Indoor and outdoor determination method and device, model training method and device

Abstract

The present disclosure provides an indoor and outdoor determination method and device, a model training method and device, and relates to the field of big data processing technology. The indoor and outdoor determination method includes: obtaining multiple MR data of a first user within at least one first time window; constructing a first matrix including a first element, the first element including RSRPs of multiple service cells in each MR data, and the multiple service cells include a main service cell and multiple neighboring service cells; constructing a second matrix including a second element, the second element including at least one of a first group and a second group of elements, the first group of elements including a first distance proportional to a TA in each MR data, and the second group of elements including a second distance between the main service cell in each MR data and the main service cell in the earliest MR data; and inputting input parameters into a trained machine learning model to determine whether the first user is located indoors or outdoors in each first time window, the input parameters including the first and second matrices.

Description

Technical Field

[0001] The present disclosure relates to the technical field of big data processing, and in particular to an indoor and outdoor determination method and device, and a model training method and device. Background Art

[0002] With the rapid development of mobile communication technology, there is an urgent need to carry out network planning and optimization, formulate accurate network optimization or base station coverage plans, improve network quality, and thus enhance user satisfaction.

[0003] In the related art, indoor / outdoor characteristics of users are distinguished through a single measurement report (MR) data collected at the base station side. Summary of the invention

[0004] The inventors note that in actual environments, due to the influence of various factors such as path loss, wall penetration loss, and signal refraction, the same signal strength may exist in multiple different geographical locations. Similarly, the signal strength at the same geographical location is not always constant and may fluctuate. Therefore, it is not accurate to determine whether a user is indoors or outdoors based on a single piece of MR data.

[0005] In order to solve the above problems, the embodiments of the present disclosure propose the following solutions.

[0006] According to one aspect of an embodiment of the present disclosure, a method for determining indoors and outdoors is provided, including: obtaining multiple measurement report MR data of a first user within at least one first time window; constructing a first matrix including a first element, the first element including the reference signal received power RSRP of multiple service cells in each MR data, the multiple service cells including a main service cell and multiple neighboring service cells; constructing a second matrix including a second element, the second element including at least one of a first group of elements and a second group of elements, the first group of elements including a first distance proportional to a maximum time advance TA in each MR data, the second group of elements including a second distance between the main service cell in each MR data and the main service cell in the earliest MR data among the multiple MR data; and inputting input parameters into a trained machine learning model to determine whether the first user is located indoors or outdoors in each first time window, the input parameters including the first matrix and the second matrix.

[0007] According to another aspect of an embodiment of the present disclosure, a model training method is provided, including: acquiring multiple groups of sample MR data, wherein each group of sample MR data includes multiple sample MR data and a first label corresponding to each sample MR data, the first label being indoor or outdoor; determining a second label corresponding to the group of sample MR data according to the first label corresponding to each sample MR data in each group of sample MR data, the second label being indoor or outdoor; constructing a first sample matrix for each group of sample MR data, the first sample matrix including a first sample element, the first sample element including a reference signal received power of multiple service cells in each sample MR data in the group of sample MR data, the multiple service cells including a primary service cell and multiple neighboring service cells; constructing a second sample matrix for each group of sample MR data, the second sample matrix includes second sample elements, the second sample elements include at least one group of the first group of sample elements and the second group of sample elements, the first group of sample elements includes a first sample distance proportional to the maximum time advance in each sample MR data, the second group of sample elements includes a second sample distance between the main service cell in each sample MR data and the main service cell in the earliest sample MR data among the multiple sample MR data; and training the machine learning model with the first sample matrix and the second sample matrix of each group of sample MR data as input and the second label corresponding to the group of sample MR data as output.

[0008] According to another aspect of an embodiment of the present disclosure, an indoor and outdoor determination device is provided, including: an acquisition module, configured to acquire multiple measurement report MR data of a first user within at least one first time window; a first construction module, configured to construct a first matrix including a first element, the first element including the reference signal received power RSRP of multiple service cells in each MR data, the multiple service cells including a main service cell and multiple neighboring service cells; a second construction module, configured to construct a second matrix including a second element, the second element including at least one of a first group of elements and a second group of elements, the first group of elements including a first distance proportional to a maximum time advance in each MR data, the second group of elements including a second distance between the main service cell in each MR data and the main service cell in the earliest MR data among the multiple MR data; and an input module, configured to input input parameters into a trained machine learning model to determine whether the first user is located indoors or outdoors in each first time window, the input parameters including the first matrix and the second matrix.

[0009] According to another aspect of an embodiment of the present disclosure, a model training device is provided, including: an acquisition module, configured to acquire multiple groups of sample MR data, wherein each group of sample MR data includes multiple sample MR data and a first label corresponding to each sample MR data, and the first label is indoor or outdoor; a determination module, configured to determine a second label corresponding to the group of sample MR data according to the first label corresponding to each sample MR data in each group of sample MR data, and the second label is indoor or outdoor; a first construction module, configured to construct a first sample matrix for each group of sample MR data, the first sample matrix includes a first sample element, and the first sample element includes a reference signal received power of multiple service cells in each sample MR data in the group of sample MR data, and the multiple service cells include a main A service cell and multiple neighboring service cells; a second construction module, configured to construct a second sample matrix for each group of sample MR data, the second sample matrix includes second sample elements, the second sample elements include at least one group of the first group of sample elements and the second group of sample elements, the first group of sample elements includes a first sample distance proportional to the maximum time advance in each sample MR data, the second group of sample elements includes a second sample distance between the main service cell in each sample MR data and the main service cell in the earliest sample MR data among the multiple sample MR data; and a training module, configured to train the machine learning model with the first sample matrix and the second sample matrix of each group of sample MR data as input and the second label corresponding to the group of sample MR data as output.

[0010] According to another aspect of an embodiment of the present disclosure, there is provided an indoor and outdoor determination device, comprising: a memory; and a processor coupled to the memory, wherein the processor is configured to execute the indoor and outdoor determination method described in any one of the above embodiments based on instructions stored in the memory.

[0011] According to another aspect of an embodiment of the present disclosure, a model training device is provided, comprising: a memory; and a processor coupled to the memory, wherein the processor is configured to execute the model training method described in any one of the above embodiments based on instructions stored in the memory.

[0012] According to another aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided, comprising computer program instructions, wherein when the computer program instructions are executed by a processor, the method described in any one of the above embodiments is implemented.

[0013] In the disclosed embodiment, the first matrix is ​​constructed based on the RSRP of the primary serving cell and the neighboring serving cell in the multiple MR data, and the second matrix is ​​constructed based on at least one of the first distance and the second distance. In this way, both the RSRP in the multiple MR data and the spatial factor (i.e., at least one of the first distance and the second distance) are considered, so that it is possible to more accurately distinguish whether the user is indoors or outdoors.

[0014] The technical solution of the present disclosure is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present disclosure 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, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0016] Figure 1 is a flowchart of an indoor and outdoor determination method according to some embodiments of the present disclosure;

[0017] Figure 2 is a schematic diagram of a process for determining a first time window according to some implementations;

[0018] Figure 3 is a flowchart of a training method of a machine learning model according to some embodiments of the present disclosure;

[0019] Figure 4 is a flowchart of a model training method according to some embodiments of the present disclosure;

[0020] Figure 5 is a schematic diagram of a process for acquiring multiple sets of sample MR data according to some implementations;

[0021] Figure 6 is a schematic diagram of the structure of an indoor and outdoor determination device according to some embodiments of the present disclosure;

[0022] Figure 7 is a schematic diagram of the structure of a model training device according to some embodiments of the present disclosure;

[0023] Figure 8 is a schematic diagram of the structure of an electronic device according to some embodiments of the present disclosure. DETAILED DESCRIPTION

[0024] The following will be combined with the drawings in the embodiments of the present disclosure to clearly and completely describe the technical solutions in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, rather than all the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present disclosure.

[0025] Unless specifically stated otherwise, the relative arrangement of components and steps, the numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present disclosure.

[0026] At the same time, it should be understood that for the convenience of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship.

[0027] Technologies, methods, and apparatus known to ordinary technicians in the relevant field may not be discussed in detail, but where appropriate, such technologies, methods, and apparatus should be considered part of the authorization specification.

[0028] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limiting. Therefore, other examples of the exemplary embodiments may have different values.

[0029] It should be noted that like reference numerals and letters refer to similar items in the following figures, and therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0030] Figure 1 It is a flowchart of the indoor and outdoor determination method according to some embodiments of the present disclosure.

[0031] In step 102, a plurality of MR data of a first user within at least one first time window is acquired.

[0032] It should be understood that when acquiring MR data in multiple first time windows, each first time window corresponds to multiple MR data. For example, multiple first time windows can form a continuous time period. As an example, the first time window can be about 5 minutes.

[0033] For example, different first users may be distinguished according to mobile phone numbers or International Mobile Subscriber Identity (IMSI), and then MR data of each first user in different first time windows may be obtained.

[0034] In some embodiments, the time intervals between any two MR data pieces that are temporally adjacent in the plurality of MR data pieces are the same. For example, if one MR data piece is generated every 10 seconds, there are a total of 30 MR data pieces in the first time window of 5 minutes.

[0035] In step 104, a first matrix including a first element is constructed. The first element includes reference signal received powers (RSRP) of a plurality of serving cells in each piece of MR data.

[0036] Here, the multiple serving cells include a primary serving cell and multiple neighbor serving cells. That is, the first element in the first matrix includes the RSRP of the primary serving cell in each MR data and the RSRP of multiple neighbor serving cells in each MR data.

[0037] In some embodiments, the multiple serving cells in each MR data include a primary serving cell and four neighboring serving cells. For example, the RSRP of the four neighboring serving cells may be the top four neighboring serving cells in terms of RSRP from largest to smallest among all the neighboring serving cells in the MR data.

[0038] It should be understood that the RSRP of multiple service cells in each MR data can be distributed in the first matrix in a variety of ways. For example, there are a total of m MR data in each first time window, each MR data includes n neighboring service cells, and the first matrix includes m rows of first elements. The i-th first element in the i-th row of first elements is the RSRP of the main service cell in the i-th MR data, the (i+m)th to (i+m+n)th first elements in the i-th row of first elements are the RSRPs of the n neighboring service cells in the i-th MR data, and the other first elements in the first matrix are 0. Here, 1≤i≤m.

[0039] For example, the second first element in the first elements of the second row is the RSRP of the primary service cell in the second MR data, and the (2+m)th to (2+m+n)th first elements in the first elements of the second row are the RSRP of the n neighboring service cells in the second MR data; the third first element in the first elements of the third row is the RSRP of the primary service cell in the third MR data, and the (3+m)th to (3+m+n)th first elements in the first elements of the third row are the RSRP of the n neighboring service cells in the third MR data; and so on.

[0040] At step 106, a second matrix including second elements is constructed.

[0041] Here, the second element includes at least one of the first group of elements and the second group of elements. The first group of elements includes a first distance proportional to the maximum time advance (TA) in each MR data, and the second group of elements includes a second distance between the main service cell in each MR data and the main service cell in the earliest MR data among the multiple MR data.

[0042] It should be understood that TA can represent the distance between the terminal and the antenna port. For example, taking the Long Term Evolution (LTE) network as an example, the first distance can be TA*16*4.89. It should be noted that the embodiments of the present disclosure are not limited to this, and the first distance can also be in other proportional relationships with TA.

[0043] It should also be understood that, for the earliest MR data, the second distance between the primary serving cell and itself in the MR data is 0.

[0044] In the case where the second elements include the first group of elements and the second group of elements, the first group of elements and the second group of elements may be distributed in the second matrix in a variety of ways, some implementations of which will be described below.

[0045] It should be understood that according to step 104 and step 106 , the first matrix and the second matrix corresponding to each first time window can be obtained according to the multiple pieces of MR data in each first time window.

[0046] In step 108, the input parameters are input into the trained machine learning model to determine whether the first user is located indoors or outdoors in each first time window. Here, the input parameters include a first matrix and a second matrix.

[0047] It should be understood that the trained machine learning model can output a label corresponding to the input parameter, i.e., indoor or outdoor, according to the input parameter. For example, the machine learning model may include a convolutional neural network model, etc.

[0048] In the above embodiment, the first matrix is ​​constructed based on the RSRP of the primary serving cell and the neighboring serving cell in the multiple MR data, and the second matrix is ​​constructed based on at least one of the first distance and the second distance. In this way, both the RSRP in the multiple MR data and the spatial factor (i.e., at least one of the first distance and the second distance) are considered, so that it is possible to more accurately distinguish whether the user is indoors or outdoors.

[0049] In some embodiments, the second elements in the second matrix include the first group of elements and the second group of elements. In this case, the second distance is calculated based on each primary service cell and the primary service cell in the earliest MR data, that is, the second distance combines the time factor and the space factor, which can further improve the accuracy of distinguishing indoor and outdoor.

[0050] In some embodiments, the second elements in the second matrix include a first group of elements and a second group of elements. In addition, the second matrix includes a first submatrix and a second submatrix that does not overlap with the first submatrix, the first group of elements belongs to the first submatrix, and the second group of elements belongs to the second submatrix. In this way, the first group of elements and the second group of elements do not affect each other, thereby further improving the accuracy of distinguishing indoor and outdoor.

[0051] It should be understood that the first submatrix and the second submatrix do not overlap, that is, the first submatrix does not include any element in the second group of elements, and the second submatrix does not include any element in the first group of elements. For example, the first submatrix is ​​located on the left, the second submatrix is ​​located on the right, and the two submatrices constitute the second matrix. As an example, the first submatrix is ​​a diagonal matrix, the second submatrix is ​​also a diagonal matrix, the first group of elements are located on the diagonal of the first submatrix, and the second group of elements are located on the diagonal of the second submatrix.

[0052] In some embodiments, the number of multiple MR data, the number of first distances in the first group of elements, and the number of second distances in the second group of elements are m, the jth MR data is earlier than the j+1th MR data, 1≤j≤m-1.

[0053] When i is any value from 1 to m, the second matrix satisfies the following three conditions.

[0054] Condition 1: The i-th first distance is proportional to the TA of the i-th MR data.

[0055] Condition 2: The i-th second distance is the second distance between the primary serving cell in the i-th MR data and the primary serving cell in the first MR data.

[0056] Condition 3: The i-th second element in the second element in the i-th row is one of the i-th first distance and the i-th second distance, and the (i+m)-th second element in the second element in the i-th row is the other of the i-th first distance and the i-th second distance. For example, the 2nd second element in the second element in the 2nd row is one of the 2nd first distance and the 2nd second distance, and the (2+m)-th second element in the second element in the 2nd row is the other of the 2nd first distance and the 2nd second distance; the 3rd second element in the second element in the 3rd row is one of the 3rd first distance and the 3rd second distance, and the (3+m)-th second element in the second element in the 3rd row is the other of the 3rd first distance and the 3rd second distance; and so on.

[0057] In some embodiments, in addition to satisfying the above three conditions, other second elements in the second matrix may be 0.

[0058] By using the second matrix including the first group of elements and the second group of elements satisfying the above conditions as input parameters, the accuracy of distinguishing indoor and outdoor can be further improved.

[0059] The inventors have noticed that in some cases, the user may move from indoors to outdoors or from outdoors to indoors in a short period of time. In this case, how to determine the first time window will affect the accuracy of the final distinction between indoors and outdoors. Figure 2 Provide explanation.

[0060] Figure 2 is a flowchart of determining a first time window according to some implementations.

[0061] In step 112, a plurality of initial MR data of the first user within a first preset time period is acquired.

[0062] It should be understood that the first preset time period includes a first time window. It should also be understood that, in order to distinguish the MR data in the first time period from the MR data in the first time window, the MR data in the first time period is referred to as initial MR data.

[0063] In step 122, the plurality of initial MR data are sorted in chronological order.

[0064] It is understandable that each piece of initial MR data is obtained at a different time. For example, each piece of initial MR data may carry a corresponding timestamp.

[0065] In step 132, the sorted multiple pieces of initial MR data are divided into multiple groups of initial MR data according to multiple continuous initial time windows. Here, the multiple initial time windows constitute a first preset time period.

[0066] After the multiple pieces of initial MR data are sorted in chronological order, the multiple pieces of initial MR data can be divided into multiple groups of initial MR data according to multiple initial time windows. For example, the first group of initial MR data includes a first portion of initial MR data, and the second group of initial MR data includes a second portion of initial MR data, and the second portion of initial MR data is later than the first portion of initial MR data.

[0067] In step 142 , a plurality of first reference values ​​corresponding one-to-one to each set of initial MR data and a plurality of first segmentation modes are determined.

[0068] It should be understood that for each set of initial MR data, there are corresponding multiple first reference values ​​under multiple first segmentation modes.

[0069] In each first segmentation mode, a first initial MR data other than the first initial MR data and the last initial MR data in the group of initial MR data is used as a segmentation point to segment the group of initial MR data into a first subgroup of initial MR data in a first initial sub-time window and a second subgroup of initial MR data in a second initial sub-time window. It should be understood that the first initial sub-time window and the second initial sub-time window constitute the initial time window where the group of initial MR data is located.

[0070] In a plurality of first segmentation modes, corresponding first subgroup initial MR data and second subgroup initial MR data can be obtained. Each first reference value reflects the degree of discreteness between the first subgroup initial MR data and the second subgroup initial MR data in the corresponding first segmentation mode.

[0071] In some embodiments, each first reference value iv is determined according to the following formula:

[0072]

[0073] in, is the average value of the first RSRP of the primary serving cell in the first subgroup initial MR data under the first segmentation mode corresponding to the first reference value iv, is the average value of the second RSRP of the primary serving cell in the second subgroup initial MR data under the first segmentation mode corresponding to the first reference value iv, σ x is the standard deviation of the first RSRP, σ y is the standard deviation of the second RSRP.

[0074] The first reference value obtained according to the above formula can more accurately reflect the discreteness between the first subgroup initial MR data and the second subgroup initial MR data, so that each group of initial MR data can be segmented more accurately.

[0075] In some embodiments, if the time interval between two temporally adjacent MR data is not a preset interval (for example, 10 seconds), the average value of the RSRP of the main service cell in the two MR data can be used as the RSRP of the MR data between the two MR data, and then the average value of the RSRP of the main service cell in the first or second sub-group initial MR data is calculated, so that the time window can be divided more accurately.

[0076] At step 152 , a first group of initial MR data whose maximum value among a plurality of first reference values ​​is greater than a threshold value and a second group of initial MR data whose maximum value is not greater than the threshold value are determined from the plurality of groups of initial MR data.

[0077] For a certain group of initial MR data, if the maximum value among the multiple first reference values ​​of the group of initial MR data is greater than a threshold, the group of initial MR data is the first group of initial MR data; otherwise, the group of initial MR data is the second group of initial MR data.

[0078] For the first group of initial MR data, according to the first segmentation method corresponding to the maximum value among the multiple first reference values ​​of the first group of initial MR data, the first group of initial MR data is segmented into a first subgroup of initial MR data in a first initial sub-time window and a second subgroup of initial MR data in a second initial sub-time window. In other words, the initial time window where the first group of initial MR data is located is segmented into a first initial sub-time window and a second initial sub-time window.

[0079] For the second set of initial MR data, the initial time window in which the second set of initial MR data is located is retained.

[0080] Therefore, the initial time window where the second group of initial MR data is located, the first initial sub-time window where the first sub-group of initial MR data is divided according to the first segmentation method corresponding to the maximum value of multiple first reference values ​​of the first group of initial MR data, and the second initial sub-time window where the second sub-group of initial MR data is divided according to the first segmentation method corresponding to the maximum value of multiple first reference values ​​of the first group of initial MR data can each be used as a time window.

[0081] In some embodiments, the above thresholds can be dynamically adjusted according to actual conditions to improve the accuracy of dividing the time window.

[0082] In the above implementation, the initial sub-time window dynamically divided from a part of the initial time window and the initial time window from another part are used as the time window, which can improve the accuracy of the time window division, thereby more accurately distinguishing whether the first user is indoors or outdoors in each time window.

[0083] In order to more accurately distinguish whether the first user is located indoors or outdoors, the input parameters input into the machine learning model may also include other parameters, which are explained below in conjunction with different embodiments.

[0084] In some embodiments, the input parameters may also include at least one of the following: a first number of multiple service cells (as described above, including main service cells and neighboring service cells) of multiple MR data located indoors; a second number of multiple service cells of multiple MR data located outdoors; a ratio of the first number to the second number; the number of main service cells; the number of multiple neighboring service cells.

[0085] For example, the first matrix and the second matrix can be convolved and pooled, and the features obtained after region of interest pooling (ROI polling) and one or more of the above features are input into a fully connected layer (for example, 2 or 3) to obtain the result of whether the first user is indoors or outdoors.

[0086] In some other embodiments, the input parameter may further include a reference signal received quality (RSRQ) of the primary serving cell in each piece of MR data.

[0087] In some other embodiments, the input parameters may further include a time attribute of each first time window, the time attribute indicating a working day or a holiday. In addition, the input parameters may further include a feature indicating whether the serving cell is a suburban area.

[0088] It can be understood that, in some embodiments, the input parameters may include all the parameters mentioned above, so as to more accurately distinguish whether the first user is located indoors or outdoors.

[0089] Figure 3 1 is a flowchart of a training method of a machine learning model according to some embodiments of the present disclosure. For example, the machine learning model may be trained at predetermined intervals in the manner described below.

[0090] In step 302, a plurality of initial sample MR data of each user in a corresponding second preset time period and a first label corresponding to each initial sample MR data are obtained. Here, the first label is indoor or outdoor.

[0091] Here, the second preset time period corresponding to each user may be the same or different. In some embodiments, the second preset time period corresponding to each user is the same, so that the machine learning model can be trained more accurately.

[0092] For example, after the initial sample MR data of multiple users within the second preset time are acquired, the initial sample MR data may be grouped according to the user's identifier to obtain multiple pieces of initial sample MR data of each user within the second preset time.

[0093] For example, it may be determined by way of drive testing whether the first tag corresponding to each piece of initial sample MR data is indoors or outdoors.

[0094] In step 304, the plurality of initial sample MR data are sorted in chronological order.

[0095] It should be understood that the multiple initial samples MR here may belong to any one of the multiple users.

[0096] In step 306, the sorted plurality of initial sample MR data are divided into a plurality of groups of initial sample MR data according to a plurality of continuous initial sample time windows, and the plurality of initial sample time windows constitute a second preset time period.

[0097] In step 308 , a plurality of second reference values ​​corresponding one-to-one to each set of initial sample MR data and a plurality of second segmentation modes are determined.

[0098] It should be understood that for each set of initial sample MR data, there are corresponding multiple second reference values ​​under multiple second segmentation modes.

[0099] In each second segmentation mode, a first initial sample MR data other than the first initial sample MR data and the last initial sample MR data in the group of initial sample MR data is used as a segmentation point to segment the group of initial sample MR data into a first subgroup of initial sample MR data in the first initial sample sub-time window and a second subgroup of initial sample MR data in the second initial sample sub-time window. It should be understood that the first initial sample sub-time window and the second initial sample sub-time window constitute the initial sample time window where the group of initial sample MR data is located.

[0100] Each second reference value reflects the discreteness between the first subset of initial sample MR data and the second subset of initial sample MR data in the corresponding second segmentation mode.

[0101] In some embodiments, the second reference value may be determined in a manner similar to the first reference value described above, that is, each second reference value iv is determined according to the following formula, namely:

[0102]

[0103] in, is the average value of the third RSRP of the primary serving cell in the first subgroup initial sample MR data under the second segmentation mode corresponding to the second reference value iv, is the average value of the fourth RSRP of the primary serving cell in the second subgroup initial sample MR data under the second segmentation mode corresponding to the second reference value iv, σ x is the standard deviation of the third RSRP, σ y is the standard deviation of the fourth RSRP.

[0104] In step 310 , a first group of initial sample MR data in which the maximum value of the second reference values ​​is greater than a threshold and a second group of initial sample MR data in which the maximum value of the second reference values ​​is not greater than the threshold are determined from the plurality of groups of initial sample MR data.

[0105] At step 312, multiple sets of sample MR data are determined.

[0106] Here, the multiple groups of sample MR data include the second group of initial sample MR data, the first subgroup of initial sample MR data and the second subgroup of initial sample MR data obtained by dividing the first group of initial sample MR data according to the second division method corresponding to the maximum value among the multiple second reference values ​​of the first group of initial sample MR data. In other words, the second group of initial sample MR data, the first subgroup of initial sample MR data and the second subgroup of initial sample MR data obtained by dividing the first group of initial sample MR data are each regarded as a group of sample MR data.

[0107] Each group of sample MR data includes a plurality of sample MR data. For example, the second group of initial sample MR data as a group of sample MR data includes a plurality of sample MR data, namely, the initial sample MR data in the second group of initial sample MR data.

[0108] In step 314, according to the first label corresponding to each piece of sample MR data in each set of sample MR data, the second label corresponding to the set of sample MR data is determined, and the second label is indoor or outdoor.

[0109] As some implementation methods, if the number of first labels of indoor in the group of sample MR data is greater than the number of outdoor, then the second label corresponding to the group of sample MR data is indoor; if the number of first labels of indoor in the group of sample MR data is less than the number of outdoor, then the second label corresponding to the group of sample MR data is outdoor.

[0110] In addition, considering that the probability of the first label being indoor (e.g., 90%) is much greater than the probability of being outdoor (e.g., about 10%), in some embodiments, if the number of first labels being indoor in the set of sample MR data is equal to the number of first labels being outdoor, then the second label corresponding to the set of sample MR data may also be indoor. Of course, the disclosed embodiments are not limited thereto, and in other embodiments, if the number of first labels being indoor in the set of sample MR data is equal to the number of first labels being outdoor, then the second label corresponding to the set of sample MR data may also be outdoor.

[0111] In step 316, a first sample matrix is ​​constructed for each set of sample MR data. The first sample matrix includes first sample elements, and the first sample elements include RSRPs of multiple serving cells (including a primary serving cell and multiple neighboring serving cells) in each piece of sample MR data in the set of sample MR data.

[0112] For example, the RSRP of multiple service cells can be distributed as the first sample element in the first sample matrix according to a preset rule, and when the machine learning model is subsequently used for prediction, the RSRP of multiple service cells can still be distributed as the first element in the first matrix according to the same preset rule.

[0113] At step 318, a second sample matrix is ​​constructed for each set of sample MR data.

[0114] Here, the second sample matrix includes second sample elements, and the second sample elements include at least one of the first group of sample elements and the second group of sample elements.

[0115] The first group of sample elements includes a first sample distance proportional to the TA in each sample MR data, and the second group of sample elements includes a second sample distance between the main service cell in each sample MR data and the main service cell in the earliest sample MR data among the multiple sample MR data.

[0116] For example, the proportionality coefficient between the first sample distance and TA is the same as the proportionality coefficient between the first distance and TA. For example, the first sample distance may be TA*16*4.89.

[0117] In addition, it should be understood that, for the earliest sample MR data, the second sample distance between the primary serving cell and itself in the sample MR data is 0.

[0118] The second sample elements in the second sample matrix may be distributed in the same manner as the second elements in the second matrix described above, which will not be described in detail herein.

[0119] In step 320, the machine learning model is trained with the first sample matrix and the second sample matrix of each group of sample MR data as input and the second label corresponding to the group of sample MR data as output.

[0120] After training, the machine learning model can obtain the prediction result of whether the user is indoors or outdoors based on the input first matrix and second matrix.

[0121] Figure 4 is a flow chart of a model training method according to some embodiments of the present disclosure. It should be noted that: Figure 4 and Figure 5 Some of the steps introduced have been explained above, so they will not be repeated here.

[0122] In step 402, multiple groups of sample MR data are acquired. Here, each group of sample MR data includes multiple pieces of sample MR data and a first label corresponding to each piece of sample MR data, and the first label is indoor or outdoor.

[0123] In step 404, according to the first label corresponding to each piece of sample MR data in each group of sample MR data, the second label corresponding to the group of sample MR data is determined, and the second label is indoor or outdoor.

[0124] For example, if the number of samples in this group of sample MR data with the first label of indoor is greater than the number of samples in this group of sample MR data with the first label of outdoor, then the second label corresponding to this group of sample MR data is indoor; if the number of samples in this group of sample MR data with the first label of indoor is less than the number of samples in this group of sample MR data with the first label of outdoor, then the second label corresponding to this group of sample MR data is outdoor.

[0125] In step 406 , a first sample matrix of each set of sample MR data is constructed.

[0126] The first sample matrix includes a first sample element, and the first sample element includes a reference signal received power RSRP of multiple serving cells in each sample MR data in the group of sample MR data. Here, the multiple serving cells include a primary serving cell and multiple neighboring serving cells.

[0127] At step 408, a second sample matrix is ​​constructed for each set of sample MR data.

[0128] The second sample matrix includes second sample elements, and the second sample elements include at least one of the first group of sample elements and the second group of sample elements.

[0129] The first group of sample elements includes a first sample distance proportional to the TA in each sample MR data, and the second group of sample elements includes a second sample distance between the main service cell in each sample MR data and the main service cell in the earliest sample MR data among the multiple sample MR data.

[0130] In step 410, the machine learning model is trained with the first sample matrix and the second sample matrix of each group of sample MR data as input and the second label corresponding to the group of sample MR data as output.

[0131] In the above embodiment, the first sample matrix is ​​constructed based on the RSRP of the primary service cell and the neighboring service cell in the multiple MR data, and the second sample matrix is ​​constructed based on at least one of the first sample distance and the second sample distance. In this way, both the RSRP in the multiple MR data and the spatial factor (i.e., at least one of the first sample distance and the second sample distance) are considered, so that the machine learning model can be trained more accurately to more accurately classify whether the user is located indoors or outdoors.

[0132] Figure 5 It is a schematic diagram of a process of acquiring multiple sets of sample MR data according to some implementations.

[0133] In step 412, a plurality of pieces of initial sample MR data of each user in a corresponding preset time period and a first label corresponding to each piece of initial sample MR data are obtained.

[0134] Here, the preset time period corresponding to each user may be the same or different. In some embodiments, the preset time period corresponding to each user is the same, so that the machine learning model can be trained more accurately.

[0135] In step 422, the plurality of initial sample MR data are sorted in chronological order.

[0136] In step 432, the sorted plurality of initial sample MR data are divided into a plurality of groups of initial sample MR data according to a plurality of continuous initial sample time windows, and the plurality of initial sample time windows constitute a preset time period.

[0137] In step 442, a plurality of reference values ​​corresponding one-to-one to each set of initial sample MR data and a plurality of segmentation modes are determined.

[0138] In each segmentation method, a first initial sample MR data other than the first initial sample MR data and the last initial sample MR data in the group of initial sample MR data is used as a segmentation point to segment the group of initial sample MR data into a first subgroup of initial sample MR data within the first initial sample sub-time window and a second subgroup of initial sample MR data within the second initial sample sub-time window.

[0139] Each reference value reflects the degree of discreteness between the first subset of initial sample MR data and the second subset of initial sample MR data in the corresponding segmentation mode. The reference value can be determined in the manner of determining the first reference value and the second reference value described above.

[0140] At step 452 , a first group of initial sample MR data in which the maximum value among the plurality of reference values ​​is greater than a threshold value and a second group of initial sample MR data in which the maximum value among the plurality of second reference values ​​is not greater than the threshold value are determined from the plurality of groups of initial sample MR data.

[0141] The obtained multiple groups of sample MR data may include a second group of initial sample MR data, a first subgroup of initial sample MR data divided by a division method corresponding to a maximum value among multiple reference values ​​of the first group of initial sample MR data, and a second subgroup of initial sample MR data.

[0142] The multiple groups of sample MR data obtained by the above implementation method are more reasonable, and the machine learning model can be trained more accurately based on them, so that the indoor and outdoor distinction results can be more accurate.

[0143] It should be noted that in Figure 1-Figure 5 Certain steps shown in a sequential order do not necessarily imply that these steps should be performed in a sequential order.

[0144] Each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device embodiment, since it basically corresponds to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0145] Figure 6 It is a structural schematic diagram of an indoor and outdoor determination device according to some embodiments of the present disclosure.

[0146] like Figure 6 As shown, the indoor and outdoor determination device includes an acquisition module 601 , a first construction module 602 , a second construction module 603 and an input module 604 .

[0147] The acquisition module 601 is configured to acquire a plurality of measurement report MR data of a first user within at least one first time window.

[0148] The first construction module 602 is configured to construct a first matrix including a first element, where the first element includes RSRPs of a plurality of serving cells in each piece of MR data, where the plurality of serving cells include a primary serving cell and a plurality of neighbor serving cells.

[0149] The second construction module 603 is configured to construct a second matrix including second elements, the second elements including at least one of the first group of elements and the second group of elements, the first group of elements including a first distance proportional to the TA in each MR data, and the second group of elements including a second distance between the main service cell in each MR data and the main service cell in the earliest MR data among the multiple MR data.

[0150] The input module 604 is configured to input input parameters into the trained machine learning model to determine whether the first user is located indoors or outdoors in each first time window, and the input parameters include a first matrix and a second matrix.

[0151] Figure 7 It is a structural schematic diagram of a model training device according to some embodiments of the present disclosure.

[0152] like Figure 7 As shown, the model training device includes an acquisition module 701, a determination module 702, a first construction module 703, a second construction module 704 and a training module 705.

[0153] The acquisition module 701 is configured to acquire multiple groups of sample MR data, wherein each group of sample MR data includes multiple pieces of sample MR data and a first label corresponding to each piece of sample MR data, and the first label is indoor or outdoor.

[0154] The determination module 702 is configured to determine the second label corresponding to each set of sample MR data according to the first label corresponding to each piece of sample MR data in each set of sample MR data, where the second label is indoor or outdoor.

[0155] The first construction module 703 is configured to construct a first sample matrix for each group of sample MR data, the first sample matrix includes a first sample element, the first sample element includes the reference signal received power RSRP of multiple service cells in each sample MR data in the group of sample MR data, and the multiple service cells include a primary service cell and multiple neighboring service cells.

[0156] The second construction module 704 is configured to construct a second sample matrix for each group of sample MR data, the second sample matrix includes second sample elements, the second sample elements include at least one of the first group of sample elements and the second group of sample elements, the first group of sample elements includes a first sample distance proportional to the TA in each sample MR data, and the second group of sample elements includes a second sample distance between the main service cell in each sample MR data and the main service cell in the earliest sample MR data among the multiple sample MR data.

[0157] The training module 705 is configured to train the machine learning model using the first sample matrix and the second sample matrix of each group of sample MR data as input and the second label corresponding to the group of sample MR data as output.

[0158] Figure 8 1 is a schematic diagram of the structure of an electronic device according to some embodiments of the present disclosure. The electronic device may be an indoor or outdoor determination device or a model training device.

[0159] like Figure 8 As shown, the electronic device 800 includes a memory 801 and a processor 802 coupled to the memory 801 , and the processor 802 is configured to execute the method of any one of the aforementioned embodiments based on the instructions stored in the memory 801 .

[0160] The memory 801 may include, for example, a system memory, a fixed non-volatile storage medium, etc. The system memory may store, for example, an operating system, an application program, a boot loader, and other programs.

[0161] The electronic device 800 may also include an input / output interface 803, a network interface 804, a storage interface 808, etc. These interfaces 803, 804, 808, and the memory 801 and the processor 802 may be connected, for example, via a bus 806. The input / output interface 803 provides a connection interface for input / output devices such as a display, a mouse, a keyboard, and a touch screen. The network interface 804 provides a connection interface for various networked devices. The storage interface 808 provides a connection interface for external storage devices such as an SD card and a USB flash drive.

[0162] The embodiments of the present disclosure further provide a computer-readable storage medium, including computer program instructions, which implement the method of any one of the above embodiments when executed by a processor.

[0163] In some embodiments, the distributed computing architecture used for MR longitude and latitude prediction can be used to assist in implementing the indoor and outdoor determination method to reduce the performance requirements of the system.

[0164] It is understandable that after using the solution of the embodiment of the present disclosure to more accurately determine whether the user is indoors or outdoors, network planning and optimization can be performed more accurately, which helps to formulate more accurate network optimization and base station coverage plans. In addition, using more accurate indoor and outdoor distinction results helps to analyze the user's network usage habits, build a more accurate user portrait, and improve user experience.

[0165] So far, various embodiments of the present disclosure have been described in detail. In order to avoid obscuring the concept of the present disclosure, some details known in the art are not described. Based on the above description, those skilled in the art can fully understand how to implement the technical solution disclosed here.

[0166] Those skilled in the art will appreciate that the embodiments of the present disclosure may be provided as methods, systems, or computer program products. Therefore, the present disclosure may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present disclosure may take the form of a computer program product implemented on one or more computer-usable non-transient storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0167] The present disclosure is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present disclosure. It should be understood that the functions specified in one or more processes in the flowchart and / or one or more blocks in the block diagram 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, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the functions specified in the flowchart. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0168] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0169] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0170] Although some specific embodiments of the present disclosure have been described in detail by way of examples, it should be understood by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the present disclosure. It should be understood by those skilled in the art that the above embodiments may be modified or some technical features may be replaced by equivalents without departing from the scope and spirit of the present disclosure. The scope of the present disclosure is defined by the appended claims.

Claims

1. A method for determining indoor and outdoor, comprising: Acquire a plurality of measurement report MR data of a first user within at least one first time window; Constructing a first matrix including a first element, wherein the first element includes reference signal received powers (RSRP) of multiple serving cells in each piece of MR data, wherein the multiple serving cells include a primary serving cell and multiple neighboring serving cells; Constructing a second matrix including second elements, the second elements including a first group of elements and a second group of elements, the first group of elements including a first distance proportional to a maximum time advance TA in each piece of MR data, the second group of elements including a second distance between a primary serving cell in each piece of MR data and a primary serving cell in the earliest MR data among the multiple pieces of MR data; and Input parameters are input into the trained machine learning model to determine whether the first user is located indoors or outdoors in each first time window, the input parameters including the first matrix and the second matrix.

2. The method according to claim 1, wherein: The acquiring of a plurality of measurement report MR data of a first user within at least one first time window comprises: Acquire a plurality of initial MR data of the first user within a first preset time period; sorting the plurality of initial MR data in chronological order; Dividing the sorted multiple pieces of initial MR data into multiple groups of initial MR data according to multiple continuous initial time windows, wherein the multiple initial time windows constitute the first preset time period; Determine a plurality of first reference values ​​corresponding one-to-one to each set of initial MR data and a plurality of first segmentation modes, wherein: In each first segmentation mode, a first initial MR data other than the first initial MR data and the last initial MR data in the group of initial MR data is used as a segmentation point to segment the group of initial MR data into a first subgroup of initial MR data within a first initial sub-time window and a second subgroup of initial MR data within a second initial sub-time window, Each first reference value reflects the discreteness between the first subgroup initial MR data and the second subgroup initial MR data in the corresponding first segmentation mode; From the plurality of groups of initial MR data, determine a first group of initial MR data in which a maximum value among the plurality of first reference values ​​is greater than a threshold value and a second group of initial MR data in which the maximum value is not greater than the threshold value; Among them, the at least one first time window includes the initial time window where the second group of initial MR data is located, the first initial sub-time window where the first sub-group of initial MR data is divided according to the first segmentation method corresponding to the maximum value of multiple first reference values ​​of the first group of initial MR data, and the second initial sub-time window where the second sub-group of initial MR data is divided according to the first segmentation method corresponding to the maximum value of multiple first reference values ​​of the first group of initial MR data.

3. The method according to claim 2, wherein: Each first reference value Determined according to the following formula: in, The first reference value The average value of the first RSRP of the primary serving cell in the first subgroup initial MR data under the corresponding first segmentation mode, The first reference value The average value of the second RSRP of the primary serving cell in the second subgroup initial MR data under the corresponding first segmentation mode, is the standard deviation of the first RSRP, is the standard deviation of the second RSRP.

4. The method according to claim 1, wherein: The second matrix includes a first submatrix and a second submatrix that does not overlap with the first submatrix, the first group of elements belongs to the first submatrix, and the second group of elements belongs to the second submatrix.

5. The method according to claim 4, wherein: The number of the plurality of MR data, the number of first distances in the first group of elements, and the number of second distances in the second group of elements is m, the jth MR data is earlier than the j+1th MR data, 1≤j≤m-1; When i is any value from 1 to m: The i-th first distance is proportional to the TA of the i-th MR data. The i-th second distance is the second distance between the primary service cell in the i-th MR data and the primary service cell in the first MR data, The i-th second element in the i-th row of second elements is one of the i-th first distance and the i-th second distance, and the (i+m)-th second element in the i-th row of second elements is the other of the i-th first distance and the i-th second distance.

6. The method according to any one of claims 1 to 5, wherein: The input parameters also include at least one of the following: The plurality of serving cells of the plurality of MR data are located in a first number indoors; The plurality of serving cells of the plurality of MR data are located in a second number outdoors; a ratio of the first amount to the second amount; The number of the primary serving cells; The number of the plurality of neighbor serving cells.

7. The method according to claim 6, wherein: The input parameters also include the reference signal reception quality of the primary serving cell in each piece of MR data.

8. The method according to claim 6, wherein: The input parameters further include a time attribute of each first time window, where the time attribute indicates a working day or a holiday.

9. The method according to claim 2, wherein: The machine learning model is trained as follows: Acquire a plurality of initial sample MR data of each user among a plurality of users within a corresponding second preset time period and a first label corresponding to each initial sample MR data, wherein the first label is indoor or outdoor; Sorting the plurality of initial sample MR data in chronological order; Dividing the sorted plurality of initial sample MR data into a plurality of groups of initial sample MR data according to a plurality of continuous initial sample time windows, wherein the plurality of initial sample time windows constitute the second preset time period; Determine a plurality of second reference values ​​corresponding one-to-one to each group of initial sample MR data and a plurality of second segmentation modes, wherein: In each second segmentation mode, a first initial sample MR data other than the first initial sample MR data and the last initial sample MR data in the group of initial sample MR data is used as a segmentation point to segment the group of initial sample MR data into a first subgroup of initial sample MR data within the first initial sample sub-time window and a second subgroup of initial sample MR data within the second initial sample sub-time window. Each second reference value reflects the discreteness between the first subgroup initial sample MR data and the second subgroup initial sample MR data under the corresponding second segmentation mode; Determine, from the plurality of groups of initial sample MR data, a first group of initial sample MR data in which the maximum value of the plurality of second reference values ​​is greater than the threshold value and a second group of initial sample MR data in which the maximum value of the plurality of second reference values ​​is not greater than the threshold value; Determine a plurality of groups of sample MR data, the plurality of groups of sample MR data comprising the second group of initial sample MR data, a first subgroup of initial sample MR data divided by the first group of initial sample MR data according to a second division method corresponding to a maximum value among a plurality of second reference values ​​of the first group of initial sample MR data, and a second subgroup of initial sample MR data, wherein each group of sample MR data comprises a plurality of sample MR data; Determine, according to the first label corresponding to each piece of sample MR data in each group of sample MR data, a second label corresponding to the group of sample MR data, wherein the second label is indoor or outdoor; Constructing a first sample matrix for each group of sample MR data, wherein the first sample matrix includes a first sample element, and the first sample element includes the RSRP of the multiple serving cells in each sample MR data in the group of sample MR data; Constructing a second sample matrix for each group of sample MR data, the second sample matrix comprising second sample elements, the second sample elements comprising at least one of a first group of sample elements and a second group of sample elements, the first group of sample elements comprising a first sample distance proportional to the TA in each sample MR data, the second group of sample elements comprising a second sample distance between a primary serving cell in each sample MR data and a primary serving cell in the earliest sample MR data among the multiple sample MR data; and The machine learning model is trained by taking the first sample matrix and the second sample matrix of each group of sample MR data as input and taking the second label corresponding to the group of sample MR data as output.

10. The method according to claim 9, wherein: Determining the second label corresponding to each piece of sample MR data in each set of sample MR data includes: If the number of samples with the first label of indoor in the group of sample MR data is greater than the number of samples with the first label of outdoor, then the second label corresponding to the group of sample MR data is indoor; If the number of samples with the first label of indoor in the set of sample MR data is less than the number of samples with the first label of outdoor, the second label corresponding to the set of sample MR data is outdoor.

11. A model training method, comprising: Acquire multiple groups of sample MR data, wherein each group of sample MR data includes multiple pieces of sample MR data and a first label corresponding to each piece of sample MR data, and the first label is indoor or outdoor; Determine, according to the first label corresponding to each piece of sample MR data in each group of sample MR data, a second label corresponding to the group of sample MR data, wherein the second label is indoor or outdoor; Constructing a first sample matrix for each group of sample MR data, the first sample matrix comprising a first sample element, the first sample element comprising reference signal received powers of multiple serving cells in each piece of sample MR data in the group of sample MR data, the multiple serving cells comprising a primary serving cell and multiple neighboring serving cells; Constructing a second sample matrix for each group of sample MR data, the second sample matrix includes second sample elements, the second sample elements include a first group of sample elements and a second group of sample elements, the first group of sample elements includes a first sample distance proportional to a maximum time advance in each sample MR data, the second group of sample elements includes a second sample distance between a primary serving cell in each sample MR data and a primary serving cell in the earliest sample MR data among the multiple sample MR data; and The machine learning model is trained by taking the first sample matrix and the second sample matrix of each group of sample MR data as input and taking the second label corresponding to the group of sample MR data as output.

12. The method according to claim 11, wherein: Acquiring multiple sets of sample MR data includes: Acquire a plurality of initial sample MR data of each user among a plurality of users within a corresponding preset time period and a first label corresponding to each piece of initial sample MR data; Sorting the plurality of initial sample MR data in chronological order; Dividing the sorted plurality of initial sample MR data into a plurality of groups of initial sample MR data according to a plurality of continuous initial sample time windows, wherein the plurality of initial sample time windows constitute the preset time period; Determine multiple reference values ​​corresponding to each group of initial sample MR data and multiple segmentation methods, where: In each segmentation mode, a first piece of initial sample MR data other than the first piece of initial sample MR data and the last piece of initial sample MR data in the group of initial sample MR data is used as a segmentation point, and the group of initial sample MR data is segmented into a first subgroup of initial sample MR data within the first initial sample sub-time window and a second subgroup of initial sample MR data within the second initial sample sub-time window. Each reference value reflects the degree of discreteness between the first subgroup initial sample MR data and the second subgroup initial sample MR data under the corresponding segmentation mode; Determine, from the plurality of groups of initial sample MR data, a first group of initial sample MR data in which the maximum value of the plurality of reference values ​​is greater than a threshold value and a second group of initial sample MR data in which the maximum value of the plurality of reference values ​​is not greater than the threshold value; The multiple groups of sample MR data include the second group of initial sample MR data, a first subgroup of initial sample MR data obtained by dividing the first group of initial sample MR data in a manner corresponding to a maximum value among multiple reference values ​​of the first group of initial sample MR data, and a second subgroup of initial sample MR data.

13. The method according to claim 11 or 12, wherein: Determining the second label corresponding to each piece of sample MR data in each set of sample MR data includes: If the number of samples with the first label of indoor in the group of sample MR data is greater than the number of samples with the first label of outdoor, then the second label corresponding to the group of sample MR data is indoor; If the number of samples with the first label of indoor in the set of sample MR data is less than the number of samples with the first label of outdoor, the second label corresponding to the set of sample MR data is outdoor.

14. An indoor and outdoor determination device, comprising: An acquisition module, configured to acquire a plurality of measurement report MR data of a first user within at least one first time window; A first construction module is configured to construct a first matrix including a first element, wherein the first element includes reference signal received power RSRP of multiple serving cells in each MR data, wherein the multiple serving cells include a primary serving cell and multiple neighboring serving cells; A second construction module is configured to construct a second matrix including second elements, wherein the second elements include a first group of elements and a second group of elements, wherein the first group of elements includes a first distance proportional to a maximum time advance in each piece of MR data, and the second group of elements includes a second distance between a primary serving cell in each piece of MR data and a primary serving cell in the earliest MR data among the plurality of pieces of MR data; and An input module is configured to input input parameters into a trained machine learning model to determine whether the first user is indoors or outdoors in each first time window, and the input parameters include the first matrix and the second matrix.

15. A model training device, comprising: An acquisition module is configured to acquire multiple groups of sample MR data, wherein each group of sample MR data includes multiple pieces of sample MR data and a first label corresponding to each piece of sample MR data, and the first label is indoor or outdoor; A determination module is configured to determine a second label corresponding to each set of sample MR data according to a first label corresponding to each piece of sample MR data in each set of sample MR data, wherein the second label is indoor or outdoor; A first construction module is configured to construct a first sample matrix for each group of sample MR data, wherein the first sample matrix includes a first sample element, and the first sample element includes a reference signal received power of multiple serving cells in each sample MR data in the group of sample MR data, wherein the multiple serving cells include a primary serving cell and multiple neighboring serving cells; A second construction module is configured to construct a second sample matrix for each group of sample MR data, wherein the second sample matrix includes second sample elements, the second sample elements include a first group of sample elements and a second group of sample elements, the first group of sample elements includes a first sample distance proportional to a maximum time advance in each sample MR data, and the second group of sample elements includes a second sample distance between a primary serving cell in each sample MR data and a primary serving cell in the earliest sample MR data among the multiple sample MR data; and The training module is configured to train the machine learning model using the first sample matrix and the second sample matrix of each group of sample MR data as input and the second label corresponding to the group of sample MR data as output.

16. An indoor and outdoor determination device, comprising: Memory; as well as A processor coupled to the memory, configured to execute the method according to any one of claims 1 to 10 based on instructions stored in the memory.

17. A model training device, comprising: Memory; as well as A processor coupled to the memory is configured to execute the method of any one of claims 11-13 based on instructions stored in the memory.

18. A computer-readable storage medium comprising computer program instructions, wherein: When the computer program instructions are executed by a processor, the method according to any one of claims 1 to 13 is implemented.

Images