WO2023216882A1

WO2023216882A1 - Radio-frequency fingerprint library updating method and apparatus, and computer device and computer storage medium

Info

Publication number: WO2023216882A1
Application number: PCT/CN2023/090782
Authority: WO
Inventors: 许正一
Original assignee: 中兴通讯股份有限公司
Priority date: 2022-05-07
Filing date: 2023-04-26
Publication date: 2023-11-16
Also published as: CN117053784A

Abstract

A radio-frequency fingerprint library updating method. Historical fingerprint data in a radio-frequency fingerprint library comprises cluster information and sub-cluster information. The method comprises: acquiring real-time fingerprint data, and dividing the real-time fingerprint data into different types of data sets according to cluster information, wherein the data sets at least comprise a boundary data set (S1); and updating a radio-frequency fingerprint library according to the various types of data sets; and for first fingerprint data in the boundary data set, according to sub-cluster information, determining the first sub-cluster closest to the position where the first fingerprint data is located, and according to information of the first sub-cluster, updating sub-cluster information of historical first fingerprint data, in the radio-frequency fingerprint library, corresponding to the first fingerprint data (S2). Further provided are a radio-frequency fingerprint library updating apparatus, and a computer device and a computer-readable storage medium.

Description

Radio frequency fingerprint database update method and device, computer equipment and computer storage medium

Cross-references to related applications

This application claims priority from Chinese patent application No. 202210493455.6 submitted on May 7, 2022. The content of this Chinese patent application is incorporated herein by reference in its entirety.

Technical field

The present disclosure relates to the field of communication technology, and specifically to a radio frequency fingerprint database updating method, a radio frequency fingerprint database updating device, computer equipment and computer readable storage media.

Background technique

The implementation of terminal indoor positioning relies on the fingerprint data in the radio frequency fingerprint database. The fingerprint database is divided into several areas. After obtaining the fingerprint data of the point to be located in the positioning stage, the point to be located is determined to be in a certain area, and finally in this area Determine the accuracy of the position of the point to be located. The accuracy of terminal indoor positioning depends to a large extent on the reliability of fingerprint data in the radio frequency fingerprint database. Therefore, how to improve the reliability of fingerprint data is particularly important.

One of the existing radio frequency fingerprint database update solutions focuses on the processing of the crowdsourcing data itself, that is, over-reliance on the crowdsourcing data itself, leaving the reliability judgment of the data completely to the user, resulting in poor fingerprint data reliability. Difference. The other is to update the radio frequency fingerprint database based on historical data and incremental data. The regional division is inaccurate, which also greatly reduces the reliability of the fingerprint data.

public content

In a first aspect, the present disclosure provides a method for updating a radio frequency fingerprint database. The historical fingerprint data in the radio frequency fingerprint database includes cluster information and sub-cluster information. The method includes: acquiring real-time fingerprint data and updating all the fingerprint data according to the cluster information. The real-time fingerprint data is divided into different types of data sets, the data sets at least include boundary data sets; and according to the data sets of each type Update the radio frequency fingerprint database based on the data set; for the first fingerprint data in the boundary data set, determine the first subcluster closest to the location of the first fingerprint data based on the subcluster information, and determine the first subcluster based on the first fingerprint data. The sub-cluster information updates the sub-cluster information of the historical first fingerprint data corresponding to the first fingerprint data in the radio frequency fingerprint database.

In another aspect, the present disclosure also provides a radio frequency fingerprint library updating device configured to update a radio frequency fingerprint library. The historical fingerprint data in the radio frequency fingerprint library includes cluster information and sub-cluster information. The radio frequency fingerprint library updating device includes a data collection device. module, a data processing module and a data update module; the data collection module is configured to obtain real-time fingerprint data; the data processing module is configured to divide the real-time fingerprint data into different types of data sets according to the cluster information, the The data set at least includes a boundary data set; the data update module is configured to update the radio frequency fingerprint database according to each type of data set; for the first fingerprint data in the boundary data set, determine the relationship between the first fingerprint data in the boundary data set and the sub-cluster information. The first sub-cluster where the first fingerprint data is located is the closest, and the sub-cluster information corresponding to the historical first fingerprint data of the first fingerprint data in the radio frequency fingerprint database is updated according to the information of the first sub-cluster.

In another aspect, the present disclosure also provides a computer device, including: at least one processor; and a storage device with at least one computer program stored thereon; when the at least one computer program is executed by the at least one processor, The at least one processor is caused to implement the radio frequency fingerprint database updating method as described above.

In another aspect, the present disclosure also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed, the radio frequency fingerprint library updating method as described above is implemented.

Description of the drawings

Figure 1 is a schematic diagram of the radio frequency fingerprint database update method provided by the present disclosure;

Figure 2 is a schematic flow chart of determining the first sub-cluster closest to the location of the first fingerprint data provided by the present disclosure;

Figure 3 is a schematic flow chart of determining the first sub-cluster closest to the location of the first fingerprint data provided by the present disclosure;

Figure 4 is a schematic flow chart of dividing each cluster of historical fingerprint data in the radio frequency fingerprint database into sub-clusters provided by the present disclosure;

Figure 5 is a schematic flowchart of determining the center point of a non-first subcluster of a cluster provided by the present disclosure;

Figure 6 is a schematic flowchart of dividing real-time fingerprint data into different types of data sets provided by the present disclosure;

Figure 7 is a schematic flowchart of marking real-time fingerprint data provided by the present disclosure;

Figure 8 is a schematic diagram of the radio frequency fingerprint database updating method provided by the present disclosure;

Figure 9 is a schematic structural diagram of the radio frequency fingerprint database updating device provided by the present disclosure; and

Figure 10 is a schematic structural diagram of the radio frequency fingerprint database updating device provided by the present disclosure.

Detailed ways

Example embodiments will be described more fully below with reference to the accompanying drawings, although they may be embodied in different forms and the disclosure should not be construed as limited to the embodiments set forth herein. These embodiments are provided so that this disclosure will be thorough and complete, and will fully understand the scope of the disclosure to those skilled in the art.

As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed items.

The terminology used herein is used to describe particular embodiments only and is not intended to limit the disclosure. As used herein, the singular forms "a," "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that when the terms "comprising" and/or "made of" are used in this specification, it is specified that particular features, integers, steps, operations, elements and/or components are present but not excluded. Add one or more other features, entities, steps, operations, elements, components, and/or groups thereof.

Embodiments described herein may be described with reference to plan and/or cross-sectional illustrations, with the aid of idealized schematic illustrations of the present disclosure. Accordingly, example illustrations may be modified based on manufacturing techniques and/or tolerances. Therefore, the embodiments are not limited to those shown in the drawings but include modifications of configurations formed based on the manufacturing process. Therefore, the regions illustrated in the figures are of a schematic nature, and the shapes of the regions shown in the figures are illustrative of the specific shapes of the regions of the element, but are not limiting.

Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be construed as having the same meaning as in the relevant technical meaning consistent with the meaning within the art and context of the present disclosure, and is not to be construed as having an idealized or overly formal meaning unless expressly so limited herein.

The existing RF fingerprint database update method based on crowdsourced data mainly collects crowdsourced data and establishes an updated fingerprint database, and then filters the actively and passively updated fingerprint databases to update the fingerprint database. This solution mainly focuses on crowdsourcing The processing of the data itself, that is, over-reliance on the crowdsourced data itself, leaves the judgment of the reliability of the data completely to the users. In addition, there is also a fingerprint database update method based on historical data and incremental data, which uses distance thresholds to determine the neighborhoods of clusters. That is, each time a cluster is formed, distance thresholds are used to find the next possible cluster area. ,The disadvantage of this scheme is that when clusters have ,different densities, the setting of the distance threshold will ,change with different clustering densities. This shortcoming also occurs in very high-dimensional data, so the distance threshold becomes difficult to estimate, greatly reducing the reliability of fingerprint data.

In order to solve the problem of poor fingerprint data reliability in the above radio frequency fingerprint database update scheme, embodiments of the present disclosure provide a radio frequency fingerprint database update method. The historical fingerprint data in the radio frequency fingerprint database includes cluster information and sub-cluster information. It should be noted that, One cluster corresponds to one location area. As shown in Figure 1, the radio frequency fingerprint database updating method includes the following steps S1 and S2.

Step S1: Obtain real-time fingerprint data, and divide the real-time fingerprint data into different types of data sets according to cluster information. The data sets at least include boundary data sets.

Real-time fingerprint data is crowdsourcing data (Crowdsourcing Data), and the data in the radio frequency fingerprint database is a crowdsourcing data set. The crowdsourcing data set represents a data set constructed in cooperation with users. Collecting crowdsourced data in real time during the fingerprint data collection process can give users active feedback on positioning results to a certain extent and improve the timeliness of fingerprint data.

In this step, the user terminal updates the real-time fingerprint data to the network side (RF fingerprint database update device), triggering the RF fingerprint database update. The RF fingerprint database update device updates the real-time fingerprint data according to the pre-divided clusters and sub-clusters in the RF fingerprint database. divided into different types of data sets.

In some implementations, the data sets may include strongly correlated data sets, anomaly data sets, and boundary data sets.

Step S2, update the radio frequency fingerprint database according to various types of data sets; for boundary data Concentrate the first fingerprint data, determine the first sub-cluster closest to the location of the first fingerprint data based on the sub-cluster information, and update the historical first fingerprint corresponding to the first fingerprint data in the radio frequency fingerprint database based on the information of the first sub-cluster. Subcluster information of the data.

In this step, for the first fingerprint data in the boundary data set, find the first sub-cluster closest to the location of the first fingerprint data. If the cluster to which the first sub-cluster belongs is the same as the historical first fingerprint data in the radio frequency fingerprint database. clusters are the same, then the first fingerprint data is included in the first sub-cluster, that is, part of the boundary of the first sub-cluster is slightly moved to the location of the first fingerprint data (that is, the boundary of the first sub-cluster is slightly adjusted), Until the first fingerprint data falls into the first sub-cluster, the radio frequency fingerprint database is updated according to the boundary-adjusted information of the first sub-cluster.

In the radio frequency fingerprint database update method provided by the embodiments of the present disclosure, the historical fingerprint data in the radio frequency fingerprint database includes cluster information and sub-cluster information. The method includes: obtaining real-time fingerprint data, and dividing the real-time fingerprint data into different types according to the cluster information. The data set at least includes a boundary data set; and updating the radio frequency fingerprint database according to each type of data set; for the first fingerprint data in the boundary data set, determining the first fingerprint data closest to the location of the first fingerprint data based on the sub-cluster information. sub-cluster, updating the sub-cluster information of the historical first fingerprint data corresponding to the first fingerprint data in the radio frequency fingerprint database according to the information of the first sub-cluster. Embodiments of the present disclosure can perform sub-cluster boundary movement optimization processing on fingerprint data at sub-cluster boundaries that are prone to misjudgment, thereby improving the reliability of fingerprint data in the radio frequency fingerprint database, thereby improving the indoor positioning accuracy of the terminal.

In some embodiments, as shown in Figure 2, determining the first sub-cluster closest to the location of the first fingerprint data based on the sub-cluster information (i.e. step S2) includes the following steps S21 and S22.

Step S21: Determine a second subcluster whose distance from the location of the first fingerprint data is less than a preset first threshold based on the subcluster information.

In this step, the first fingerprint data in the boundary data set are selected in turn. For the current first fingerprint data, multiple second sub-clusters near the location of the current first fingerprint data are found based on the sub-cluster information. That is to say , the distance between the center point of the second sub-cluster and the first fingerprint data is less than the preset first threshold.

Step S22: Calculate the first distance between the first fingerprint data and the center point of each second sub-cluster, and determine based on the first distance and the cluster to which the first fingerprint data belongs in the radio frequency fingerprint database. The first sub-cluster closest to the location of the first fingerprint data.

In this step, the first distance between the center point of each second sub-cluster and the current first fingerprint data is calculated respectively, and based on the first distance, it is determined one by one whether the corresponding sub-cluster is the same as the current first fingerprint data. The first sub-cluster with the closest location.

In some embodiments, as shown in Figure 3, the first sub-cluster closest to the location of the first fingerprint data is determined based on the first distance and the cluster to which the first fingerprint data belongs in the radio frequency fingerprint database (i.e. step S22) includes the following steps S221 to S225.

Step S221: Sort the first distances from small to large to obtain a distance sequence.

Step S222: Determine the current first distance in the order of the first distance in the distance sequence, and determine the cluster to which the third sub-cluster corresponding to the current first distance belongs.

In this step, the current first distance is determined in the order of the first distance in the distance sequence. The current first distance corresponds to the third sub-cluster. The third sub-cluster is determined based on the cluster information and sub-cluster information of the historical fingerprint data in the radio frequency fingerprint database. The cluster to which the cluster belongs. The first fingerprint data includes location information. The clusters and sub-clusters in the radio frequency fingerprint database are divided according to the location information of the fingerprint data. Therefore, in this step, the corresponding cluster can be determined based on the location information of the first fingerprint data.

Step S223, if the cluster to which the third sub-cluster belongs is the same as the cluster to which the historical first fingerprint data in the radio frequency fingerprint database belongs, then step S224 is executed; if the cluster to which the third sub-cluster belongs is the same as the cluster to which the historical first fingerprint data in the radio frequency fingerprint database belongs. clusters are different, execute step S225.

In this step, the cluster to which the third sub-cluster belongs is compared with the cluster to which the historical first fingerprint data in the radio frequency fingerprint database belongs. If the two are consistent, step S224 is executed. The third sub-cluster is the same as the first fingerprint data. The nearest first sub-cluster; if the two are inconsistent, step S225 is executed, that is, the next first distance is selected from the distance sequence, the cluster to which the third sub-cluster corresponding to the currently selected first distance belongs is determined, and the cluster is compared with the radio frequency The cluster to which the corresponding historical first fingerprint data in the fingerprint database belongs is compared until the cluster to which the third sub-cluster corresponding to the currently selected first distance belongs is the same as the cluster to which the historical first fingerprint data in the radio frequency fingerprint database belongs. In this way In this case, the third sub-cluster corresponding to the currently selected first distance is the first sub-cluster closest to the first fingerprint data.

Step S224: Determine the first sub-cluster closest to the first fingerprint data as the third sub-cluster.

Step S225, select the next first distance according to the distance sequence until the cluster to which the third sub-cluster corresponding to the currently selected first distance belongs is the same as the cluster to which the historical first fingerprint data in the radio frequency fingerprint database belongs, and it is determined that it is the same as the first fingerprint. The first sub-cluster with the closest data distance is the third sub-cluster corresponding to the currently selected first distance.

In this step, after selecting the next first distance from the distance sequence, steps S222 and S223 are executed in a loop until the cluster to which the third sub-cluster corresponding to the currently selected first distance belongs matches the historical first distance in the radio frequency fingerprint database. The loop is terminated when the clusters to which the fingerprint data belong are the same. The third sub-cluster corresponding to the first distance selected when the loop terminates is the first sub-cluster closest to the first fingerprint data.

Through the above method, the sub-cluster to which the first fingerprint data in the boundary data set belongs can be slightly adjusted, that is, the sub-cluster information of the first fingerprint data can be adjusted, and the adjusted first fingerprint data can be used to update the radio frequency fingerprint database, thereby improving the accuracy of the radio frequency fingerprint database. Reliability of fingerprint data.

The process of adjusting the sub-cluster information of the first fingerprint data will be described in detail below with reference to a specific example.

The cluster to which a certain first fingerprint data in the boundary data set belongs in the radio frequency fingerprint database is C1, and the first sub-cluster closest to the location of the first fingerprint data is C25. However, due to defects in the radio frequency fingerprint database or changes in actual scenarios, the first sub-cluster C25 belongs to the C2 cluster and does not belong to the C1 cluster. Therefore, it is necessary to continue to search for the closest first sub-cluster to the location of the first fingerprint data. That is, select the next first distance from the distance sequence and determine the third sub-cluster corresponding to the currently selected first distance. Assume that the third sub-cluster at this time is C15, and the cluster to which the third sub-cluster C15 belongs is C1. The historical first fingerprint data corresponding to the first fingerprint data belongs to the same cluster in the radio frequency fingerprint database. Therefore, the first sub-cluster closest to the first fingerprint data is the third sub-cluster C15. At this time, the first fingerprint data needs to be included in the third sub-cluster C15, and the result is that the boundary of the third sub-cluster C15 is finely adjusted until the first fingerprint data is included.

In some embodiments, the radio frequency fingerprint database updating method further includes the following steps: dividing historical fingerprint data in the radio frequency fingerprint database into clusters and dividing each cluster into sub-clusters to generate cluster information and sub-cluster information. It should be noted that the step of dividing the historical fingerprint data in the initial radio frequency fingerprint database into clusters and dividing each cluster into sub-clusters is performed in the initialization phase, which is a preprocessing step for the initial radio frequency fingerprint database. In this step, the initial radio frequency fingerprint data is divided into clusters and each cluster is divided into sub-clusters. Historical fingerprints in the fingerprint database The data is clustered and multiple clusters are obtained, each cluster corresponding to a location area. The user terminal turns on the wireless positioning system and connects to the WIFI signal. The radio frequency fingerprint database update device obtains the fingerprint data to generate an initial radio frequency fingerprint database, and performs initialization processing on the initial radio frequency fingerprint database. The content of the initialization process includes: clustering the initial radio frequency fingerprint database. First, the messy fingerprint data in the initial radio frequency fingerprint database is clustered into multiple clusters. The clusters here can be two-dimensional or three-dimensional. Each cluster corresponds to a location area, that is, clusters and locations form a one-to-one correspondence. For each cluster, secondary clustering is performed on each cluster with the help of differential thinking, and the sub-clusters of each cluster are determined in turn.

In some embodiments, as shown in Figure 4, for each cluster, the step of determining sub-clusters of the cluster includes steps S41 to S43.

Step S41: determine the center points of each sub-cluster of the cluster in sequence.

In some implementations, the center point of each sub-cluster C _ij within the C _i cluster can be calculated according to the following formula (1):

μ _ij is the center point of sub-cluster C _ij , N is the total number of historical fingerprint data in sub-cluster C _ij , X _n is the historical fingerprint data in sub-cluster C _ij , X _n (x, y, z) is _the Location coordinates, n is the identifier of historical fingerprint data, n=(1,2,...,N).

Step S42: For each sub-cluster, determine the range of the sub-cluster based on the center point of the sub-cluster and the preset second threshold.

The preset second threshold is the radius of the sub-cluster, and the area range of the sub-cluster can be determined based on the center point of the sub-cluster and the second threshold.

Step S43: Determine the historical fingerprint data contained in the sub-cluster based on the range of the sub-cluster and the location information of the historical fingerprint data in the cluster.

In this step, historical fingerprint data whose location information falls within the scope of the sub-cluster is selected from the historical fingerprint data of the cluster. These historical fingerprint data are members of the sub-cluster.

In some embodiments, as shown in Figure 5, when the current subcluster is not the first subcluster of the cluster to which it belongs, the step of determining the center point of the current subcluster includes steps S411 to S413.

Step S411, determine the location of each remaining historical fingerprint data in the cluster and the selected The shortest distance between the center points of each sub-cluster.

The remaining historical fingerprint data in the cluster refers to other historical fingerprint data in the cluster except the center point of the sub-cluster. The cluster mentioned here refers to the current sub-cluster C _ij (that is, the sub-cluster to which the sub-cluster center point is to be determined) belongs. Cluster C _i . In this step, for each remaining historical fingerprint data in the cluster, calculate the location of the remaining historical fingerprint data and the center point of each selected sub-cluster (the selected (j-1) sub-cluster center point ), D _kp = (D _k1 , D _k2 ,..., D _k(j-1) ₎ , k is the identification of the remaining historical fingerprint data in the cluster, k = (1,2 ,...,M), M is the total number of remaining historical fingerprint data in the cluster, and p is the identifier of the sub-cluster in the cluster. Determine the minimum value D _k of D _kp , D _k =MIN (D _k1 , D _k2 ,..., D _{k (j-1)} ).

Step S412: According to the shortest distance between the location of each remaining historical fingerprint data in the cluster and the center point of each selected sub-cluster and the distance between each remaining historical fingerprint data in the cluster and the center point of each selected sub-cluster, respectively Calculate the probability that each remaining historical fingerprint data in the cluster serves as the center point of the current sub-cluster.

In this step, first calculate the sum of the distances between each remaining historical fingerprint data in the cluster and the center point of each selected sub-cluster. Then calculate the probability that each remaining historical fingerprint data in the cluster becomes the center point of the current sub-cluster according to the following formula (2):

P _kj is the probability that the remaining historical fingerprint data k in the cluster becomes the center point of the current sub-cluster j.

Step S413: Determine the center point of the current sub-cluster based on the probability that each remaining historical fingerprint data in the cluster is the center point of the current sub-cluster.

In this step, the historical fingerprint data corresponding to the maximum probability value is selected as the center point of the current sub-cluster. In this way, in the process of selecting the center point of the sub-cluster, the larger the distance between the center points of the sub-clusters, the better, so as to avoid adjacent sub-clusters. Clusters overlap.

Taking the remaining two historical fingerprint data in the cluster (historical fingerprint data 1, historical fingerprint data 2) and the selected two sub-cluster center points (sub-cluster center point A, sub-cluster center point B) as an example, determine the third sub-cluster. The process of center point is as follows: for historical fingerprint data 1, calculate the distance D _1A between historical fingerprint data 1 and sub-cluster center point A and the distance D _1B between historical fingerprint data 1 and sub-cluster center point B respectively. Historical fingerprint data 1 with each selected sub-cluster The shortest distance between the center points is D1, D1=MIN(D _1A , D _1B ). For the historical fingerprint data 2, the distance D _2A between the historical fingerprint data 2 and the sub-cluster center point A and the distance D _2B between the historical fingerprint data 2 and the sub-cluster center point B are respectively calculated. The shortest distance between the center points of the sub-cluster is D2, D2=MIN(D _2A , D _2B ). The sum of the distances between the remaining historical fingerprint data in the cluster and the center point of each selected sub-cluster is D=D _1A + D _1B + D _2A + D _2B , and the historical fingerprint data 1 is calculated as the center point of the third sub-cluster. The probability P ₁₃ and _{the probability P 23 of the historical fingerprint data 2 as the center point of the third sub-cluster, P 13} ₌ D1 ² /D ² , P ₂₃ =D2 ² /D ² . Compare P ₁₃ and P ₂₃ , and take the historical fingerprint data corresponding to the maximum probability of P ₁₃ and P ₂₃ as the center point of the third subcluster.

In the embodiment of the present disclosure, for each cluster C _i , sub-clusters are divided through secondary clustering. The purpose is to improve the resolution of fingerprint data within the entire cluster and facilitate subsequent analysis at the junction or boundary of multiple location areas. Make reliable judgments on fingerprint data where misjudgments have occurred.

In some embodiments, when the current subcluster is the first subcluster of the cluster to which it belongs, one historical fingerprint data is randomly selected from the historical fingerprint data in the cluster as the center point of the first subcluster of the cluster. That is to say, if the center point of the first sub-cluster in the cluster is determined, one is directly selected randomly from each historical fingerprint data in the cluster as the center point of the first sub-cluster in the cluster.

In some embodiments, the real-time fingerprint data includes at least location information. As shown in Figure 6, the division of real-time fingerprint data into different types of data sets according to cluster information (ie, step S1) includes the following steps S11 to S15.

Step S11: Determine the cluster to which the historical fingerprint data corresponding to the real-time fingerprint data belongs based on the location information and cluster information of the real-time fingerprint data.

Step S12: Calculate the center point of the cluster and calculate the second distance between the location of the real-time fingerprint data and the center point of the cluster.

In this step, the calculation formula for calculating the center point of the cluster is similar to formula (1). It should be noted that N in formula (1) is the total number of historical fingerprint data in the cluster, and X _n is the total number of historical fingerprint data in the cluster. historical fingerprint data. After calculating the center point of the cluster, calculate the second distance d between the location of the real-time fingerprint data and the center point.

Step S13: Calculate the third distance between the historical fingerprint data of the cluster boundary position and the center point of the cluster.

In this step, multiple historical fingerprint data at the cluster boundary position are determined, and the second distance d between each historical fingerprint data at the cluster boundary position and the center point of the cluster is calculated respectively. The maximum value of the second distance d corresponds to The historical fingerprint data is the historical fingerprint data at the boundary position of the cluster, and the maximum value of the second distance d is the third distance d _{i, furthest} .

Step S14: Mark the real-time fingerprint data according to at least the second distance and the third distance.

In this step, the real-time fingerprint data is marked according to the comparison result of the second distance d and the third distance d _i,furthest . In some implementations, real-time fingerprint data can be labeled as strongly relevant data, anomaly data, and boundary data.

Step S15: Generate a data set based on the labels of the real-time fingerprint data.

The same labeled real-time fingerprint data forms the same data set. Therefore, in some implementations, the real-time fingerprint data can generate a strongly correlated data set, an anomaly data set, and a boundary data set.

In some embodiments, as shown in Figure 7, before marking the real-time fingerprint data according to at least the second distance and the third distance (ie step S14), the radio frequency fingerprint library updating method may further include the following steps: Step S14': Obtain the user's evaluation results on the correctness of the location of the real-time fingerprint data sent by the user terminal.

After the radio frequency fingerprint database is initialized, the wireless positioning system interacts with the baseband of the user terminal to obtain the fingerprint feature information (such as wireless signal strength, amplitude frequency and phase frequency information, etc.) and the corresponding location information in the area where the user terminal is located, and then The location information is presented on the user interface of the user terminal, and the user can make a judgment on the location information. When the location information is consistent with the actual location, the check is correct, and when the location information is inconsistent with the actual location, the check is incorrect. The user terminal uploads the location information and fingerprint feature information of the above crowdsourced data (real-time fingerprint data), as well as the location evaluation results given by the user, to the server on the network side (i.e., the radio frequency fingerprint database update device).

In some embodiments, as shown in FIG. 7 , marking the real-time fingerprint data according to at least the second distance and the third distance (ie, step S14 ) includes the following steps S141 and S142 .

Step S141, in response to the evaluation result being wrong and the second distance being greater than the third distance (d>d _{i, furthest} ), mark the real-time fingerprint data as first abnormal data.

Step S142, filter out second abnormal data from the first abnormal data whose distance from the center point of the cluster to which the real-time fingerprint data belongs is less than a preset third threshold, and mark the second abnormal data as boundary data. The data is boundary data, and the data in the abnormal data set is the data in the first abnormal data except the boundary data.

In some embodiments, as shown in FIG. 7 , marking the real-time fingerprint data according to at least the second distance and the third distance (ie, step S14) also includes the following step S143.

Step S143, in response to the evaluation result being correct and the second distance being less than the third distance (d<d _{i, furthest} ), mark the real-time fingerprint data as strong correlation data, and the data in the strong correlation data set are strong correlation data.

In the embodiment of the present disclosure, a double confirmation mechanism is used to divide the fingerprint data into data sets. If the real-time fingerprint data falls into the corresponding cluster in the radio frequency fingerprint database through double confirmation, this type of real-time fingerprint data is marked as strongly relevant data. A strong correlation data set is formed; if the real-time fingerprint data does not fall into the corresponding cluster in the radio frequency fingerprint database through double confirmation, this type of real-time fingerprint data is marked as the first abnormal data. A second judgment is required for the first abnormal data. If the first abnormal data is at the boundary of the cluster, that is, the distance d between the first abnormal data and the center point of the cluster to which it belongs is approximately equal to the historical fingerprint data at the cluster boundary and the cluster center point. If the third distance d _i,furthest (that is, the second distance d is less than the preset third threshold), this type of real-time fingerprint data is marked as boundary data to form a boundary data set. After the boundary data set is formed, the boundary data set is eliminated from the first abnormal data to obtain the abnormal data set.

In some embodiments, when the data set is an abnormal data set, updating the radio frequency fingerprint database according to each type of data set (ie step S2) includes the following steps: deleting the history corresponding to the abnormal data set in the radio frequency fingerprint database Fingerprint data, that is, the first abnormal data from which the second abnormal data has been eliminated is deleted from the radio frequency fingerprint database.

In some embodiments, when the data set is a strongly correlated data set, updating the radio frequency fingerprint database according to each type of data set (i.e. step S2) includes the following step S2': according to the strongly correlated data in the strongly correlated data set Update the RF fingerprint database, that is, replace the corresponding historical fingerprint data in the RF fingerprint database with strongly correlated data in the strongly correlated data set.

As shown in Figure 8, after updating the radio frequency fingerprint library according to the strong correlation data in the strong correlation data set (ie step S2'), the radio frequency fingerprint library updating method further includes: Go to step S3.

Step S3: Divide the historical fingerprint data in the updated radio frequency fingerprint database into clusters and divide each cluster into sub-clusters to update cluster information and sub-cluster information.

In this step, after updating the RF fingerprint database using the strongly correlated data set, the fingerprint data in the RF fingerprint database is re-divided into clusters and sub-clusters to update the cluster information and sub-cluster information for the next update of the RF fingerprint database. .

Embodiments of the present disclosure collect crowdsourced data during the radio frequency fingerprint data collection process, giving full play to users' active feedback on positioning results to a certain extent, and improving the timeliness of fingerprint data. Compare the user's location judgment results in the crowdsourcing data with the secondary differential clustering results in the preprocessing process of the radio frequency fingerprint database, and perform correlation classification on the crowdsourcing data, especially sub-clustering fingerprint data that is prone to misjudgment. Boundary differential movement optimization processing, thereby improving the reliability of fingerprint data in the radio frequency fingerprint database and the indoor positioning accuracy of the terminal.

Based on the same technical concept, embodiments of the present disclosure also provide a radio frequency fingerprint database update device. As shown in Figure 9, the radio frequency fingerprint database update device is configured to update the radio frequency fingerprint database. The historical fingerprint data in the radio frequency fingerprint database Including cluster information and sub-cluster information, the radio frequency fingerprint database update device includes a data acquisition module 101, a data processing module 102 and a data update module 103.

The data collection module 101 is configured to obtain real-time fingerprint data.

The data processing module 102 is configured to divide the real-time fingerprint data into different types of data sets according to the cluster information, where the data sets at least include boundary data sets.

The data update module 103 is configured to update the radio frequency fingerprint database according to various types of data sets; for the first fingerprint data in the boundary data set, determine the closest fingerprint data to the location of the first fingerprint data based on the sub-cluster information. The first sub-cluster updates the sub-cluster information of the historical first fingerprint data corresponding to the first fingerprint data in the radio frequency fingerprint database according to the information of the first sub-cluster.

In some embodiments, the data update module 103 is configured to determine, based on the sub-cluster information, a second sub-cluster whose distance from the location of the first fingerprint data is less than a preset first threshold; and calculate the relationship between the first fingerprint data and The first distance between the center points of each second sub-cluster is determined based on the first distance and the cluster to which the first fingerprint data belongs in the radio frequency fingerprint database, and the distance from the location of the first fingerprint data is determined. The nearest first subcluster.

In some embodiments, the data update module 103 is configured to sort the first distances from small to large to obtain a distance sequence; determine the current first distance according to the order of the first distances in the distance sequence, and determine the current first distance. The cluster to which the third sub-cluster corresponding to the first distance belongs; in response to the cluster to which the third sub-cluster belongs is the same as the cluster to which the historical first fingerprint data in the radio frequency fingerprint database belongs, it is determined that the cluster to which the first fingerprint data belongs The closest first sub-cluster is the third sub-cluster; and in response to the cluster to which the third sub-cluster belongs is different from the cluster to which the historical first fingerprint data in the radio frequency fingerprint database belongs, select according to the distance sequence The next first distance is determined to be the closest to the first fingerprint data until the cluster to which the third sub-cluster corresponding to the currently selected first distance belongs is the same as the cluster to which the historical first fingerprint data in the radio frequency fingerprint database belongs. The first subcluster of is the third subcluster corresponding to the currently selected first distance.

In some embodiments, as shown in Figure 10, the radio frequency fingerprint database updating device also includes a preprocessing module 104. The preprocessing module 104 is configured to divide the historical fingerprint data in the radio frequency fingerprint database into clusters and divide each cluster into sub-clusters to generate the cluster information and sub-cluster information; for each cluster, determine the sub-clusters of the cluster in turn; determine the sub-clusters of the cluster in the following manner: determine the center point of each sub-cluster of the cluster in turn ; For each sub-cluster, determine the range of the sub-cluster based on the center point of the sub-cluster and the preset second threshold; determine the range of the sub-cluster based on the range of the sub-cluster and the location information of historical fingerprint data in the cluster. Subclusters contain historical fingerprint data.

In some embodiments, the preprocessing module 104 is configured to determine the center point of the current subcluster in the following manner: when the current subcluster is not the first subcluster of the cluster to which it belongs: determining the location of each remaining historical fingerprint data in the cluster. The shortest distance between the location and the center point of each selected sub-cluster; according to the shortest distance between the location of each remaining historical fingerprint data in the cluster and the center point of each selected sub-cluster and each remaining history in the cluster The distance between the fingerprint data and the center point of each selected sub-cluster is used to calculate the probability that each remaining historical fingerprint data in the cluster is the center point of the current sub-cluster; and based on the remaining historical fingerprint data in the cluster, the probability is calculated as the center point of the current sub-cluster. The probability of the center point of a subcluster determines the center point of the current subcluster.

In some embodiments, the preprocessing module 104 is configured to randomly select one historical fingerprint data from the historical fingerprint data in the cluster as the first subcluster of the cluster when the current subcluster is the first subcluster of the cluster to which it belongs. The center point of the cluster.

In some embodiments, the real-time fingerprint data at least includes location information, and the data processing module 102 is configured to determine, based on the location information of the real-time fingerprint data and the cluster information, the historical fingerprint data corresponding to the real-time fingerprint data. Cluster; calculate the center point of the cluster, and calculate the second distance between the location of the real-time fingerprint data and the center point of the cluster; calculate the distance between the historical fingerprint data of the boundary position of the cluster and the center point of the cluster a third distance; labeling the real-time fingerprint data based on at least the second distance and the third distance; and generating a data set based on the labeling of the real-time fingerprint data.

In some embodiments, the data set further includes an abnormal data set, and the data collection module 101 is further configured to mark the real-time fingerprint data according to at least the second distance and the third distance before the data processing module 102 , obtain the user's evaluation result of the correctness of the real-time fingerprint data location sent by the user terminal.

The data processing module 102 is configured to mark the real-time fingerprint data as first abnormal data in response to the evaluation result being wrong and the second distance being greater than the third distance; and filtering from the first abnormal data. Find the second abnormal data whose distance from the center point of the cluster to which the real-time fingerprint data belongs is less than a preset third threshold, and mark the second abnormal data as boundary data, and the data in the boundary data set is the Boundary data, the data in the abnormal data set is the data in the first abnormal data except the boundary data.

In some embodiments, the data update module 103 is configured to delete the historical fingerprint data corresponding to the abnormal data set in the radio frequency fingerprint database when the data set is an abnormal data set.

In some embodiments, the data set further includes a strongly correlated data set, and the data processing module 102 is further configured to, in response to the evaluation result being correct and the second distance being less than the third distance, convert the real-time fingerprint The data is marked as strongly correlated data, and the data in the strongly correlated data set is the strongly correlated data.

In some implementations, the data update module 103 is configured to update the radio frequency fingerprint database according to the strong correlation data in the strongly correlated data set when the data set is a strongly correlated data set.

The preprocessing module 104 is also configured to update the updated radio frequency fingerprint database after the data update module 103 updates the radio frequency fingerprint database according to the strong correlation data in the strong correlation data set. The historical fingerprint data in the texture library is divided into clusters and each cluster is divided into sub-clusters to update the cluster information and sub-cluster information.

An embodiment of the present disclosure also provides a computer device, including: at least one processor; and a storage device; at least one computer program is stored on the storage device, and when the at least one computer program is executed by the at least one processor, the above-mentioned At least one processor implements the radio frequency fingerprint database updating method as provided above.

Embodiments of the present disclosure also provide a computer-readable storage medium on which a computer program is stored. When the computer program is executed, the above-mentioned radio frequency fingerprint database updating method is implemented.

Those of ordinary skill in the art can understand that all or some steps in the methods disclosed above and functional modules/units in the devices can be implemented as software, firmware, hardware, and appropriate combinations thereof. In hardware implementations, the division between functional modules/units mentioned in the above description does not necessarily correspond to the division of physical components; for example, one physical component may have multiple functions, or one function or step may consist of several physical components. Components execute cooperatively. Some or all of the physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or as hardware, or as an integrated circuit, such as an application specific integrated circuit. circuit. Such software may be distributed on computer-readable media, which may include computer storage media (or non-transitory media) and communication media (or transitory media). As is known to those of ordinary skill in the art, the term computer storage media includes volatile and nonvolatile media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. removable, removable and non-removable media. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, Digital Versatile Disk (DVD) or other optical disk storage, magnetic cassettes, tapes, disk storage or other magnetic storage devices, or may Any other medium used to store the desired information and that can be accessed by a computer. Additionally, it is known to those of ordinary skill in the art that communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media .

Example embodiments have been disclosed herein, and although specific terminology is employed, it They are used and shall be construed in a general illustrative sense only and not for purposes of limitation. In some instances, it will be apparent to those skilled in the art that, unless expressly stated otherwise, features, characteristics, and/or elements described in connection with a particular embodiment may be used alone, or may be used with features, characteristics, and/or elements described in connection with other embodiments. and/or used in combination with components. Accordingly, it will be understood by those skilled in the art that various changes in form and details may be made without departing from the scope of the present disclosure as set forth in the appended claims.

Claims

A radio frequency fingerprint database updating method. The historical fingerprint data in the radio frequency fingerprint database includes cluster information and sub-cluster information. The method includes:

Obtaining real-time fingerprint data, and dividing the real-time fingerprint data into different types of data sets according to the cluster information, the data sets at least include boundary data sets; and

The radio frequency fingerprint database is updated according to various types of data sets; wherein, for the first fingerprint data in the boundary data set, the first sub-cluster closest to the location of the first fingerprint data is determined based on the sub-cluster information. , updating the sub-cluster information of the historical first fingerprint data corresponding to the first fingerprint data in the radio frequency fingerprint database according to the information of the first sub-cluster.
The method of claim 1, wherein determining the first sub-cluster closest to the location of the first fingerprint data based on the sub-cluster information includes:

Determine a second subcluster whose distance from the location of the first fingerprint data is less than a preset first threshold based on the subcluster information; and

Calculate the first distance between the first fingerprint data and the center point of each second sub-cluster, and determine the distance between the first fingerprint data and the center point of each second sub-cluster according to the first distance and the cluster to which the first fingerprint data belongs in the radio frequency fingerprint database. The first sub-cluster where the first fingerprint data is located is closest to the first sub-cluster.
The method of claim 2, wherein the closest distance to the location of the first fingerprint data is determined based on the first distance and the cluster to which the first fingerprint data belongs in the radio frequency fingerprint database. The first subcluster includes:

Sort the first distances from small to large to obtain a distance sequence;

Determine the current first distance according to the order of the first distance in the distance sequence, and determine the cluster to which the third sub-cluster corresponding to the current first distance belongs;

In response to the cluster to which the third sub-cluster belongs is the same as the cluster to which the historical first fingerprint data in the radio frequency fingerprint database belongs, it is determined that the first sub-cluster closest to the first fingerprint data is the third sub-cluster. ;as well as

In response to the cluster to which the third sub-cluster belongs and the first historical index in the radio frequency fingerprint database If the cluster to which the fingerprint data belongs is different, the next first distance is selected according to the distance sequence until the cluster to which the third sub-cluster corresponding to the currently selected first distance belongs is the same as the cluster to which the historical first fingerprint data in the radio frequency fingerprint database belongs. Until the same, the first sub-cluster closest to the first fingerprint data is determined to be the third sub-cluster corresponding to the currently selected first distance.
The method of claim 1, further comprising: dividing historical fingerprint data in the radio frequency fingerprint database into clusters and dividing each cluster into sub-clusters to generate the cluster information and sub-cluster information; Each cluster of historical fingerprint data in the database is divided into sub-clusters including:

For each cluster, the sub-clusters of the cluster are determined in turn;

Among them, the sub-clusters of the cluster are determined in the following way:

Determine the center point of each sub-cluster of the cluster in turn;

For each sub-cluster, determine the range of the sub-cluster according to the center point of the sub-cluster and a preset second threshold; and

The historical fingerprint data contained in the sub-cluster is determined according to the range of the sub-cluster and the location information of the historical fingerprint data in the cluster.
The method of claim 4, wherein when the current subcluster is not the first subcluster of the cluster to which it belongs, the center point of the current subcluster is determined in the following manner:

Determine the shortest distance between the location of each remaining historical fingerprint data in the cluster and the center point of each selected sub-cluster;

According to the shortest distance between the location of each remaining historical fingerprint data in the cluster and the center point of each selected sub-cluster and the distance between each remaining historical fingerprint data in the cluster and the center point of each selected sub-cluster, Calculate respectively the probability that each remaining historical fingerprint data in the cluster serves as the center point of the current sub-cluster; and

The center point of the current sub-cluster is determined based on the probability that each remaining historical fingerprint data in the cluster serves as the center point of the current sub-cluster.
The method of claim 4, wherein when the current sub-cluster is the first sub-cluster of the cluster to which it belongs, one historical fingerprint data is randomly selected from the historical fingerprint data in the cluster as the first sub-cluster of the cluster. center point.
The method according to any one of claims 1 to 6, wherein the real-time fingerprint data includes at least location information, and the dividing the real-time fingerprint data into different types of data sets according to the cluster information includes:

According to the location information of the real-time fingerprint data and the cluster information, determine the cluster to which the historical fingerprint data corresponding to the real-time fingerprint data belongs;

Calculate the center point of the cluster, and calculate a second distance between the location of the real-time fingerprint data and the center point of the cluster;

Calculate a third distance between the historical fingerprint data of the boundary position of the cluster and the center point of the cluster;

labeling the real-time fingerprint data based on at least the second distance and the third distance; and

A data set is generated based on the tags of the real-time fingerprint data.
The method of claim 7, wherein the data set further includes an anomaly data set, and the method further includes: performing an operation on the real-time fingerprint data based on at least the second distance and the third distance. Before marking, obtain the user's evaluation results on the correctness of the location of the real-time fingerprint data sent by the user terminal;

Marking the real-time fingerprint data according to at least the second distance and the third distance includes:

In response to the evaluation result being wrong and the second distance being greater than the third distance, marking the real-time fingerprint data as first abnormal data; and

Filter out the second abnormal data from the first abnormal data whose distance from the center point of the cluster to which the real-time fingerprint data belongs is less than a preset third threshold, and mark the second abnormal data as boundary data, so The data in the boundary data set is the boundary data, and the data in the abnormal data set is the data in the first abnormal data except the boundary data.
The method of claim 8, wherein when the data set is an abnormal data set, updating the radio frequency fingerprint database according to each type of data set includes: deleting Remove historical fingerprint data corresponding to the abnormal data set in the radio frequency fingerprint database.
The method of claim 8, wherein the data set further includes a strongly correlated data set, and marking the real-time fingerprint data according to at least the second distance and the third distance further includes:

In response to the evaluation result being correct and the second distance being smaller than the third distance, the real-time fingerprint data is marked as strong correlation data, and the data in the strong correlation data set is the strong correlation data.
The method of claim 10, wherein, when the data set is a strongly correlated data set, updating the radio frequency fingerprint database according to each type of data set includes: according to the strongly correlated data set. Relevant data updates the radio frequency fingerprint database;

The method also includes:

After the radio frequency fingerprint database is updated according to the strong correlation data in the strong correlation data set, the historical fingerprint data in the updated radio frequency fingerprint database is divided into clusters and each cluster is divided into sub-clusters to update the cluster information and sub-clusters. information.
A radio frequency fingerprint library updating device configured to update a radio frequency fingerprint library. The historical fingerprint data in the radio frequency fingerprint library includes cluster information and sub-cluster information. The radio frequency fingerprint library updating device includes a data collection module, a data processing module and a data update module;

The data collection module is configured to obtain real-time fingerprint data;

The data processing module is configured to divide the real-time fingerprint data into different types of data sets according to the cluster information, where the data sets at least include boundary data sets;

The data update module is configured to update the radio frequency fingerprint database according to various types of data sets; wherein, for the first fingerprint data in the boundary data set, the location of the first fingerprint data is determined based on the sub-cluster information. The closest first sub-cluster updates the sub-cluster information of the historical first fingerprint data corresponding to the first fingerprint data in the radio frequency fingerprint database according to the information of the first sub-cluster.
A computer device consisting of:

at least one processor; and

a storage device having at least one computer program stored thereon;

When the at least one computer program is executed by the at least one processor, the at least one processor is caused to implement the radio frequency fingerprint database updating method according to any one of claims 1 to 11.
A computer-readable storage medium on which a computer program is stored. When the computer program is executed, the radio frequency fingerprint database updating method as described in any one of claims 1 to 11 is implemented.