Position feature determination method and apparatus, device, storage medium, and program product

Abstract

The application discloses a position feature determination method, device, equipment, a storage medium and a program product. The position feature determination method comprises the following steps: acquiring mobile position information of a mobile object; performing feature extraction on the mobile position information to obtain a base station position vector corresponding to each time step of the mobile object, wherein the mobile position information comprises a plurality of time steps; processing each base station position vector to obtain a hidden state vector corresponding to each time step; and determining a position feature vector according to each hidden state vector. In the application, the extraction efficiency of the position feature is improved.

Description

[0001] Technology Neighborhood

[0002] This application belongs to the field of data processing technology, and particularly relates to a method, apparatus, device, storage medium and program product for determining location features. Background Technology

[0003] The movement trajectory of mobile objects such as terminals and vehicles can help models predict the position of the moving object in the next moment. The movement trajectory can be characterized by the positional features of the moving object.

[0004] Location features contain the movement patterns and location movement rules of moving objects, which can be used to train models, enabling the models to more accurately predict the position of moving objects in the next moment.

[0005] In the exemplary technology, the location features are extracted using manual feature engineering. However, manual extraction of location features suffers from low efficiency. Summary of the Invention

[0006] This application provides a method, apparatus, device, storage medium, and program product for determining location features, in order to solve the problem of low efficiency in extracting location features.

[0007] In a first aspect, embodiments of this application provide a method for determining location features, including:

[0008] Get the movement position information of the moving object;

[0009] Feature extraction is performed on the mobile location information to obtain the base station location vector corresponding to the mobile object at each time step, wherein the mobile location information includes multiple time steps;

[0010] The location vectors of each base station are processed to obtain the hidden state vector corresponding to each time step.

[0011] The location feature vector is determined based on each of the hidden state vectors.

[0012] In one embodiment, the step of determining the location feature vector based on each of the hidden state vectors includes:

[0013] Based on the movement location information, determine the grid corresponding to the moving object at each time step;

[0014] Determine the spatial feature vector of the moving object based on each of the grids;

[0015] The location feature vector is determined based on the spatial feature vector and each of the hidden state vectors.

[0016] In another embodiment, the step of determining the spatial feature vector of the moving object based on each of the grids includes:

[0017] Determine the neighboring grids corresponding to the current grid, and obtain the occupancy information and movement trend information of the neighboring grids;

[0018] Each of the neighborhood grids is clustered to obtain multiple clusters of neighborhood grids, and the neighborhood grid clustering features corresponding to each cluster of neighborhood grids are determined.

[0019] The occupancy characteristics of the neighboring grid are determined based on the occupancy information, and the movement trend characteristics of the neighboring grid are determined based on the movement trend information.

[0020] Based on the occupancy characteristics, movement trend characteristics, and clustering characteristics of the neighboring grid corresponding to the current grid, the feature vector of the neighboring grid corresponding to the current grid is determined.

[0021] The spatial feature vector is determined based on the neighborhood raster feature vector.

[0022] In another embodiment, the step of determining the location feature vector based on the neighborhood grid feature vector includes:

[0023] Obtain the first category distribution features of the interest points corresponding to the current grid, and determine the second category distribution features of the interest points of the neighboring grids corresponding to the current grid;

[0024] Based on the distribution characteristics of the first category and the distribution characteristics of the second category, the feature vector of interest points is determined;

[0025] The spatial feature vector is determined based on the feature vector of the interest point and the feature vector of the neighborhood raster.

[0026] In another embodiment, the step of determining the location feature vector based on the neighborhood grid feature vector includes:

[0027] Obtain the first category attribute of the current raster and determine the second category attribute of the neighboring raster of the current raster;

[0028] Based on the first category attribute and the second category attribute, construct a geographic semantic feature vector;

[0029] The spatial feature vector is determined based on the geographic semantic feature vector and the neighborhood raster feature vector.

[0030] In another embodiment, the step of determining the location feature vector based on each of the hidden state vectors includes:

[0031] The hidden state vector corresponding to the latest time step is determined as the location feature vector.

[0032] In another embodiment, the step of determining the location feature vector based on each of the hidden state vectors includes:

[0033] The average value or weighted sum of each of the hidden state vectors is determined as the location feature vector.

[0034] In another embodiment, the step of processing each of the base station location vectors to obtain the hidden state vector corresponding to each time step includes:

[0035] Obtain the first hidden state vector corresponding to the previous time step;

[0036] Based on the first hidden state vector and the base station location vector corresponding to the current time step, the second hidden state vector corresponding to the current time step is determined.

[0037] In another embodiment, the step of determining the second hidden state vector corresponding to the current time step based on the first hidden state vector and the base station location vector corresponding to the current time step includes:

[0038] Based on the first hidden state vector, the base station location vector corresponding to the current time step, and the first mapping relationship, the input gate parameters are determined;

[0039] Based on the first hidden state vector, the base station location vector corresponding to the current time step, and the second mapping relationship, the forget gate parameters are determined;

[0040] Based on the first hidden state vector, the base station location vector corresponding to the current time step, and the third mapping relationship, the output gate parameters are determined;

[0041] Based on the first hidden state vector, the base station location vector corresponding to the current time step, and the fourth mapping relationship, the state parameters of the candidate memory cells are determined.

[0042] The second hidden state vector corresponding to the current time step is determined based on the input gate parameters, the forget gate parameters, the output gate parameters, and the candidate memory cell state parameters.

[0043] In another embodiment, the step of determining the second hidden state vector corresponding to the current time step based on the first hidden state vector and the base station location vector corresponding to the current time step includes:

[0044] Based on the first hidden state vector, the base station location vector corresponding to the current time step, and the fifth mapping relationship, the update gate parameters are determined.

[0045] Based on the first hidden state vector, the base station location vector corresponding to the current time step, and the sixth mapping relationship, the reset gate parameters are determined.

[0046] Based on the first hidden state vector, the base station location vector corresponding to the current time step, and the seventh mapping relationship, the candidate hidden state parameters are determined.

[0047] The second hidden state vector corresponding to the current time step is determined based on the update gate parameter, the reset gate parameter, and the candidate hidden state parameter.

[0048] Secondly, embodiments of this application provide a location feature determination device, comprising:

[0049] The acquisition module is used to obtain the movement location information of the moving object;

[0050] The extraction module is used to extract features from the mobile location information to obtain the base station location vector corresponding to the mobile object at each time step, wherein the mobile location information includes multiple time steps;

[0051] The processing module is used to process each of the base station location vectors to obtain the hidden state vector corresponding to each time step;

[0052] The determination module is used to determine the position feature vector based on each of the hidden state vectors.

[0053] Thirdly, embodiments of this application provide a location feature determination device, including: a processor and a memory storing computer program instructions;

[0054] When the processor executes the computer program instructions, it implements a position feature determination method as described in one aspect.

[0055] Fourthly, embodiments of this application provide a computer storage medium on which computer program instructions are stored. When the computer program instructions are executed by a processor, they implement the location feature determination method as described in the first aspect.

[0056] Fifthly, embodiments of this application provide a computer program product in which instructions, when executed by a processor of an electronic device, cause the electronic device to perform the position feature determination method as described in the first aspect.

[0057] The location feature determination method, apparatus, device, storage medium, and program product provided in this application acquire the mobile location information of a mobile object, extract features from the mobile location information to obtain the base station location vector corresponding to the mobile object at each time step, process each base station location vector to obtain the hidden state vector corresponding to each time step, and determine the location feature vector based on each hidden state vector. In this application, multiple base station location vectors are obtained in a time series by extracting features from the mobile location information, and the location feature vector is automatically extracted from the hidden state vector obtained by processing each base station location vector, eliminating the need for manual location feature extraction and improving the efficiency of location feature extraction. Attached Figure Description

[0058] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments of this application will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0059] Figure 1 This is a schematic diagram of a scenario in which the location feature determination method of this application can be implemented;

[0060] Figure 2 This is a flowchart illustrating the first embodiment of the location feature determination method of this application;

[0061] Figure 3 This is a flowchart illustrating the second embodiment of the location feature determination method of this application;

[0062] Figure 4 This is a flowchart illustrating the third embodiment of the location feature determination method of this application;

[0063] Figure 5 This is a flowchart illustrating the fourth embodiment of the location feature determination method of this application;

[0064] Figure 6 This is a flowchart illustrating the fifth embodiment of the location feature determination method of this application;

[0065] Figure 7 This is a flowchart illustrating the sixth embodiment of the location feature determination method of this application;

[0066] Figure 8 This is a schematic diagram of the functional modules of the location feature determination device provided in the embodiments of this application;

[0067] Figure 9 This is a schematic diagram of the location feature determination device provided in the embodiments of this application. Detailed Implementation

[0068] The features and exemplary embodiments of various aspects of this application will be described in detail below. To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are intended only to explain this application and not to limit it. For those skilled in the art, this application can be implemented without some of these specific details. The following description of the embodiments is merely to provide a better understanding of this application by illustrating examples.

[0069] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0070] The movement trajectory of mobile objects such as terminals and vehicles can help models predict the position of the moving object in the next moment. The movement trajectory can be characterized by the positional features of the moving object.

[0071] Location features contain the movement patterns and location movement rules of moving objects, which can be used to train models, enabling the models to more accurately predict the position of moving objects in the next moment.

[0072] The inventors of this application have discovered that the location features are extracted using manual feature engineering, which results in low efficiency in the extraction of location features.

[0073] The inventors of this application therefore conceived of obtaining multiple base station location vectors in a time series by performing feature extraction on mobile location information, and then automatically extracting location feature vectors from the hidden state vectors obtained by processing each base station location vector, without the need for manual location feature extraction, thus improving the efficiency of location feature extraction.

[0074] This application provides a method, apparatus, device, storage medium, and program product for determining location features. The following section first describes a framework for applying the location feature determination method provided in this application.

[0075] like Figure 1As shown, the location feature determination device 100 obtains the movement location information of a moving object from the database 200. The device extracts features from the movement location information to obtain the base station location vector corresponding to the moving object at each time step. The device processes the base station location vector to obtain the hidden state vector corresponding to each time step, thereby determining the location feature vector through each hidden state vector. The device outputs the location feature vector; for example, it transmits the location feature vector to the model training device 300 so that the model training device 300 can train a model to predict the location of other moving objects based on the location feature vector.

[0076] The following combination Figure 1 The method for determining location features provided in the embodiments of this application will be introduced.

[0077] Figure 2 A flowchart illustrating a first embodiment of the location feature determination method provided in this application is shown. Figure 2 As shown, the location feature determination methods include:

[0078] Step S201: Obtain the movement location information of the moving object.

[0079] In this embodiment, the executing entity is a location feature determination device. For ease of description, the term "device" will be used to refer to the location feature determination device below. The device can be a server or any terminal device with location feature extraction capabilities.

[0080] The device acquires the location information of a moving object. The moving object can be a user terminal or vehicle-mounted terminal connected to a base station signal and in motion. Location information refers to the information comprised of the moving object's location at different time steps. Each time step is a timestamp, and location information includes the location at different time steps; that is, location information includes multiple time steps.

[0081] Step S202: Feature extraction is performed on the mobile location information to obtain the base station location vector corresponding to the mobile object at each time step. The mobile location information includes multiple time steps.

[0082] After obtaining the mobile location information, feature extraction is performed on the mobile location information to obtain the base station location vector corresponding to the mobile object at each time step. For example, the device obtains the base station ID or base station location coordinates from the record corresponding to each timestamp, performs one-hot encoding on the base station ID, and converts the base station location coordinates into real-valued vectors. Each base station location vector is represented as follows: ,in Let represent the base station location vector at time step t.

[0083] Furthermore, after obtaining the location information, the data is cleaned and preprocessed, specifically to remove noise and outliers, as follows:

[0084] Remove duplicate records: Delete duplicate data from the location information;

[0085] Remove invalid records: Remove records that are missing key fields or are in abnormal positions;

[0086] Timestamp normalization: converting timestamps of different formats or time zones into a unified timestamp representation;

[0087] Location coordinate denoising: The location coordinates of the base station are smoothed to remove high-frequency noise.

[0088] Step S203: Process the location vectors of each base station to obtain the hidden state vector corresponding to each time step.

[0089] After obtaining the location vectors of each base station, the location vectors of each base station are processed to obtain the hidden state vector corresponding to each time step.

[0090] For example, the location vectors of each base station As input, the data is fed into a recurrent neural network for feature extraction, thereby obtaining the hidden state vector corresponding to each time step. The hidden state vector contains the temporal dependence and contextual information of the trajectory sequence up to the current time step.

[0091] Step S204: Determine the position feature vector based on each hidden state vector.

[0092] After obtaining each hidden state vector, the location feature vector is determined based on each hidden state vector.

[0093] In one example, the hidden state vector corresponding to the latest time step is used as the location feature vector.

[0094] In another instance, the average or weighted sum of the individual hidden state vectors is used to determine the location feature vector.

[0095] In this embodiment, the location feature vector is the base station zipper trajectory feature vector, which contains the movement pattern and location transfer rules of the moving object at the base station level.

[0096] In this embodiment, the movement location information of a mobile object is acquired, and feature extraction is performed on the movement location information to obtain the base station location vector corresponding to the mobile object at each time step. Then, each base station location vector is processed to obtain the hidden state vector corresponding to each time step, and the location feature vector is determined based on each hidden state vector. In this embodiment, multiple base station location vectors are obtained from the movement location information over a time series by feature extraction. The location feature vector is then automatically extracted from the hidden state vectors obtained by processing each base station location vector, eliminating the need for manual location feature extraction and improving the efficiency of location feature extraction.

[0097] Reference Figure 3 , Figure 3 The flowchart of the second embodiment of the location feature determination method of this application is shown. Based on the first embodiment, step S204 includes:

[0098] Step S301: Determine the grid corresponding to the moving object at each time step based on the moving position information.

[0099] In this embodiment, the constructed location feature vector is used to train the model, therefore the location feature vector needs to be extracted accurately.

[0100] To address this, the device determines the corresponding grid for each time step of the moving object based on its location information. Specifically, the area on the device's map is divided into multiple grids, each defined as a grid, and the moving object moves between these grids.

[0101] The device can map MR (Measurement Report) fingerprint positioning sequences (mobility location information) into raster sequences. ,in This represents the grid ID where the moving object is located at time step t.

[0102] Step S302: Determine the spatial feature vector of the moving object based on each grid.

[0103] A raster ID is a unique identifier for a raster. A unique identifier can be, for example, the row and column number of the raster or the hash value of the row and column number.

[0104] Each raster ID is mapped to a fixed-length embedding vector, represented as follows: ,in It is a raster embedding matrix. It is the one-hot vector representation of the raster ID at time step t, and the raster embedding vector. This represents the feature representation of the grid where the user is located at time step t. After obtaining the feature vectors of each grid, the device can determine the spatial feature vector based on these feature vectors. For example, the feature vectors of each grid are fused to obtain a fused feature. Since the fused feature is obtained by fusing grid feature vectors, and grid feature vectors have spatial properties, the fused feature can be used as a spatial feature vector.

[0105] Step S303: Determine the location feature vector based on the spatial feature vector and each hidden state vector.

[0106] After obtaining the spatial feature vector, the device constructs a temporal feature vector based on each hidden state vector. The spatial and temporal feature vectors are then fused to obtain the location feature vector. For example, fusing the hidden state vectors yields the temporal feature vector.

[0107] For example, each hidden state vector is represented as The device performs pooling operations on each hidden state vector to obtain a fixed-length temporal feature vector, which represents... Common pooling operations include max pooling and average pooling. These operations are used to transform temporal feature vectors. By concatenating the spatial feature vector with the location feature vector, we can obtain the positional feature vector. This positional feature vector is the MR fingerprint localization feature vector.

[0108] In this embodiment, the location feature vector can be used for subsequent construction of a large location model and location sensing applications. Location sensing applications are used in offline scenarios, such as commuting route planning. The aforementioned use of the hidden state vector at the latest time step, or the average or weighted sum of all hidden state vectors as the location feature vector, is applied to real-time scenarios, such as tracking moving objects.

[0109] In this embodiment, the location feature vector includes multi-dimensional information such as the temporal dependence and spatial relationship of the moving object's location, which can comprehensively characterize the location behavior pattern of the moving object. The location feature vector can be used for the subsequent construction of a large location model and location sensing applications.

[0110] Reference Figure 4 , Figure 4 This paper illustrates a flowchart of a third embodiment of the location feature determination method provided in this application, based on... Figure 3 In the embodiment shown, step S302 includes:

[0111] Step S401: Determine the neighboring grids corresponding to the current grid, and obtain the occupancy information and movement trend information of the neighboring grids.

[0112] In this embodiment, each grid has adjacent grids, which are defined as neighborhood grids.

[0113] Define the neighborhood range as the surrounding grid cells centered on the currently located grid cell of the moving object. There are N grid cells, where N is an odd number. The neighborhood range can be represented as: ,in This represents the grid cell where the object is located at time step t. Relative to The (i,j)th neighboring raster.

[0114] The location information includes the occupancy status and movement trend of each grid cell. Occupancy status refers to whether the current grid cell is occupied by a moving object, and movement trend refers to the trend of the moving object moving from the current grid cell to neighboring grid cells. The device obtains the occupancy information and movement trend information of neighboring grid cells from the location information.

[0115] Step S402: Cluster each neighborhood grid to obtain multiple clusters of neighborhood grids, and determine the neighborhood grid clustering features corresponding to each cluster of neighborhood grids.

[0116] After identifying the neighborhood grids, the device clusters these grids. Specifically, it groups grids with similar functions or behavioral patterns into a single category. Functional similarity could be based on the grid's location being suitable for parking, while behavioral similarity could be based on the presence of moving objects within the grid.

[0117] The device clusters each neighborhood grid to obtain multiple clusters of neighborhood grids and determines the neighborhood grid clustering features corresponding to each cluster of neighborhood grids.

[0118] Neighborhood grid clustering features can be represented as ,in This represents a clustering function that maps neighborhood raster clustering features to corresponding category labels.

[0119] Step S403: Determine the occupancy characteristics of the neighboring grid based on the occupancy information, and determine the movement trend characteristics of the neighboring grid based on the movement trend information.

[0120] After obtaining occupancy information, the device determines the occupancy characteristics of neighboring grid cells based on this information. Specifically, for each neighboring grid cell... Calculate the occupancy characteristics of neighboring grid cells This indicates whether the grid is occupied by a user at the current time step.

[0121] The neighborhood grid occupancy feature can be represented as ,in This is an indicator function that takes the value 1 when a certain condition is met, and 0 otherwise. For example, it is 1 when the neighboring raster is occupied, and 0 when it is not occupied.

[0122] The device can also determine the movement trend characteristics of neighboring grids based on movement trend information. Specifically, for each neighboring grid... Calculate its movement trend characteristics This indicates the tendency for a moving object to move from the current grid cell to a neighboring grid cell.

[0123] The movement trend characteristics of neighboring grids can be calculated by statistically analyzing the frequency with which moving objects move from the current grid to the neighboring grid in historical trajectories. ,in Indicates that the moving object is from the grid. Move to grid The number of times.

[0124] Step S404: Determine the feature vector of the neighboring grid corresponding to the current grid based on the occupancy characteristics, movement trend characteristics, and clustering characteristics of the neighboring grid.

[0125] After obtaining the neighborhood grid occupancy features, neighborhood grid movement trend features, and neighborhood grid clustering features, the device determines the neighborhood grid feature vector based on these features.

[0126] Specifically, the neighborhood grid occupancy features, neighborhood grid movement trend features, and neighborhood grid clustering features are combined into a single feature vector, which is represented as follows: The current grid has multiple neighboring grids. The feature vectors of each neighboring grid corresponding to the current grid are concatenated to obtain the neighboring grid feature vector. .

[0127] Step S405: Determine the spatial feature vector based on the neighborhood raster feature vector.

[0128] After obtaining the feature vectors of the neighboring raster cells, the spatial feature vector is determined based on the feature vectors of each neighboring raster cell. For example, the spatial feature vector can be obtained by concatenating the feature vectors of the neighboring raster cells corresponding to each raster cell.

[0129] In this embodiment, the device acquires multiple features of the neighborhood grid based on the current grid, thereby acquiring spatial features with precise spatial relationships based on each feature.

[0130] Reference Figure 5 , Figure 5 This paper illustrates a flowchart of a fourth embodiment of the location feature determination method provided in this application, based on... Figure 4In the embodiment shown, step S405 includes:

[0131] Step S501: Obtain the first category distribution features of the interest points corresponding to the current grid, and determine the second category distribution features of the interest points of the neighboring grids corresponding to the current grid.

[0132] In this embodiment, for each grid cell, the distribution characteristics of POI (Point of Interest) categories within its neighborhood need to be considered. A POI includes information such as name, category, coordinates, and classification. POIs are pre-marked on a map; after the map is divided into grids, each grid cell has corresponding POI category information. Therefore, the device obtains the first category distribution characteristics of the POI corresponding to the current grid cell and determines the second distribution characteristics of the POIs in the neighboring grid cells corresponding to the current grid cell. The device can directly obtain the category distribution characteristics from the grid information corresponding to the current grid cell.

[0133] Step S502: Determine the feature vector of interest points based on the distribution characteristics of the first category and the distribution characteristics of the second category.

[0134] After obtaining the distribution features of the first category and the distribution features of the second category, the feature vector of the interest point can be determined.

[0135] For example, the device will generate a feature vector of POI category distribution within the grid. and neighborhood POI category distribution feature vector The features are fused to obtain a comprehensive POI category distribution feature vector. The fusion method employs an attention mechanism. ,in For weight fusion.

[0136] The device normalizes the comprehensive POI category distribution feature vector to obtain the interest point feature vector. The interest point feature vector satisfies certain numerical range and distribution requirements. For example, the normalization is as follows: ,in This represents the L2 norm.

[0137] Step S503: Determine the spatial feature vector based on the feature vector of the interest point and the feature vector of the neighboring raster.

[0138] After obtaining the feature vectors of interest points, spatial feature vectors can be constructed based on the feature vectors of interest points and the feature vectors of neighboring grids. For example, the spatial feature vector can be obtained by concatenating the feature vectors of interest points and neighboring grids corresponding to each grid.

[0139] In this embodiment, the feature vector of the point of interest is determined based on the category distribution characteristics of the point of interest in the current grid and the category distribution characteristics of the neighboring grids. Based on the feature vector of the point of interest and the feature vector of the neighboring grids, spatial features with accurate spatial relationships are obtained.

[0140] Reference Figure 6 , Figure 6 This paper illustrates a flowchart of the fifth embodiment of the location feature determination method provided in this application, based on... Figure 4 In the embodiment shown, step S405 includes:

[0141] Step S601: Obtain the first category attribute of the current raster and determine the second category attribute of the neighboring raster of the current raster.

[0142] In this embodiment, the spatial feature vector needs to consider geographic semantic features. The device obtains the first category attribute of the current raster and the second category attribute of the neighboring raster corresponding to the current raster.

[0143] Specifically, define a set of geographic semantic categories. Where M represents the total number of geographic semantic categories. Geographic semantic categories can include commercial areas, residential areas, industrial areas, parks, schools, etc. For example, if the current raster's location on the map is in a commercial area, then the first category attribute corresponds to the semantic category corresponding to commercial areas; if the neighboring raster's location on the map is in an industrial area, then the second category attribute corresponds to the semantic category corresponding to industrial areas.

[0144] For example, for each raster g, it is mapped to a corresponding geographic semantic category using a mapping function f to obtain a first category attribute and a second category attribute. The mapping function can be represented as follows: ,in , represents the geographic semantic category to which raster g belongs. The mapping function f can be constructed based on features such as the composition of POIs, demographics, and land use of the raster, through methods such as rules, clustering, and classification.

[0145] Step S602: Construct a geographic semantic feature vector based on the first category attribute and the second category attribute.

[0146] For each raster g, construct an M-dimensional geographic semantic feature vector. This indicates the geographic semantic category to which the raster belongs. The geographic semantic feature vector is represented using one-hot encoding. The i-th element being 1 indicates that the raster g belongs to the geographic semantic category. Other elements have a value of 0. The element represents the first category attribute.

[0147] For each raster g, consider the geographic semantic features within its neighborhood. Define the neighborhood of raster g as... The geographic semantic feature vector of the raster within the neighborhood is Aggregate the geographic semantic feature vectors of all grid cells within the neighborhood to obtain the neighborhood geographic semantic feature vector. ,in This indicates the number of grid cells in the neighborhood.

[0148] Geographic semantic feature vectors of the raster and neighborhood geographic semantic feature vector The data is then fused to obtain comprehensive geographic semantic features. The fusion method employs an attention mechanism. ,in For weight fusion.

[0149] The fused geographic semantic features are then embedded by mapping the high-dimensional one-hot encoded vectors to a low-dimensional dense vector space to obtain the final geographic semantic feature vectors. Embedding can be achieved by training a feedforward neural network or using a pre-trained word embedding model, for example... ,in For the embedding matrix, This is the embedded geographic semantic feature vector.

[0150] Step S603: Determine the spatial feature vector based on the geographic semantic feature vector and the neighborhood raster feature vector.

[0151] After obtaining the geographic semantic feature vector, the spatial feature vector can be obtained by concatenating the geographic semantic feature vector with the neighborhood raster feature vector.

[0152] In addition, the spatial feature vector can be obtained by concatenating the semantic feature vector, the neighborhood grid feature vector, and the interest point feature vector.

[0153] In this embodiment, a geographic semantic feature vector is determined based on the category attributes of the raster and the category attributes of the neighborhood raster. Then, spatial features with precise spatial relationships are obtained based on the semantic feature vector and the feature vector of the neighborhood raster.

[0154] Reference Figure 7 , Figure 7 This paper illustrates a flowchart of the sixth embodiment of the location feature determination method provided in this application, based on... Figures 2 to 6 In any of the embodiments shown, step S203 includes:

[0155] Step S701: Obtain the first hidden state vector corresponding to the previous time step.

[0156] Step S702: Determine the second hidden state vector corresponding to the current time step based on the first hidden state vector and the base station location vector corresponding to the current time step.

[0157] In this embodiment, the device obtains the first hidden state vector corresponding to the previous time step, and then determines the second hidden state vector corresponding to the current time step based on the first hidden state vector and the base station location vector corresponding to the current time step. The previous time step and the current time step are adjacent timestamps, and the previous time step is earlier than the current time step.

[0158] In one example, the device determines the input gate parameters based on a first hidden state vector, the base station location vector corresponding to the current time step, and a first mapping relationship.

[0159] Input gate parameters , It is the sum of the base station location vectors at the current time step. It is the hidden state of the previous time step. , and Since is a constant, the above formula constitutes the first mapping relationship.

[0160] The device determines the forget gate parameters based on the first hidden state vector, the base station location vector corresponding to the current time step, and the second mapping relationship. Forget gate parameters

[0161] , , and Since is a constant, the above formula constitutes the second mapping relationship.

[0162] The device determines the output gate parameters based on the first hidden state vector, the base station location vector corresponding to the current time step, and the third mapping relationship. Output gate parameters , , and Since it is a constant, the above formula constitutes the third mapping relationship.

[0163] The device determines the state parameters of candidate memory cells based on the first hidden state vector, the base station location vector corresponding to the current time step, and the fourth mapping relationship.

[0164] Select memory cell state parameters , Since it is a constant, the above formula constitutes the fourth mapping relationship.

[0165] The device determines the second hidden state vector corresponding to the current time step based on the input gate parameters, forget gate parameters, output gate parameters, and candidate memory cell state parameters. Specifically, the memory cell state parameter C is determined based on the input gate parameters, forget gate parameters, and candidate memory cell state parameters. The device determines the second hidden state vector using the output gate parameters and the memory cell state parameters. The second hidden state vector... .

[0166] In another example, the device determines the update gate parameters based on the first hidden state vector, the base station location vector corresponding to the current time step, and the fifth mapping relationship. For example, the update gate parameters... ,in, , Since is a constant, the formula formed is the fifth mapping relation.

[0167] The device determines the reset gate parameters based on the first hidden state vector, the base station location vector corresponding to the current time step, and the sixth mapping relationship.

[0168] For example, resetting the gate parameters ,in, , Since is a constant, the formula formed is the sixth mapping relation.

[0169] The device determines candidate hidden state parameters based on the first hidden state vector, the base station location vector corresponding to the current time step, and the seventh mapping relationship.

[0170] For example, candidate hidden state parameters , Since is a constant, the formula formed is the seventh mapping relation.

[0171] The device determines the second hidden state vector corresponding to the current time step based on the updated gate parameters, the reset gate parameters, and the candidate hidden state parameters.

[0172] For example, the second hidden state parameter .

[0173] In this embodiment, the device can accurately determine the second hidden state vector corresponding to the current time step based on the first hidden state parameter of the previous time step and the base station location vector corresponding to the current time step.

[0174] Based on the above embodiments, this application has the following advantages:

[0175] 1. In view of the spatiotemporal characteristics of mobile location information, a gridded spatiotemporal partitioning (grid) and data preprocessing method is proposed to transform the original discrete trajectory data into regular time series data, which reduces data redundancy and noise and improves the efficiency and accuracy of location feature extraction.

[0176] 2. Gated recurrent units (such as the input gates and output gates mentioned above) adaptively learn temporal features to mine the temporal dependencies and regional association patterns of moving objects' trajectories. Compared with traditional static feature extraction methods, they have stronger feature representation and generalization capabilities.

[0177] 3. An end-to-end spatiotemporal feature extraction framework is proposed, which realizes automated extraction and optimization from raw location data to high-level feature representation. This framework is universal and scalable, and can be flexibly embedded into different data processing, feature learning, and model optimization modules to adapt to different data formats and application scenarios.

[0178] The method proposed in this application is expected to break through the bottlenecks of traditional location big data analysis, fully explore the value of spatiotemporal trajectory data, and empower operators' intelligent operation and refined management. At the same time, this method can also be applied to location feature extraction in fields such as transportation, e-commerce, and public safety.

[0179] like Figure 8 As shown, this application embodiment also provides a location feature determination device 800, which includes:

[0180] The acquisition module 810 is used to acquire the movement location information of the moving object;

[0181] The extraction module 820 is used to extract features from the mobile location information to obtain the base station location vector corresponding to the mobile object at each time step. The mobile location information includes multiple time steps.

[0182] The processing module 830 is used to process the location vectors of each base station to obtain the hidden state vector corresponding to each time step;

[0183] The determination module 840 is used to determine the position feature vector based on each hidden state vector.

[0184] In one embodiment, the location feature determination device 800 is further configured to:

[0185] Based on the location information, determine the grid corresponding to the moving object at each time step;

[0186] Determine the spatial feature vector of the moving object based on each grid cell;

[0187] The location feature vector is determined based on the spatial feature vector and each hidden state vector.

[0188] In one embodiment, the location feature determination device 800 is further configured to:

[0189] Determine the neighboring rasters corresponding to the current raster, and obtain the occupancy information and movement trend information of the neighboring rasters;

[0190] Clustering of each neighborhood grid results in multiple neighborhood grids, and the clustering features of each neighborhood grid are determined.

[0191] The occupancy characteristics of the neighboring grid are determined based on the occupancy information, and the movement trend characteristics of the neighboring grid are determined based on the movement trend information.

[0192] Based on the occupancy characteristics, movement trend characteristics, and clustering characteristics of the neighboring grid corresponding to the current grid, determine the feature vector of the neighboring grid corresponding to the current grid.

[0193] Determine the spatial feature vector based on the feature vector of the neighborhood raster.

[0194] In one embodiment, the location feature determination device 800 is further configured to:

[0195] Obtain the first category distribution features of the interest points corresponding to the current raster, and determine the second category distribution features of the interest points of the neighboring raster corresponding to the current raster;

[0196] Based on the distribution characteristics of the first category and the distribution characteristics of the second category, the feature vector of the interest point is determined;

[0197] The spatial feature vector is determined based on the feature vector of the interest point and the feature vector of the neighborhood raster.

[0198] In one embodiment, the location feature determination device 800 is further configured to:

[0199] Get the first category attribute of the current raster and determine the second category attribute of the neighboring raster of the current raster;

[0200] Construct a geographic semantic feature vector based on the first category attribute and the second category attribute;

[0201] The spatial feature vector is determined based on the geographic semantic feature vector and the neighborhood raster feature vector.

[0202] In one embodiment, the location feature determination device 800 is further configured to:

[0203] The hidden state vector corresponding to the latest time step is determined as the location feature vector.

[0204] In one embodiment, the location feature determination device 800 is further configured to:

[0205] Determine the average or weighted sum of each hidden state vector as the location feature vector.

[0206] In one embodiment, the location feature determination device 800 is further configured to:

[0207] Obtain the first hidden state vector corresponding to the previous time step;

[0208] Based on the first hidden state vector and the base station location vector corresponding to the current time step, determine the second hidden state vector corresponding to the current time step.

[0209] In one embodiment, the location feature determination device 800 is further configured to:

[0210] The input gate parameters are determined based on the first hidden state vector, the base station location vector corresponding to the current time step, and the first mapping relationship.

[0211] Based on the first hidden state vector, the base station location vector corresponding to the current time step, and the second mapping relationship, the forget gate parameters are determined.

[0212] The output gate parameters are determined based on the first hidden state vector, the base station location vector corresponding to the current time step, and the third mapping relationship.

[0213] Based on the first hidden state vector, the base station location vector corresponding to the current time step, and the fourth mapping relationship, the state parameters of the candidate memory cells are determined.

[0214] The second hidden state vector corresponding to the current time step is determined based on the input gate parameters, forget gate parameters, output gate parameters, and candidate memory cell state parameters.

[0215] In one embodiment, the location feature determination device 800 is further configured to:

[0216] Based on the first hidden state vector, the base station location vector corresponding to the current time step, and the fifth mapping relationship, the update gate parameters are determined.

[0217] Based on the first hidden state vector, the base station location vector corresponding to the current time step, and the sixth mapping relationship, the reset gate parameters are determined;

[0218] Based on the first hidden state vector, the base station location vector corresponding to the current time step, and the seventh mapping relationship, the candidate hidden state parameters are determined.

[0219] Based on the updated gate parameters, reset gate parameters, and candidate hidden state parameters, determine the second hidden state vector corresponding to the current time step.

[0220] Figure 9 A schematic diagram of the hardware structure of the location feature determination device provided in an embodiment of this application is shown.

[0221] The location feature determination device may include a processor 901 and a memory 902 storing computer program instructions.

[0222] Specifically, the processor 901 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0223] Memory 902 may include mass storage for data or instructions. For example, and not limitingly, memory 902 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 902 may include removable or non-removable (or fixed) media. Where appropriate, memory 902 may be internal or external to the integrated gateway disaster recovery device. In a particular embodiment, memory 902 is non-volatile solid-state memory.

[0224] In a particular embodiment, memory 902 may include read-only memory (ROM), random access memory (RAM), disk storage media device, optical storage media device, flash memory device, electrical, optical, or other physical / tangible memory storage device. Thus, generally, memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method according to one aspect of this disclosure.

[0225] The processor 901 reads and executes computer program instructions stored in the memory 902 to implement any of the location feature determination methods in the above embodiments.

[0226] In one example, the location feature determination device may further include a communication interface 903 and a bus 910. Wherein, as... Figure 9 As shown, the processor 901, memory 902, and communication interface 903 are connected through bus 910 and complete communication with each other.

[0227] The communication interface 903 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.

[0228] Bus 910 includes hardware, software, or both, that couples components of an online data traffic metering device together. For example, and not limitingly, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 910 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, any suitable bus or interconnect is contemplated herein.

[0229] Furthermore, in conjunction with the terminal device motion and position feature determination methods in the above embodiments, this application embodiment can provide a computer storage medium for implementation. This computer storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement any of the position feature determination methods in the above embodiments.

[0230] This application also provides a computer program product, including a computer program, which, when executed, implements any of the location feature determination methods described in the above embodiments.

[0231] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0232] The functional blocks shown in the above block diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. Programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0233] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0234] The aspects of this disclosure have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable terminal device motion device to produce a machine such that these instructions, executable via the processor of the computer or other programmable terminal device motion device, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowchart illustrations, and combinations of blocks in the block diagrams and / or flowchart illustrations, can also be implemented by special-purpose hardware performing the specified functions or actions, or can be implemented by a combination of special-purpose hardware and computer instructions.

[0235] The above are merely specific embodiments of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the scope of the technology disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application.

Claims

1. A method for determining location features, characterized in that, include: Get the movement position information of the moving object; Feature extraction is performed on the mobile location information to obtain the base station location vector corresponding to the mobile object at each time step, wherein the mobile location information includes multiple time steps; The location vectors of each base station are processed to obtain the hidden state vector corresponding to each time step. Determine the location feature vector based on each of the hidden state vectors; The step of determining the position feature vector based on each of the hidden state vectors includes: Based on the movement location information, determine the grid corresponding to the moving object at each time step; Determine the spatial feature vector of the moving object based on each of the grids; The location feature vector is determined based on the spatial feature vector and each of the hidden state vectors.

2. The location feature determination method according to claim 1, characterized in that, The step of determining the spatial feature vector of the moving object based on each of the grids includes: Determine the neighboring grids corresponding to the current grid, and obtain the occupancy information and movement trend information of the neighboring grids; Each of the neighborhood grids is clustered to obtain multiple clusters of neighborhood grids, and the neighborhood grid clustering features corresponding to each cluster of neighborhood grids are determined. The occupancy characteristics of the neighboring grid are determined based on the occupancy information, and the movement trend characteristics of the neighboring grid are determined based on the movement trend information. Based on the occupancy characteristics, movement trend characteristics, and clustering characteristics of the neighboring grid corresponding to the current grid, the feature vector of the neighboring grid corresponding to the current grid is determined. The spatial feature vector is determined based on the neighborhood raster feature vector.

3. The location feature determination method according to claim 2, characterized in that, The step of determining the location feature vector based on the neighborhood grid feature vector includes: Obtain the first category distribution features of the interest points corresponding to the current grid, and determine the second category distribution features of the interest points of the neighboring grids corresponding to the current grid; Based on the distribution characteristics of the first category and the distribution characteristics of the second category, the feature vector of interest points is determined; The spatial feature vector is determined based on the feature vector of the interest point and the feature vector of the neighborhood raster.

4. The location feature determination method according to claim 2, characterized in that, The step of determining the location feature vector based on the neighborhood grid feature vector includes: Obtain the first category attribute of the current raster and determine the second category attribute of the neighboring raster of the current raster; Based on the first category attribute and the second category attribute, construct a geographic semantic feature vector; The spatial feature vector is determined based on the geographic semantic feature vector and the neighborhood raster feature vector.

5. The method for determining location features according to claim 1, characterized in that, The step of determining the position feature vector based on each of the hidden state vectors includes: The hidden state vector corresponding to the latest time step is determined as the location feature vector.

6. The method for determining location features according to claim 1, characterized in that, The step of determining the position feature vector based on each of the hidden state vectors includes: The average value or weighted sum of each of the hidden state vectors is determined as the location feature vector.

7. The method for determining location features according to any one of claims 1-6, characterized in that, The step of processing each of the base station location vectors to obtain the hidden state vector corresponding to each time step includes: Obtain the first hidden state vector corresponding to the previous time step; Based on the first hidden state vector and the base station location vector corresponding to the current time step, the second hidden state vector corresponding to the current time step is determined.

8. The method for determining location features according to claim 7, characterized in that, The step of determining the second hidden state vector corresponding to the current time step based on the first hidden state vector and the base station location vector corresponding to the current time step includes: Based on the first hidden state vector, the base station location vector corresponding to the current time step, and the first mapping relationship, the input gate parameters are determined; Based on the first hidden state vector, the base station location vector corresponding to the current time step, and the second mapping relationship, the forget gate parameters are determined; Based on the first hidden state vector, the base station location vector corresponding to the current time step, and the third mapping relationship, the output gate parameters are determined; Based on the first hidden state vector, the base station location vector corresponding to the current time step, and the fourth mapping relationship, the state parameters of the candidate memory cells are determined. The second hidden state vector corresponding to the current time step is determined based on the input gate parameters, the forget gate parameters, the output gate parameters, and the candidate memory cell state parameters. The first mapping relationship is as follows: in, It is the sum of the base station location vectors at the current time step. It is the hidden state of the previous time step. , and It is a constant; The second mapping relationship is: in, , and It is a constant; The third mapping relationship is as follows: in, , and It is a constant; The fourth mapping relationship is: in, It is a constant.

9. The method for determining location features according to claim 7, characterized in that, The step of determining the second hidden state vector corresponding to the current time step based on the first hidden state vector and the base station location vector corresponding to the current time step includes: Based on the first hidden state vector, the base station location vector corresponding to the current time step, and the fifth mapping relationship, the update gate parameters are determined. Based on the first hidden state vector, the base station location vector corresponding to the current time step, and the sixth mapping relationship, the reset gate parameters are determined; Based on the first hidden state vector, the base station location vector corresponding to the current time step, and the seventh mapping relationship, the candidate hidden state parameters are determined. Based on the update gate parameter, the reset gate parameter, and the candidate hidden state parameter, determine the second hidden state vector corresponding to the current time step; The fifth mapping relationship is as follows: in, , It is a constant; The sixth mapping relationship is as follows: in, , It is a constant; The seventh mapping relationship is as follows: in, It is a constant.

10. A location feature determination device, characterized in that, include: The acquisition module is used to obtain the movement location information of the moving object; The extraction module is used to extract features from the mobile location information to obtain the base station location vector corresponding to the mobile object at each time step, wherein the mobile location information includes multiple time steps; The processing module is used to process each of the base station location vectors to obtain the hidden state vector corresponding to each time step; The determination module is used to determine the position feature vector based on each of the hidden state vectors; The determining module is specifically used for: Based on the movement location information, determine the grid corresponding to the moving object at each time step; Determine the spatial feature vector of the moving object based on each of the grids; The location feature vector is determined based on the spatial feature vector and each of the hidden state vectors.

11. A location feature determination device, characterized in that, include: Processor and memory storing computer program instructions; When the processor executes the computer program instructions, it implements the position feature determination method as described in any one of claims 1-9.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer program instructions that, when executed by a processor, implement the location feature determination method as described in any one of claims 1-9.

13. A computer program product, characterized in that, When the instructions in the computer program product are executed by the processor of the electronic device, the electronic device is able to perform the position feature determination method as described in any one of claims 1-9.

Images