is a schematic diagram of the module structure of the positioning device according to an embodiment of the present application;

[0070] Figure 7 Schematic diagram of the device structure of the hardware operating environment involved in the positioning method in the embodiment of the present application.

[0071] The purpose, features and advantages of the embodiments of the present application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0072] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the embodiments of the present application and are not used to limit the embodiments of the present application.

[0073] In order to better understand the technical solutions of the embodiments of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0074] Mobile phone signaling data records the connection interaction relationship and timestamp information between the mobile terminal carried by the user and the communication base station, and also includes the location of the base station. Through the mobile phone signaling data and the location of the base station, the user's spatiotemporal trajectory can be extracted to determine the user's location. For example, the location of the base station to which the mobile terminal is connected can be directly used as the user's location.

[0075] However, since each base station usually covers a certain spatial service range, the user's location may also frequently switch between multiple base stations or drift to a distant base station in a short period of time due to abnormal signal connection and other reasons. Directly using the location of the base station to which the mobile terminal is connected as the user's location will result in a certain spatial error, which will lead to inaccurate user location positioning. In other words, the accuracy of the user's spatiotemporal location extracted based on signaling data can only reach the granularity of the base station, and it is impossible to accurately determine the user's actual location, and it is also difficult to determine the user's actual activity trajectory.

[0076] To this end, in order to solve the above defects, the embodiment of the present application trains a preset trajectory reconstruction model, based on the spatiotemporal coding sequence corresponding to the first signaling of the user to be located and the preset trajectory reconstruction model, the actual position of the user corresponding to the first signaling can be accurately mapped. Among them, the spatiotemporal coding sequence is composed of multiple spatiotemporal coding vectors, each of which is composed of a time coding vector and a space coding vector, and the space coding vector can describe the distribution of POI points of interest around the base station, and can reflect the distance of each POI point of interest from the base station. The preset trajectory reconstruction model can predict the most likely location of the user to be located based on the spatiotemporal coding sequence, which can be the location of the POI point of interest or the location of the base station, for example, the closer the POI point of interest is to the base station, the greater the probability that the user will appear at the POI point of interest, and the time coding vector can reflect the periodic activities of the user, for example, the possible location of n users per week, the possible location of users per day, etc., and n is 1 to 7. In this embodiment, the distance relationship between the POI point of interest and the base station is one of the influencing factors for determining the actual location of the target user to be located. The above is only an example, and this embodiment is not specifically limited to this. The specific implementation method for determining the actual location of the target can refer to the following embodiment. Furthermore, this embodiment can reconstruct the signaling trajectory of the user to be located according to the actual positions of multiple targets of the user to be located.

[0077] Based on this, the present application embodiment provides a positioning method, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the positioning method of the present application. The positioning method includes steps S10 to S30:

[0078] Step S10, obtaining a signaling sequence of a first signaling of a user to be located, wherein the signaling sequence includes a plurality of sub-signals, each of which is a first signaling and a time-associated signaling that is time-continuous with the first signaling of the first signaling;

[0079] It should be noted that the user to be located refers to the user whose location needs to be determined. The signaling records the connection interaction relationship between the mobile terminal and the base station and the location information of the base station. The signaling includes the user ID holding the mobile terminal, the base station ID, the longitude and latitude of the base station, and the start time and end time of the signaling. The latitude and longitude of the base station are used to describe the location information of the base station. The start time refers to the time when the signaling starts recording, and the end time is the time when the signaling ends recording.

[0080] The first signaling is the signaling generated by the mobile terminal carried by the user to be located. The first signaling also refers to the signaling of the actual target location of the user to be located that needs to be located. The first signaling time refers to the start time and end time of the first signaling. The time-associated signaling is also the signaling generated by the mobile terminal carried by the user to be located. The associated signaling time of the time-associated signaling is associated with the first signaling time. The associated signaling time refers to the start time and end time of the time-associated signaling.

[0081] The signaling sequence includes a first signaling and a time-associated signaling that is time-continuous with the first signaling of the first signaling. The sub-signaling may be the first signaling or the time-associated signaling. The time-associated signaling that is time-continuous with the first signaling refers to one or more time-associated signalings that are closest to the first signaling time. The associated signaling time of the time-associated signaling may be continuous with the first signaling time or may be discontinuous. There may be multiple time-associated signalings included in the signaling sequence. In this embodiment, in order to ensure the accuracy of position prediction, multiple time-associated signaling, for example, 10 signaling, are provided in a signaling sequence, with 5 signaling before and after the first signaling time of the first signaling as time-associated signaling. In the signaling of the user to be located, N signalings that are sequentially arranged before the start time of the first signaling time are determined, and all N signalings are used as time-associated signalings. Also, M signalings that are sequentially arranged after the end time of the first signaling time are determined, and all M signalings are used as time-associated signalings. Generally, N is the same as M, and N and M are integers, so that the activity trajectory of the user within a certain range of the first signaling time can be described, thereby facilitating more accurate determination of the user's actual target position in the first signaling in the subsequent process.

[0082] Exemplarily, the first signaling of the user to be located can be obtained, and the N signalings that are sequentially arranged before the first signaling time of the first signaling, and the N signalings that are sequentially arranged after the first signaling time are all used as time-associated signaling of the first signaling, and a signaling sequence including the first signaling and each time-associated signaling is determined, and the user identifiers of the first signaling and each time-associated signaling are the same. In other embodiments, multiple signalings of the user to be located that are sequentially arranged according to the signaling time can also be directly obtained, and an odd number of signalings greater than 1 is selected from the multiple signalings as the signaling sequence, and the signaling arranged in the middle is used as the first signaling, and the signalings in the signaling sequence except the first signaling are used as time-associated signaling.

[0083] Step S20, determining the spatiotemporal coding vector of each sub-signaling according to the acquired POI points of interest and each sub-signaling;

[0084] It should be noted that POI refers to a place, venue or building that can arouse people's interest. POI can be obtained from a map system. POI includes a POI number, POI longitude and latitude, and POI type. POI number is used to distinguish different POI points of interest, POI longitude and latitude are used to describe the location of the POI point of interest, and POI type is used to describe the type of the POI point of interest. There are existing POI types such as shopping malls, car services, financial services, and scenic spots, which are not specifically limited in this embodiment.

[0085] The POI point of interest refers to a POI associated with a sub-base station, and the distance between the POI point of interest corresponding to the sub-signaling and the sub-base station is less than the preset distance threshold, that is, the POI point of interest corresponding to the sub-signaling is a POI within the preset distance threshold range of the sub-base station of the sub-signaling. The preset distance threshold can be determined based on actual conditions. The sub-base station refers to the base station of the sub-signaling.

[0086] The spatiotemporal coding vector is obtained by concatenating the spatial coding vector and the temporal coding vector. The spatial coding vector represents the distribution of the base station and the POI points of interest around the base station. The temporal coding vector can comprehensively represent the duration and absolute time of the sub-signaling. The duration is used to describe the duration between the start time and the end time of the sub-signaling. The absolute time is represented by the month, week, date and hour of the start time of the sub-signaling, and the month, week, date and hour of the end time of the sub-signaling. The spatiotemporal coding vector is used to comprehensively describe the space and time corresponding to the sub-signaling. Each sub-signaling has its own corresponding spatiotemporal coding vector.

[0087] Exemplarily, for each sub-signaling, each POI point of interest within a preset distance range of the sub-base station of the sub-signaling is obtained, and the spatiotemporal coding vector of the sub-signaling is determined according to the sub-signaling and each POI point of interest.

[0088] Step S30, inputting the space-time coding sequence constructed by each space-time coding vector into a preset trajectory reconstruction model to obtain the target actual position of the user to be located in the first signaling mapped by the space-time coding sequence.

[0089] It should be noted that the preset trajectory reconstruction model is obtained by training the neural network model, and the space-time coding sequence is composed of various space-time coding vectors. The space-time coding vectors in the space-time coding sequence are sorted according to the chronological order of the sub-signaling time of their corresponding sub-signalings, for example, they are sorted according to the chronological order of the start time of the sub-signaling time.

[0090] The target actual position refers to the precise positioning position of the user to be located at the first signaling time. The preset trajectory reconstruction model can determine the mapping relationship between the spatiotemporal coding sequence and the target actual position, so the preset trajectory reconstruction model can map the target actual position of the user to be located at the first signaling time.

[0091] Since the spatiotemporal coding vector of the first signaling can provide less information to the preset trajectory reconstruction model, this embodiment provides more auxiliary information for the preset trajectory reconstruction model by inputting the spatiotemporal coding sequence of the first signaling into the preset trajectory reconstruction model (for example, the activity of the user to be located near the first signaling time reflected by the spatiotemporal coding sequence), so that the preset trajectory reconstruction model can more accurately predict the actual target position of the user to be located in the first signaling.

[0092] Exemplarily, each space-time coding vector is sorted according to the sub-signaling time of the sub-signaling corresponding to each space-time coding vector to obtain a sorting result, and each space-time coding vector is combined in sequence according to the sorting result to obtain a space-time coding sequence, and the space-time coding sequence is input into a preset trajectory reconstruction model so that the preset trajectory reconstruction model predicts the target actual position of the user to be located in the first signaling based on the space-time coding sequence.

[0093] In an embodiment of the present application, a signaling sequence of a first signaling of a user to be located is obtained, wherein the multiple sub-signals in the first signaling sequence are respectively the first signaling and the time-associated signaling which is continuous with the first signaling time, and then a space-time coding vector of each sub-signaling is determined according to the POI points of interest of each sub-signaling and each sub-signaling obtained, and the space-time coding vector can reflect the distribution of the surrounding POI points of interest corresponding to the sub-signaling, so that the space-time coding sequence obtained after splicing the space-time coding vectors can comprehensively reflect the distribution of the POI points of interest around all the sub-signals, that is, the space-time coding sequence can reflect the distribution of the surrounding POI points of interest corresponding to the first signaling, and at the same time can reflect the distribution of the surrounding POI points of interest corresponding to the time-associated signaling, and since the time-associated signaling is continuous with the first signaling time of the first signaling, the space-time coding sequence can also reflect the activity trajectory of the user to be located near the first signaling time of the first signaling, and then through the space-time coding sequence and the preset trajectory reconstruction model, the preset trajectory reconstruction model is used to predict the target actual position of the user to be located mapped by the space-time coding sequence in the first signaling.

[0094] Furthermore, the distribution of the first POI point of interest around the first signaling reflected by the space-time coding sequence, the distribution of the second POI point of interest around the time-related signaling, and the activity trajectory of the user to be located near the first signaling time of the first signaling are achieved, and the actual target position of the user to be located corresponding to the first signaling is accurately predicted, instead of directly taking the base station position in the first signaling as the actual position of the user to be located corresponding to the first signaling, thereby improving the accuracy of user positioning.

[0095] Further, see Figure 2 In one feasible embodiment, step S20 further includes steps S21 to S24:

[0096] Step S21, for each sub-signaling, obtaining each POI point of interest of the sub-base station in the sub-signaling;

[0097] Step S22, determining a spatial coding vector of the sub-signaling according to the sub-signaling and the corresponding POI points of interest;

[0098] The spatial coding vector of the sub-signaling may be constructed according to the sub-signaling, the position of the sub-base station in the sub-signaling, and the positions of the POI points of interest corresponding to the sub-signaling.

[0099] Exemplarily, a preset attention algorithm can be introduced, and a spatial coding vector of the sub-signaling can be constructed according to the preset attention algorithm, the sub-signaling and the corresponding points of interest. The preset attention algorithm is used to assign a corresponding attention weight to each POI point of interest according to the distance between the sub-base station and the corresponding POI point of interest, so that the distance between the POI point of interest and the sub-base station can be determined in the spatial coding vector by the attention weight of the POI point of interest. The larger the attention weight, the closer the POI point of interest is to the corresponding sub-base station, and the smaller the attention weight, the farther the POI point of interest is from the corresponding sub-base station.

[0100] In a feasible embodiment, step S22 further includes: steps S221 to S225:

[0101] Step S221, performing nonlinear changes on the sub-signaling to obtain a signaling vector, and performing nonlinear changes on each POI point of interest to obtain an interest vector corresponding to each POI point of interest;

[0102] The nonlinear change of the sub-signaling refers to processing the sub-signaling with one or more nonlinear functions to obtain a vector representation of the sub-signaling. The nonlinear change of the POI point of interest refers to processing the POI point of interest with one or more nonlinear functions to obtain a vector representation of the POI point of interest. The nonlinear function can be a function such as an activation function, a trigonometric function, etc., which is not specifically limited in this embodiment.

[0103] Step S222, determining the signaling query parameters of the signaling vector according to the preset attention algorithm, and determining the interest key and interest value corresponding to each interest vector;

[0104] It should be noted that the preset attention algorithm is an algorithm based on the attention mechanism. The signaling query parameter refers to the signaling vector as the Query in the attention mechanism. The interest key refers to the Key for determining the POI point of interest in the attention mechanism. The interest value refers to the Value of the POI point of interest in the attention mechanism.

[0105] For example, a nonlinear change is performed on the sub-signaling to obtain a 1×D signaling vector x signal , nonlinearly transform the POI interest point to obtain a 1×D interest vector x poi , all interest vectors corresponding to the sub-signaling can be comprehensively expressed as n×D x poi , where D is the length of the vector and n is the number of POI points of interest corresponding to the sub-signaling. Based on this, the formula for determining the signaling query parameter, interest key and interest value can be expressed as:

[0106] query = ReLU(W Q x signal )

[0107] key i =ReLU(W K x ipoi )

[0108] value i =ReLU(W V x ipoi )

[0109] in, are the weight matrices of query, key and value respectively. ReLU is a nonlinear activation function. Q =D K =Dv. query is the signaling query parameter, key is the interest key and value is the interest value, D Q , D K , Dv are W Q , W K and W V The dimension of . ReLU is the activation function, x ipoi For i x poi , i is an integer greater than 0 and less than n. key i is the interest key of the i-th interest vector. i is the interest value of the i-th interest vector.

[0110] Step S223, calculating the attention weights between the signaling query parameters and each interest key respectively;

[0111] The attention weight can be used to describe the location similarity between the location of the POI point of interest and the location of the corresponding sub-base station. Since the location of the POI point of interest is represented by longitude and latitude, the location of the sub-base station is also represented by longitude and latitude. When the longitude and latitude of the POI point of interest are closer to the longitude and latitude of the sub-base station, the location similarity between the POI point of interest and the sub-base station is higher, which also means that the closer the POI point of interest is to the sub-base station, the higher the attention weight. The attention weight of the interest key can be used as the attention weight of the interest vector.

[0112] Exemplarily, the formula for calculating the attention weight of each interest vector can be expressed as:

[0113]

[0114] Among them, attention(query,key i ) is the query and key i Attention weights, query and key T The similarity is calculated by dot product operation. The softmax function is used to calculate the weight, key i T For key i The transpose of .

[0115] Step S224, according to the attention weight of each interest key, weighted sum is performed on the interest value of the interest vector where each interest key is located, to obtain the interest point aggregation score of the sub-signaling;

[0116] The interest point convergence score is used to describe the overall distribution of each POI point of interest around the sub-signaling. The attention weight corresponding to the interest vector is used as the weight of the interest value, and each interest value is weighted and summed to obtain the interest point convergence score.

[0117] Exemplarily, the formula for calculating the interest point convergence score may be:

[0118]

[0119] Among them, f(q,(k1,v1),…,(k n ,v n )) is the interest point aggregation score.

[0120] Step S225, determine the longitude vector and latitude vector of the sub-signaling from the signaling vector, concatenate the longitude vector, latitude vector and the interest point aggregation score to obtain a spatial coding vector.

[0121] The longitude vector and latitude vector of the sub-signaling are respectively represented by the longitude and latitude of the sub-base station in the sub-signaling. Exemplarily, the spatial coding vector can be represented as:

[0122] x spatial =[x signal .longitude,x signal .latitude,f]

[0123] Among them, “·” represents the vector concatenation operation, x spatial is the spatial encoding vector.

[0124] This embodiment determines the point of interest convergence score between each POI point of interest and the corresponding sub-base station, and splices the point of interest convergence score and the longitude vector and latitude vector of the sub-signaling to obtain a spatial coding vector, so that the subsequent pre-trajectory reconstruction model can simultaneously utilize multiple levels of information (the distance between the POI point of interest and the sub-signaling, the POI points of interest around the sub-base station), so as to improve the prediction accuracy of the preset trajectory reconstruction model. The vector fusion method used in this embodiment is not limited to direct splicing. The longitude and latitude of the POI point of interest closer to the sub-base station are closer to the longitude and latitude of the base station. In order to increase the influence of the closer POI point of interest on the base station, a higher weight will be assigned to the interest vector of the closer POI point of interest.

[0125] Step S23, determining a time coding vector of the sub-signaling according to the sub-signaling time of the sub-signaling;

[0126] Step S24, concatenate the spatial coding vector and the temporal coding vector of the same sub-signaling to obtain the spatiotemporal coding vector of the sub-signaling.

[0127] It should be noted that the sub-signaling time includes the start time and the end time of the sub-signaling. Exemplarily, the duration of the sub-signaling is determined according to the start time and the end time of the sub-signaling, and the time coding vector is constructed according to the first absolute time represented by the start time, the second absolute time represented by the end time, and the duration. Each sub-signaling has its own corresponding space-time coding vector.

[0128] The formula for concatenating the spatial coding vector and the temporal coding vector of the same sub-signaling to obtain the spatiotemporal coding vector of the sub-signaling can be:

[0129] x spati-temp =[x spatial ,x temporal ]

[0130] Among them, x spati-temp is the space-time encoding vector, x temporal is the time encoding vector, x spatial is the spatial encoding vector.

[0131] In a feasible embodiment, step S23 includes steps S231 to S234:

[0132] Step S231, calculating the duration of the sub-signaling according to the start time and the end time;

[0133] Step S232, performing nonlinear change on the duration to obtain a time difference vector;

[0134] Step S233, performing nonlinear transformation on the start time and the end time respectively to obtain a first time semantic vector of the start time and a second time semantic vector of the end time;

[0135] Step S234, concatenating the time difference vector, the first time semantic vector and the second time semantic vector to obtain a time coding vector.

[0136] It should be noted that the duration describes the time from the start time to the end time of the sub-signaling, and the duration can include the start time and the end time. The time difference vector is a vector representation of the duration, the first time semantic vector is a vector representation of the first absolute time represented by the start time, and the second time semantic vector is a vector representation of the second absolute time represented by the end time. The first absolute time refers to the month, week, date and hour of the start time; the second absolute time refers to the month, week, date and hour of the end time. Each sub-signaling has its own corresponding time code.

[0137] Exemplarily, the process of determining the time coding vector of the sub-signaling may include: the formula for calculating the duration is:

[0138] The signaling record is calculated based on the start time and end time. The calculation formula is as follows:

[0139] duration = x signal .endTime-x signal .stratTime

[0140] Among them, duration is the duration, the unit of duration is seconds, x signal .endTime is the end time of the sub-signaling, x signal .stratTime is the start time of the sub-signaling.

[0141] The duration may be subjected to a weighted nonlinear transformation, and the calculation formula for determining the time difference vector of the sub-signaling may be:

[0142] delta=cos(W delta duration)

[0143] Among them, delta is the time difference vector, is the weight matrix, D d W delta The matrix dimension of is, and cos(·) is the cosine transform function.

[0144] The formula for determining the first time semantic vector and the second time semantic vector may be:

[0145] semantic1=e(m1)+e(w1)+e(d1)+e(h1)

[0146] semantic2=e(m2)+e(w2)+e(d2)+e(h2)

[0147] Where semantic1 is the first time semantic vector, semantic2 is the second time semantic vector, m1, w1, d1 and h1 are the month, week, date and hour of the start time, m2, w2, d2 and h2 are the month, week, date and hour of the end time, and e(·) is a nonlinear transformation that converts the data in the brackets of e into 1×D s Specifically, the month, week, date, and hour of the start time are nonlinearly transformed and concatenated to obtain a first time semantic vector, and the month, week, date, and hour of the end time are nonlinearly transformed and concatenated to obtain a second time semantic vector.

[0148] Concatenate the time difference vector, the first time semantic vector, and the second time semantic vector to obtain the time encoding vector:

[0149] x temporal =[delta,semantic1,semantic2]

[0150] Among them, x temporal is the time encoding vector.

[0151] In a feasible embodiment, the positioning method further includes steps A10 to A50:

[0152] Step A10, obtaining a training sample set, wherein the training sample set includes a plurality of training samples, each training sample includes a training signaling sequence, a training POI and an actual location tag, and each training sub-signaling included in the training signaling sequence is a signaling to be trained, and a training associated signaling that is time-continuous with the training signaling of the signaling to be trained;

[0153] It should be noted that the training sample set can be used to train the model to be trained. Each training sample in the training sample set includes a training signaling sequence, a training POI and an actual location tag. The same training sample includes a training signaling sequence, a training POI and an actual location tag that are mutually associated, that is, in the same training sample, the signaling to be trained and the training associated signaling in the training signaling sequence are associated with the actual location tag. The signaling to be trained and each training associated signaling in the training signaling sequence have their own corresponding training POIs. The training signaling sequence includes multiple training sub-signals, and the training sub-signaling can be a signaling to be trained or a training associated signaling. The training signaling time refers to the signaling time of the signaling to be trained, and the associated signaling time of the training associated signaling can be continuous with the training signaling time or discontinuous. The training signaling sequence is a signaling sequence used in the model training process. The training POI is also a POI point of interest used for model training, and the actual location tag is the actual location tag corresponding to the signaling to be trained. The user identifier of each training associated signaling in the training signaling sequence is the same as the user identifier of the signaling to be trained. The same user identity may correspond to one training signaling sequence or to multiple training signaling sequences. The signaling to be trained in each training signaling sequence of the same user identity is different, and the training-related signaling is also different.

[0154] Step A20, determining the training spatiotemporal coding vectors of the training sub-signals according to the training sub-signals and the training POIs;

[0155] It should be noted that each training sub-signaling has its own corresponding training POI, and the training space-time coding vector includes a training space coding vector and a training time coding vector. The training space-time coding vector is a space-time coding vector used to train the model to be trained, and the construction process of the training space-time coding vector is consistent with the construction process of the space-time coding vector. The training space-time coding vector is also obtained by splicing the training space coding vector and the training time coding vector. The training space coding vector is a space coding vector used for training, and the training time coding is a time coding vector used for training. The construction process of the training space coding vector is the same as the construction process of the space coding vector, and the construction process of the training time coding vector is the same as the construction process of the time coding vector, and this implementation will not be repeated here.

[0156] For example, a time coding vector of a training sub-signaling can be constructed according to the signaling time corresponding to the training sub-signaling, and a spatial coding vector of a training sub-signaling can be constructed according to the training sub-signaling and the corresponding training POI. The spatial distances of the training POIs corresponding to each training sub-signaling from the training sub-signaling are less than a preset distance threshold. Therefore, the constructed training spatial coding vector can more accurately describe the distribution of POIs near the training sub-signaling, avoid reflecting POIs far away from the training sub-signaling in the training spatial coding vector, and thus avoid reducing the accuracy of user positioning.

[0157] Step A30, inputting the spatiotemporal coding sequence constructed by each training spatiotemporal coding vector into a preset model to be trained to obtain a first prediction result, wherein the first prediction result includes a predicted position of the signaling to be trained;

[0158] It should be noted that the model to be trained is a neural network model. The training samples are used to train the model to be trained, and the training sample set can be divided into a training set, a validation set, and a test set according to a preset ratio, for example, the preset ratio can be 7:1:2, etc., which is not specifically limited in this embodiment.

[0159] In this embodiment, the model to be trained is composed of a recurrent neural network and a linear layer, and the number of layers of the recurrent neural network is l. R , the number of linear layers is l L , the number of neurons in the hidden layer of the recurrent neural network and the linear layer is shown in the formula:

[0160]

[0161] Where i is the number of layers in the model. input_size is the dimension of the input training spatiotemporal encoding sequence, hidden_size i The number of neurons in the hidden layer can be expressed as:

[0162] f(x)=max(0,x)

[0163] Among them, x can be represented as the input data of the model to be trained, for example, x can be the training spatiotemporal coding vector in the training spatiotemporal coding sequence, etc., f(x) can be represented as the input data activated by the relu function, and max(0,x) represents that when 0 is greater than x, f(x) is 0, and when x is greater than 0, f(x) is x. In this embodiment, the dimension of the output of the model to be trained is 2, so as to facilitate the output of the actual target position of the user, that is, the longitude and latitude of the user.

[0164] Step A40, calculating the prediction loss according to the first prediction result and the actual position label;

[0165] It should be noted that the first prediction result and the actual location tag can be calculated based on a preset loss function, the first prediction result includes the predicted location of the signaling to be trained, the predicted location is the longitude and latitude of the user identifier in the signaling to be trained predicted by the training model, and the first prediction result can also include the predicted time of the longitude and latitude when the user is at the predicted location. The actual location tag includes the actual latitude tag and the actual longitude tag corresponding to the signaling to be trained. In this embodiment, the preset loss function can be a mean square error function.

[0166] For example, the formula for calculating the prediction error can be:

[0167]

[0168] Where MSE is the prediction error, N is the cumulative number of training spatiotemporal encoding sequences input to the hyperparameter training model, is the preset position corresponding to the i-th training spatiotemporal coding sequence, y i is the actual position label corresponding to the i-th training spatiotemporal coding sequence. The smaller the prediction error, the better the prediction performance of the model to be trained. The larger the prediction error, the worse the prediction performance of the model to be trained.

[0169] Step A50, optimizing the model to be trained according to the predicted loss to obtain a preset trajectory reconstruction model that has been trained.

[0170] It should be noted that it is possible to determine whether the prediction loss is less than a preset training error threshold. When the prediction loss is less than the preset training error threshold, the Adam optimization algorithm can be used to update the learning parameters of the model to be trained.

[0171] For example, the learning parameters of the model to be trained can be updated using the gradient descent method, and the training spatiotemporal coding sequence can be re-acquired, and the step A30 can be returned to, and the re-acquired training spatiotemporal coding sequence is different from the training spatiotemporal coding sequence of the previous time step. Until the prediction error is less than the preset training error threshold, the model to be trained is used as the initial trajectory reconstruction model, and then the initial trajectory reconstruction model is optimized to obtain the predicted trajectory reconstruction model.

[0172] In a feasible embodiment, step A50 includes steps A51 to A53:

[0173] Step A51, optimizing the model to be trained according to the prediction loss until the prediction loss is less than a preset training error threshold, thereby obtaining an initial trajectory reconstruction model;

[0174] Step A52, verifying the initial trajectory reconstruction model according to the obtained verification set, and obtaining the verification position output by the initial trajectory reconstruction model;

[0175] It should be noted that the validation set is used to verify the performance of the initial trajectory reconstruction model so as to determine whether to update the hyperparameters later. The test set can be used to test the preset trajectory reconstruction model to ensure the accuracy of the output results of the preset trajectory reconstruction model so as to accurately predict the actual location of the user.

[0176] Exemplarily, a verification spatiotemporal coding sequence and a corresponding verification actual position label can be determined from the verification set, and the verification spatiotemporal coding sequence is input into the initial trajectory reconstruction model to obtain a verification position, which is the position corresponding to the verification spatiotemporal coding sequence output by the initial trajectory reconstruction model.

[0177] Step A53, if the verification position does not meet the preset optimal effect conditions, the hyperparameters of the model to be trained are updated, and the model to be trained is retrained until the verification position meets the preset optimal effect conditions, and a preset trajectory reconstruction model is obtained.

[0178] It should be noted that the hyperparameters include training round parameters, learning rate and data group size. The training round parameters are used to control the model to be trained. The learning rate is characterized by the update rate of the learning parameters of the model to be trained. The data group size refers to the amount of data of the samples used for each training during the training of the model to be trained. In this embodiment, the training round parameters, learning rate and training group parameters can be set to 150, 0.001 and 32 respectively, or to other values. In order to select the model with the best prediction effect, the hyperparameters of the model to be trained can be updated to select the initial trajectory reconstruction model corresponding to the hyperparameters with the best prediction effect.

[0179] It should also be noted that the verification difference between the verification position and the verification actual position label can be calculated according to the preset loss function, and the preset optimal condition for the effect can be that the verification difference corresponding to the verification position is less than or equal to the preset verification loss threshold. In this embodiment, the hyperparameter can be determined based on the actual situation, and the hyperparameter update can be reconfigured by relevant technical personnel, which is not specifically limited in this embodiment.

[0180] Exemplarily, when the verification difference of the verification training position is greater than the preset loss threshold, the hyperparameters are updated and the model to be trained is retrained until the verification difference of the verification position is less than or equal to the preset verification loss threshold. The initial trajectory reconstruction model corresponding to the verification difference less than or equal to the preset verification loss threshold is used as the preset trajectory reconstruction model, which is equivalent to saving the hyperparameters corresponding to the initial trajectory reconstruction model at the verification position that meets the preset optimal effect conditions.

[0181] In other embodiments, different hyperparameters can also be obtained, and each hyperparameter is configured in the model to be trained to obtain the hyperparameter training model corresponding to each hyperparameter; each hyperparameter training model is iteratively trained according to the training samples, and the initial trajectory reconstruction model of each hyperparameter training model is determined; each initial trajectory reconstruction model is verified according to the obtained verification set to obtain the verification position output by each initial trajectory reconstruction model; the verification difference corresponding to each verification position is determined, and the verification difference is less than or equal to the preset verification loss threshold value and the smallest target verification difference among the verification differences is selected, and the initial trajectory reconstruction model corresponding to the target verification difference is used as the preset trajectory reconstruction model. The preset trajectory reconstruction model can be tested according to the test set to ensure the accuracy of the preset trajectory reconstruction model.

[0182] This embodiment updates the hyperparameters, thereby training a preset trajectory reconstruction model with better performance and improving the accuracy of positioning.

[0183] For a better understanding of this embodiment, please refer to Figure 3 , the process of this embodiment is described, including steps Z101 to Z107, step Z101: start, step Z102: obtain a training sample set; step Z103: iteratively train the hyperparameter model based on the training sample set to obtain an initial trajectory reconstruction model; step Z104: verify and evaluate the initial trajectory reconstruction model; for example, determine a verification set from the training samples, test and evaluate the initial trajectory reconstruction model based on the verification set, obtain a verification position output by the initial trajectory reconstruction model, and evaluate the effect of the initial trajectory reconstruction model based on the verification position; step Z105: the model effect is Whether it is optimal; for example, whether the verification difference corresponding to the verification position is less than or equal to the preset verification loss threshold. If the verification difference is less than or equal to the preset verification loss threshold, it means that the effect of the initial trajectory reconstruction model is not optimal, then execute step Z106b: update the model to be trained, and then execute step Z102; if the verification difference is less than or equal to the preset verification loss threshold, it means that the model effect of the initial trajectory reconstruction model is optimal, and execute step Z106a: save the hyperparameters of the initial trajectory reconstruction model; and the initial trajectory reconstruction model can be used as the preset trajectory reconstruction model; finally, execute step Z107: end.

[0184] In a feasible embodiment, step A10 further includes steps A11 to A13:

[0185] Step A11, acquiring each signaling to be trained, each precise positioning data and each training POI to be screened, wherein the training POI to be screened includes the POI position, and the signaling to be trained includes the training base station position;

[0186] It should be noted that the precise positioning data is used to accurately determine the position. The precise positioning data can be the collected GPS (Global Positioning System) precise positioning data, or the data collected by the Beidou positioning system, etc. This embodiment does not make specific limitations on this. The precise positioning data can be obtained from the sampled MR (Measurement Report), MDT (Modular Dynamic formation Tester) data or the parsed OTT (Over-The-Top) data. The precise positioning data includes user identification, longitude, latitude and sampling time, and the sampling time is the time when the precise positioning data is sampled.

[0187] The training time coding vector of each signaling to be trained can be determined according to the training signaling time of each signaling to be trained; the training signaling time is the signaling time of the signaling to be trained, and the construction process of the training time coding vector is the same as the construction process of the time coding vector. Please refer to the description of the above steps S231 to S234, and this embodiment will not be elaborated here.

[0188] Step A12, for each signaling to be trained, in each training POI to be screened, a target POI position whose first spatial distance between the POI position and the training base station position is less than a preset distance threshold is screened out, and the signaling to be trained and the training POI at the target POI position are associated;

[0189] It should be noted that the training base station position refers to the position of the base station in the signaling to be trained, the training POI to be screened is the POI collected for training, and the training POI to be screened can also be obtained from the map. The POI position refers to the position of the training POI to be screened, and the POI position can be expressed in longitude and latitude. Since the training POI to be screened may be too far away from the training base station position in the signaling to be trained, and is therefore not suitable for describing the POI distribution around the training base station of the signaling to be trained, in this embodiment, it is necessary to separately determine the training POI associated with each signaling to be trained. For any signaling to be trained and any training POI to be screened, the first spatial distance can be calculated by the longitude and latitude of the POI position and the longitude and latitude of the training base station position.

[0190] Exemplarily, the training POIs associated with each signaling to be trained can be determined based on the respective training base station positions in each signaling to be trained and the respective POI positions of each training POI to be screened. Specifically, the first spatial distance between each signaling to be trained and each training POI to be screened is calculated based on the respective training base station positions in each signaling to be trained and the respective POI positions of each training POI to be screened; for each signaling to be trained, a target spatial distance less than a preset distance threshold is screened out from the spatial distances corresponding to the signaling to be trained, and the training POI corresponding to the signaling to be trained and the target spatial distance is associated.

[0191] Step A13, determining the actual location tag corresponding to each signaling to be trained and the training-related signaling corresponding to each signaling to be trained according to each signaling to be trained and each precise positioning data;

[0192] It should be noted that the actual location tag is the precise positioning data that has established an association relationship with the signaling to be trained, that is, the precise positioning data corresponding to the signaling to be trained. Since the precise positioning data is obtained by sampling and is highly random, each signaling to be trained may not have its own corresponding precise positioning data, that is, not all of them have actual location tags. The user identifiers of the signaling to be trained and the corresponding actual location tag are the same.

[0193] Exemplarily, the actual location label corresponding to each signaling to be trained is determined based on the user identifier and training base station location of each signaling to be trained and the user identifier of each precise positioning data. If there is no corresponding precise positioning data for a signaling to be trained, there is no need to determine the actual location label of the signaling to be trained, or the signaling to be trained can be eliminated to avoid adding it to the training sample.

[0194] In a feasible embodiment, the training-related signaling includes a first-related signaling and a second-related signaling; step A13 also includes steps A131 to A134:

[0195] Step A131, determining the training signaling of each user and the precise positioning data of each user with the same user identifier according to the respective user identifiers in the signaling to be trained and the respective user identifiers in the precise positioning data;

[0196] Step A132, determining a target positioning mapping relationship among the training signaling of each user and the precise positioning data of each user, wherein the target positioning mapping relationship is the training signaling of the target user and the precise positioning data of the target user that are temporally and spatially associated;

[0197] It should be noted that the signaling to be trained and the precise positioning data can be classified to determine the user training signaling and the user precise positioning data of the same user identity. The user training signaling and the user precise positioning data are the signaling to be trained and the precise positioning data with the same user identity.

[0198] The time-space association refers to the target mapping relationship that the target user training signaling and the target user precise positioning data are associated in time and space, the time association refers to the sampling time of the target user precise positioning data within the user signaling time of the target user training signaling, and the space association refers to the second spatial distance between the target user precise positioning data and the target user training signaling is less than the preset distance threshold. The user signaling time refers to the signaling time of the target user training signaling, and the user signaling time includes the user signaling start time and the user signaling end time.

[0199] Exemplarily, among the training signaling of each user and the precise positioning data of each user, the training signaling of the target user and the precise positioning data of the target user that are temporally and spatially correlated are determined.

[0200] In a feasible embodiment, step A132 further includes steps B10 to B40:

[0201] Step B10, determining the respective user signaling time from the training signaling of each user, and determining the respective sampling time from the precise positioning data of each user, wherein the user signaling time includes the user signaling start time and the user signaling end time;

[0202] Step B20, associating the initial user precise positioning data and the initial user training signaling within the user signaling time to obtain an initial positioning mapping relationship;

[0203] It should be noted that the positioning mapping relationship is the initial user precise positioning data and initial user training signaling with the same user ID and sampling time within the user signaling time. The sampling time within the user signaling time means that the sampling time is less than the user signaling start time and greater than the user signaling end time.

[0204] Exemplarily, for each user training signaling of the same user identifier, the user signaling time is determined from the user training signaling, and the respective sampling times are determined from the precise positioning data of each user; the initial user precise positioning data with a sampling time within the user signaling time is filtered out from the precise positioning data of each user with the same user identifier, and the user training signaling at the user signaling time is used as the initial user training signaling, the initial user training signaling and the initial user precise positioning data are associated to obtain an initial positioning mapping relationship.

[0205] Step B30, respectively calculating the distance between the initial user training signaling and the initial user precise positioning data in each initial positioning mapping relationship to obtain a second spatial distance corresponding to each initial positioning mapping relationship;

[0206] Step B40: Filter out target positioning mapping relationships whose second spatial distance is less than a preset distance threshold from each initial positioning mapping relationship.

[0207] It should be noted that the second spatial distance refers to the distance between the training base station position in the initial user training signaling and the initial user precise positioning data in the same initial positioning mapping relationship. The sampling time of the initial user precise positioning data in the same initial positioning mapping relationship is within the training signaling time of the initial user training signaling, which indicates that the user corresponding to the user identifier at this time is likely to be at the location where the initial user precise positioning data is located. However, due to possible signal connection abnormalities and other reasons, the mobile terminal that the user carries may frequently switch between multiple base stations or drift to a farther base station in a short time, so the initial user training signaling may be abnormal. Therefore, it is necessary to calculate the second spatial distance to eliminate the initial positioning mapping relationship whose second spatial distance is greater than the preset distance threshold, thereby facilitating the avoidance of collecting erroneous samples and affecting the training effect.

[0208] Exemplarily, for each initial positioning mapping relationship, the distance between the initial user training signaling and the initial user precise positioning data is calculated based on the training base station position of the initial user training signaling in the same initial positioning mapping relationship and the initial user precise positioning data, to obtain the second spatial distance of the initial positioning mapping relationship. In each initial positioning mapping relationship, the target positioning mapping relationship whose second spatial distance is less than the preset distance threshold is screened out, that is, the initial user precise positioning data in the initial positioning mapping relationship whose second spatial distance is less than the preset distance threshold is used as the target user precise positioning data, and the initial user signaling data is used as the target user signaling data. The initial user precise positioning data and initial user training signaling whose second spatial distance is greater than or equal to the preset distance threshold are eliminated.

[0209] Furthermore, if there are multiple target positioning mapping relationships with the same target user signaling data, the one with the smallest second spatial distance can be used as the final target positioning mapping relationship, thereby facilitating conflicts in actual location tags corresponding to the target user signaling data.

[0210] Step A133, using the target user precise positioning data of the target user training signaling in the target positioning mapping relationship as the actual position label of the target user training signaling;

[0211] Step A134, for any target positioning mapping relationship, according to the user signaling time of the target user training signaling in the target positioning mapping relationship, among each user training signaling corresponding to the user identifier of the target user training signaling, filter out the first associated signaling whose training signaling time is sorted before the user signaling time and whose number is a preset number, and filter out the second associated signaling whose training signaling time is sorted after the user signaling time and whose number is a preset number.

[0212] It should be noted that the first associated signaling and the second associated signaling have the same user identifier as the target user training signaling. There can be multiple first associated signaling and second associated signaling, and the preset number can be determined based on the actual situation. For example, the preset number can be 5. The first associated signaling is the user training signaling that is sorted before the user signaling time, and the second associated signaling is the user training signaling that is sorted after the user signaling time. Each target positioning mapping relationship can determine the corresponding first associated signaling and second associated signaling. The target user training signaling, the first associated signaling, and the second associated signaling with the same user identifier all correspond to the actual location tag of the target user training signaling.

[0213] Exemplarily, the user signaling start time and the user signaling end time in the user signaling time of the target user training signaling are determined, and the N user training signalings that are sequentially arranged before the user signaling start time are determined in each user training signaling with the same user identifier as the target user training signaling, and the N user training signalings are all used as the first associated signaling, and the M signalings that are sequentially arranged after the user signaling end time are determined, and the M signalings are all used as the second associated signaling; generally, N is the same as M, and N and M are integers. Each target user training signaling has its own corresponding first associated signaling and second associated signaling.

[0214] For a better understanding of this embodiment, please refer to Figure 4, briefly describe the overall process of constructing the preset trajectory reconstruction model, step Y101 obtains each signaling to be trained, each precise positioning data and each training POI to be screened; step Y102 includes determining the training POI, actual location tag and training associated signaling corresponding to each signaling to be trained; step Y103: respectively determine the training space-time coding sequence corresponding to each signaling to be trained, which may specifically include, for each signaling to be trained, determining the training space-time coding vector of the signaling to be trained, and the training space-time coding vector of each training associated signaling associated with the signaling to be trained, splicing the training space-time coding vector of the signaling to be trained, and the training space-time coding vector of each training associated signaling associated with the signaling to be trained, to obtain the training space-time coding sequence of the signaling to be trained. Step Y104: Based on each training space-time coding vector, optimize the model to be trained to obtain the preset trajectory reconstruction model that has been trained. Specifically, you can refer to the description of steps A10 to A50, which will not be repeated here.

[0215] Also refer to Figure 5 , the overall process of this embodiment is briefly described, step X101: training to obtain a preset trajectory reconstruction model; step X102: inputting the spatiotemporal coding sequence of the first signaling into the preset trajectory reconstruction model; step X103: obtaining the actual position of the target output by the preset trajectory reconstruction model.

[0216] Since the embodiment of the present application can improve the accuracy of determining the user's location, the travel chain of the user carrying the mobile terminal can be identified through this embodiment, and the user's location can be identified through signaling, so that the signaling trajectory corresponding to the user can be reconstructed, which can better characterize the user's travel characteristics such as travel distance, travel speed, etc., and help improve the identification of the user's travel mode and travel path, thereby constructing the user's full-process travel chain. Based on the user's full-process travel chain, a large-scale urban traffic survey can be implemented more conveniently and quickly, replacing the traditional questionnaire survey format, and improving the efficiency of the government and relevant departments in gaining insight into the current status of urban traffic.

[0217] The signaling data of mobile terminals has become an important data resource for conducting urban traffic and population analysis. This embodiment can further improve the accuracy of user location, and can realize road-level traffic flow statistics and more refined population distribution analysis, thereby supporting more application scenarios that require higher precision and accuracy of the results.

[0218] The present application also provides a positioning device, please refer to Figure 6 , the device comprises:

[0219] An acquisition module 10 is used to acquire a signaling sequence of a first signaling of a user to be located, wherein the signaling sequence includes a plurality of sub-signals, which are respectively the first signaling and a time-associated signaling that is time-continuous with the first signaling of the first signaling;

[0220] The coding module 20 is used to determine the spatiotemporal coding vector of each sub-signaling according to the obtained POI points of interest and each sub-signaling;

[0221] The positioning prediction module 30 is used to input the spatiotemporal coding sequence constructed by each spatiotemporal coding vector into a preset trajectory reconstruction model to obtain the target actual position of the user to be positioned in the first signaling mapped by the spatiotemporal coding sequence.

[0222] The positioning device provided in the embodiment of the present application adopts the positioning method in the above embodiment to solve the technical problem of low accuracy of user position positioning. Compared with the prior art, the beneficial effects of the positioning method provided in the embodiment of the present application are the same as the beneficial effects of the positioning method provided in the above embodiment, and other technical features in the positioning device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0223] An embodiment of the present application provides an electronic device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the positioning method in the above embodiment.

[0224] Reference below Figure 7 , which shows a schematic diagram of the structure of an electronic device suitable for implementing the embodiments of the present disclosure. The electronic device in the embodiments of the present disclosure may include, but is not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 7 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0225] like Figure 7As shown, the electronic device may include a processing device 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM) 1004. Various programs and data required for the operation of the electronic device are also stored in the RAM 1004. The processing device 1001, the ROM 1002, and the RAM 1004 are connected to each other via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus.

[0226] Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 1003 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 1009. The communication device can allow the electronic device to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows an electronic device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems may be implemented or have alternatively.

[0227] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 1009, or installed from the storage device 1003, or installed from the ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.

[0228] The electronic device provided in the embodiment of the present application adopts the positioning method in the above embodiment 1 to solve the technical problem of low accuracy of user position positioning. Compared with the prior art, the beneficial effect of product flow data distribution provided in the embodiment of the present application is the same as the beneficial effect of the positioning method provided in the above embodiment, and the other technical features in the positioning device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0229] It should be understood that the various parts of the present disclosure can be implemented with hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0230] The above are only specific implementations of the embodiments of the present application, but the protection scope of the embodiments of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or replacements within the technical scope disclosed in the embodiments of the present application, which should be included in the protection scope of the embodiments of the present application. Therefore, the protection scope of the embodiments of the present application shall be based on the protection scope of the claims.

[0231] This embodiment provides a computer-readable storage medium having computer-readable program instructions stored thereon, and the computer-readable program instructions are used to execute the positioning method in the above-mentioned embodiment 1.

[0232] The computer-readable storage medium provided in the embodiment of the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor devices, equipment or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable EPROM (Electrical Programmable Read Only Memory, read-only memory) or flash memory, an optical fiber, a portable compact disk CD-ROM (compact disc read-only memory, read-only memory), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution device, device or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency, radio frequency), etc., or any suitable combination of the above.

[0233] The computer-readable storage medium may be included in the electronic device, or may exist independently without being installed in the electronic device.

[0234] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by an electronic device, the electronic device: obtains a signaling sequence of a first signaling of a user to be located, wherein the signaling sequence includes multiple sub-signals, which are respectively a first signaling and a time-associated signaling that is time-continuous with the first signaling of the first signaling; determines a space-time coding vector of each sub-signaling according to the obtained POI points of interest and each sub-signaling; and inputs the space-time coding sequence constructed by each space-time coding vector into a preset trajectory reconstruction model to obtain the target actual position of the user to be located in the first signaling mapped by the space-time coding sequence.

[0235] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a LAN (local area network) or WAN (Wide Area Network), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0236] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the equipment, method and computer program product according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based device that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0237] The modules involved in the embodiments described in the present disclosure may be implemented by software or hardware, wherein the name of the module does not limit the unit itself in some cases.

[0238] The computer-readable storage medium provided in the embodiment of the present application stores computer-readable program instructions for executing the above positioning method, aiming to solve the technical problem of low accuracy of user position positioning. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in the embodiment of the present application are the same as the beneficial effects of the positioning method provided in the above embodiment, and will not be described in detail here.

[0239] An embodiment of the present application also provides a computer program product, including a computer program, which implements the steps of the positioning method as described above when the computer program is executed by a processor.

[0240] The computer program product provided in the embodiment of the present application is intended to solve the technical problem of low accuracy in positioning the user's position. Compared with the prior art, the beneficial effects of the computer program product provided in the embodiment of the present application are the same as the beneficial effects of the positioning method provided in the above embodiment, which will not be repeated here.

[0241] The above are only preferred embodiments of the embodiments of the present application, and are not intended to limit the patent scope of the embodiments of the present application. Any equivalent structure or equivalent process transformation made using the description and drawings of the embodiments of the present application, or directly or indirectly applied in other related technical fields, are also included in the patent processing scope of the embodiments of the present application.

Claims

1. A positioning method, characterized in that: The method includes: Acquire a signaling sequence of a first signaling of a user to be located, wherein the signaling sequence includes a plurality of sub-signals, which are respectively the first signaling and a time-associated signaling that is time-continuous with the first signaling of the first signaling; Determine the spatiotemporal coding vector of each sub-signaling respectively according to the acquired POI points of interest of each sub-signaling and each sub-signaling; The space-time coding sequence constructed by each of the space-time coding vectors is input into a preset trajectory reconstruction model to obtain the target actual position of the user to be located in the first signaling mapped by the space-time coding sequence.

2. The method according to claim 1, characterized in that The step of determining the spatiotemporal coding vector of each sub-signaling according to the acquired POI points of interest of each sub-signaling and each sub-signaling comprises: For each of the sub-signals, obtaining each POI point of interest of the sub-base station in the sub-signal; Determine a spatial coding vector of the sub-signaling according to the sub-signaling and the corresponding POI points of interest; Determining a time coding vector of the sub-signaling according to the sub-signaling time of the sub-signaling; The spatial coding vector and the temporal coding vector of the same sub-signaling are concatenated to obtain the spatiotemporal coding vector of the sub-signaling.

3. The method according to claim 2, characterized in that The step of determining the spatial coding vector of the sub-signaling according to the sub-signaling and the corresponding POI points of interest includes: Performing a nonlinear change on the sub-signaling to obtain a signaling vector, and performing a nonlinear change on each of the POI points of interest to obtain an interest vector corresponding to each of the POI points of interest; Determine the signaling query parameters of the signaling vector according to a preset attention algorithm, and determine the interest key and interest value corresponding to each of the interest vectors; Calculating the attention weights between the signaling query parameters and each of the interest keys; According to the attention weights of the interest keys, weighted summation is performed on the interest values ​​of the interest vectors where the interest keys are located, to obtain the interest point convergence score of the sub-signaling; The longitude vector and the latitude vector of the sub-signaling are determined from the signaling vector, and the longitude vector, the latitude vector and the interest point convergence score are concatenated to obtain a spatial coding vector.

4. The method according to claim 2, characterized in that The sub-signaling time includes a start time and an end time; The step of determining the time coding vector of the sub-signaling according to the sub-signaling time of the sub-signaling comprises: Calculate the duration of the sub-signaling according to the start time and the end time; Performing nonlinear change on the duration to obtain a time difference vector; Performing nonlinear transformation on the start time and the end time respectively to obtain a first time semantic vector of the start time and a second time semantic vector of the end time; The time difference vector, the first time semantic vector and the second time semantic vector are concatenated to obtain a time coding vector.

5. The method according to claim 1, characterized in that The method includes: Acquire a training sample set, wherein the training sample set includes a plurality of training samples, each of the training samples includes a training signaling sequence, a training POI and an actual location tag, and each training sub-signaling included in the training signaling sequence is a signaling to be trained, and a training associated signaling that is time-continuous with the training signaling of the signaling to be trained; Determining, according to each of the training sub-signals and the respective training POIs, respective training spatiotemporal coding vectors of each of the training sub-signals; Inputting the spatiotemporal coding sequence constructed by each of the training spatiotemporal coding vectors into a preset model to be trained to obtain a first prediction result, wherein the first prediction result includes a predicted position of the signaling to be trained; Calculating a prediction loss according to the first prediction result and the actual position label; The model to be trained is optimized according to the prediction loss to obtain a preset trajectory reconstruction model that has been trained.

6. The method according to claim 5, characterized in that The step of obtaining a training sample set comprises: Acquire each signaling to be trained, each precise positioning data and each training POI to be screened, wherein the training POI to be screened includes the POI position, and the signaling to be trained includes the training base station position; For each of the signalings to be trained, in each of the training POIs to be screened, a target POI position whose first spatial distance between the POI position and the training base station position is less than a preset distance threshold is screened out, and the signaling to be trained is associated with the training POI of the target POI position; According to each of the signalings to be trained and each of the precise positioning data, the actual location tag corresponding to each of the signalings to be trained and the training-related signaling corresponding to each of the signalings to be trained are determined.

7. The method according to claim 6, characterized in that The training-associated signaling includes a first-associated signaling and a second-associated signaling; The step of determining, according to each of the signaling to be trained and each of the precise positioning data, the actual location tag corresponding to each of the signaling to be trained and the training associated signaling corresponding to each of the signaling to be trained comprises: Determine the training signaling of each user and the precise positioning data of each user belonging to the same user identifier according to the respective user identifiers in the signaling to be trained and the respective user identifiers in the precise positioning data; Determine a target positioning mapping relationship in each of the user training signaling and each of the user precise positioning data, wherein the target positioning mapping relationship is a target user training signaling and a target user precise positioning data that are temporally and spatially associated; Using the target user precise positioning data of the target user training signaling in the target positioning mapping relationship as the actual position label of the target user training signaling; For any target positioning mapping relationship, according to the user signaling time of the target user training signaling in the target positioning mapping relationship, among each user training signaling corresponding to the user identifier of the target user training signaling, filter out the first associated signaling whose training signaling time is ranked before the user signaling time and the number thereof is a preset number, and filter out the second associated signaling whose training signaling time is ranked after the user signaling time and the number thereof is a preset number.

8. The method according to claim 7, characterized in that The step of determining the target positioning mapping relationship in each of the user training signaling and each of the user precise positioning data comprises: Determine the respective user signaling time from the respective user training signaling, and determine the respective sampling time from the respective user precise positioning data, wherein the user signaling time includes the user signaling start time and the user signaling end time; Associating the initial user precise positioning data and the initial user training signaling at the sampling time within the user signaling time to obtain an initial positioning mapping relationship; Respectively calculating the distance between the initial user training signaling and the initial user precise positioning data in each of the initial positioning mapping relationships to obtain a second spatial distance corresponding to each of the initial positioning mapping relationships; A target positioning mapping relationship whose second spatial distance is less than a preset distance threshold is selected from each of the initial positioning mapping relationships.

9. The method according to claim 5, characterized in that The step of optimizing the model to be trained according to the prediction loss to obtain a preset trajectory reconstruction model that has been trained includes: Optimizing the model to be trained according to the prediction loss until the prediction loss is less than a preset training error threshold, thereby obtaining an initial trajectory reconstruction model; Verifying the initial trajectory reconstruction model according to the obtained verification set to obtain a verification position output by the initial trajectory reconstruction model; If the verification position does not meet the preset optimal effect conditions, the hyperparameters of the model to be trained are updated, and the model to be trained is retrained until the verification position meets the preset optimal effect conditions to obtain a preset trajectory reconstruction model.

10. A positioning device, characterized in that: The device comprises: An acquisition module, used to acquire a signaling sequence of a first signaling of a user to be located, wherein the signaling sequence includes a plurality of sub-signals, which are respectively the first signaling and a time-associated signaling that is time-continuous with the first signaling of the first signaling; A coding module, configured to determine a spatiotemporal coding vector of each sub-signaling according to the acquired POI points of interest of each sub-signaling and each sub-signaling; The positioning prediction module is used to input the space-time coding sequence constructed by each of the space-time coding vectors into a preset trajectory reconstruction model to obtain the target actual position of the user to be located in the first signaling mapped by the space-time coding sequence.

11. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively coupled to the at least one processor; The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the steps of the positioning method described in any one of claims 1 to 9.

12. A medium, characterized in that The medium is a computer-readable storage medium, on which is stored a program for implementing the positioning method, and the program for implementing the positioning method is executed by a processor to implement the steps of the positioning method as claimed in any one of claims 1 to 9.

13. A program product, characterized in that The program product is a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps of the positioning method according to any one of claims 1 to 9 are implemented.

Citation Information

Cited By

  • Terminal track generation method and device, medium and product

    CN120659013A

  • Terminal trajectory generation method, apparatus, medium, and product

    CN120659013B