An interference source positioning method and device based on adaptive clustering, a medium and equipment

Abstract

The application discloses a kind of interference source positioning method, device, medium and equipment based on adaptive clustering, the method includes, obtains the interference data of several interference cells in target area, the interference data at least includes cell physical resource block data and cell location information;The interference data is preprocessed, and target training data is obtained;Based on the target training data, adaptive clustering parameters are generated using recurrent neural network;Based on the adaptive clustering parameters, the target training data is clustered using clustering algorithm, and the interference cluster information corresponding to each interference cluster is obtained;Based on each interference cluster information, the area range of each interference cluster is determined, and the interference main cell in each interference cluster is determined, to realize the positioning of interference source.The application improves the effectiveness of interference cell clustering, and improves the accuracy of interference source positioning.

Description

Technical Field

[0001] This invention relates to the fields of wireless communication and data processing technology, and in particular to a method, apparatus, medium and device for locating interference sources based on adaptive clustering. Background Technology

[0002] Currently, in clustering-based solutions for locating external interference sources in problematic cells, the clustering process typically relies on manual pre-determination of the interference type and the setting of corresponding clustering rules based on the determination results. Incorrect interference type determination directly impacts the accuracy of subsequent clustering results. Furthermore, when faced with novel or complex interference types, existing solutions struggle to adjust clustering strategies promptly, leading to low clustering effectiveness and consequently, poor accuracy in interference source location. Summary of the Invention

[0003] The purpose of this invention is to propose an interference source localization method, apparatus, medium, and device based on adaptive clustering. It utilizes a recurrent neural network to generate adaptive clustering parameters from interference data of the interfered cell, and performs adaptive clustering of the interfered cell based on the adaptive clustering parameters to determine the area range corresponding to each interference cluster and the main interfered cell, thereby improving the effectiveness of clustering and increasing the accuracy of interference source localization.

[0004] To achieve the above objectives, a first aspect of the present invention provides a method for locating interference sources based on adaptive clustering, the method comprising: Obtain interference data from several affected cells within the target area, wherein the interference data includes at least cell physical resource block data and cell location information; The interference data is preprocessed to obtain the target training data; Based on the target training data, adaptive clustering parameters are generated using a recurrent neural network. Based on the adaptive clustering parameters, the target training data is clustered using a clustering algorithm to obtain the interference cluster information corresponding to each interference cluster. Based on the information of each interference cluster, the regional range of each interference cluster is determined, and the main cell affected by the interference is identified in each interference cluster, so as to locate the source of interference.

[0005] In this embodiment, interference data of the affected cells is preprocessed, and adaptive clustering parameters are generated using a recurrent neural network. This provides effective clustering hyperparameters for subsequent clustering of affected cells in real time, improving the accuracy of clustering. Then, clustering analysis of the affected cells is performed based on the adaptive clustering parameters, thereby improving the accuracy and precision of locating the interference source.

[0006] Further, the preprocessing of the interference data to obtain the target training data includes: The interference data is cleaned, and the data format and units are standardized to obtain preprocessed data. The data cleaning includes removing outliers, filling in blanks, and removing duplicates. Physical resource block map feature data is extracted from the preprocessed data, and the physical resource block map feature data includes at least time-domain fluctuation dimension feature data and frequency-domain correlation dimension feature data. The physical resource block map feature data is standardized and mapped to a preset numerical range to obtain the target training data.

[0007] In this embodiment, by performing data cleaning, feature extraction, and standardized mapping on the interfering data, the interference of abnormal data, blank data, and duplicate data on subsequent clustering analysis can be reduced, thereby improving the standardization and usability of the target training data and providing a reliable data foundation for subsequent adaptive clustering parameter generation and clustering analysis.

[0008] Further, the adaptive clustering parameters include a spatial distance threshold, a physical resource block temporal volatility threshold, a physical resource block temporal correlation threshold, a physical resource block frequency domain volatility threshold, a physical resource block frequency domain correlation threshold, and the number of interference clusters. The step of generating adaptive clustering parameters using a recurrent neural network based on the target training data includes: The target training data is input into the recurrent neural network for learning, and the spatial distance threshold, the temporal volatility threshold of the physical resource block, the temporal correlation threshold of the physical resource block, the frequency volatility threshold of the physical resource block, the frequency correlation threshold of the physical resource block, and the number of interference clusters are obtained respectively. Wherein, the spatial distance threshold is the farthest spatial distance between each of the interfered cells within the same interference cluster; the physical resource block temporal volatility threshold and the physical resource block frequency volatility threshold are respectively the variance thresholds of the temporal and frequency characteristics of the physical resource blocks within the same interference cluster; the physical resource block temporal correlation threshold and the physical resource block frequency correlation threshold are respectively the similarity thresholds of the temporal and frequency characteristics of the physical resource blocks within the same interference cluster; and the number of interference clusters is the number of target interference clusters corresponding to the data to be clustered.

[0009] In this embodiment, a recurrent neural network is used to adaptively determine clustering constraints based on the target training data, thereby improving the accuracy of subsequent clustering.

[0010] Further, the step of clustering the target training data using a clustering algorithm based on the adaptive clustering parameters to obtain interference cluster information corresponding to each interference cluster includes: Based on spatial, temporal, and frequency domain similarity, the target training data is segmented using the clustering algorithm. Then, based on the spatial distance threshold, the temporal volatility threshold, the temporal correlation threshold, the frequency volatility threshold, and the number of interference clusters, clustering constraints are applied to the segmented data to obtain the interference cluster information for each target interference cluster.

[0011] In this embodiment, by segmenting the target training data according to spatial, temporal, and frequency domain similarity, and combining adaptive clustering parameters to constrain the clustering of the segmented data, the aggregation accuracy of the corresponding interfered cells of the same interference source can be improved from three dimensions: spatial distribution characteristics, temporal variation characteristics, and frequency domain characteristics, thereby improving the accuracy of interference clustering.

[0012] Furthermore, the interference cluster information includes at least regional clustering statistics and interference cluster member information; The regional clustering statistics include at least the province, city, frequency band, number of interfering cells, and number of clusters; The interference cluster member information includes at least the cell identifier, location information, operating frequency band, and cluster number.

[0013] In this embodiment, by dividing the interference cluster information into regional clustering statistics and interference cluster member information, the clustering results can be characterized from the regional statistical level and the intra-cluster member level, respectively. This facilitates hierarchical management and result analysis of interference clusters, thereby improving the interpretability and traceability of interference source location results.

[0014] Furthermore, the step of determining the regional range of each interference cluster based on the information of each interference cluster, and identifying the main cell affected by interference within each interference cluster to locate the interference source, includes: Based on the cluster number corresponding to the member information of each interference cluster, the interfered cells with the same cluster number are classified to obtain the cell details information corresponding to each interference cluster; and based on the location information in the cell details information corresponding to each interference cluster, the area range corresponding to each interference cluster is determined. For each interference cluster, based on the interference severity information of each interfered cell within the interference cluster, the interfered cell with the highest interference severity is selected as the primary interfered cell of the corresponding interference cluster. Based on the area and the affected primary cell, the source of interference is located.

[0015] In this embodiment, the affected cells are classified based on cluster numbers to determine the area range corresponding to each interference cluster. The affected cell with the highest degree of interference in each interference cluster is selected as the main affected cell. This can narrow down the scope of interference source investigation and improve the efficiency of interference source location and investigation.

[0016] Furthermore, the method also includes: Obtain expert annotation information corresponding to historical interference data to construct the training set of the recurrent neural network, wherein the expert annotation information includes interference categories; The expert annotation information is input into the recurrent neural network, and the model parameters of the recurrent neural network are optimized and updated using the backpropagation algorithm based on the training error.

[0017] In this embodiment, a recurrent neural network training set is constructed by introducing expert annotation information corresponding to historical interference data, and the model parameters are optimized and updated using the backpropagation algorithm. This combines historical experience with the model training process, improves the recurrent neural network's ability to learn interference features, and thus enhances the accuracy of adaptive clustering parameter generation.

[0018] To achieve the above objectives, a second aspect of the present invention further provides an interference source localization device based on adaptive clustering, used to implement the interference source localization method based on adaptive clustering as described in any of the first aspects above, the device comprising: The data acquisition module is used to acquire interference data of several interfered cells within the target area. The interference data includes at least cell physical resource block data and cell location information. The data preprocessing module is used to preprocess the interference data to obtain the target training data; The clustering parameter generation module is used to generate adaptive clustering parameters based on the target training data using a recurrent neural network. The clustering module is used to cluster the target training data based on the adaptive clustering parameters and using a clustering algorithm to obtain the interference cluster information corresponding to each interference cluster. The interference source localization module is used to determine the area range of each interference cluster based on the information of each interference cluster, and to identify the main cell affected by interference in each interference cluster, so as to locate the interference source.

[0019] A third aspect of the present invention also provides a computer-readable storage medium, the computer-readable storage medium including a stored computer program; wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform an interference source localization method based on adaptive clustering as described in any of the first aspects.

[0020] A fourth aspect of the present invention also provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements an interference source localization method based on adaptive clustering as described in any of the first aspects above. Attached Figure Description

[0021] Figure 1 This is a flowchart of a preferred embodiment of an interference source localization method based on adaptive clustering provided in the first aspect of the present invention; Figure 2 This is a schematic diagram of the preprocessing of PRB frequency domain features and hourly time domain features of an interfering cell, which is another preferred embodiment of the interference source localization method based on adaptive clustering provided in the first aspect of the present invention. Figure 3 This is a schematic diagram of interference source clustering parameter configuration, representing another preferred embodiment of the interference source localization method based on adaptive clustering provided in the first aspect of the present invention. Figure 4 This is a schematic diagram of the interference source segmentation clustering results of another preferred embodiment of the interference source localization method based on adaptive clustering provided in the first aspect of the present invention; Figure 5 This is a GIS schematic diagram of the interference source localization area, which is another preferred embodiment of the interference source localization method based on adaptive clustering provided in the first aspect of the present invention. Figure 6 This is a structural block diagram of a preferred embodiment of an interference source localization device based on adaptive clustering provided in the second aspect of the present invention; Figure 7 This is a structural block diagram of a preferred embodiment of a terminal device provided in the fourth aspect of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] It should be noted that the data involved in this invention (including but not limited to data used for analysis, data stored, data displayed, etc.) are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of related data must comply with relevant laws, regulations and standards, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0024] In this embodiment of the invention, the words "exemplarily" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplarily" or "for example" in this invention should not be construed as being more preferred or advantageous than other embodiments or designs. Rather, the use of the words "exemplarily" or "for example" is intended to present the relevant concepts in a specific manner.

[0025] In this invention description, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first," "second," "third," etc., may explicitly or implicitly include one or more of that feature. In this invention description, unless otherwise stated, "a plurality of" means two or more. In this invention description, the term "comprising" and its variations are open-ended, meaning "including but not limited to." The term "based on" means "at least partially based on." The term "according to" means "at least partially according to." The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments."

[0026] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0027] In the description of this invention, it should be noted that, unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this specification is for the purpose of describing specific embodiments only and is not intended to limit the invention. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0028] The first aspect of this invention provides a method for locating interference sources based on adaptive clustering, see [link to relevant documentation]. Figure 1 The diagram shown is a flowchart of a preferred embodiment of an interference source localization method based on adaptive clustering provided in the first aspect of the present invention. The method includes steps S1 to S5, as detailed below: Step S1: Obtain interference data of several affected cells within the target area. The interference data includes at least cell physical resource block data and cell location information. In one example, the data source platform corresponding to the target area acts as the sender, transmitting relevant data from several interfered cells within the target area to the data interface module of the interference source location system via a standard data interface. The data interface module parses and caches the received interference data for subsequent processing. The interference data includes at least cell physical resource block data and cell location information, as detailed in Tables 1 and 2.

[0029] Table 1: 4G Raw Data Tags and Indicators

[0030] Table 2: 5G Raw Data Labels and Indicators

[0031] In this embodiment, standardized interference data of the affected cells within the target area is acquired to provide a reliable data foundation for the subsequent positioning process.

[0032] Step S2: Preprocess the interference data to obtain the target training data; In one example, the data preprocessing module performs data cleaning, feature extraction, and standardization mapping on the interfering data to obtain target training data. Specifically, the data preprocessing module can first remove outliers, fill in null values, remove duplicate records, and standardize the data format and units of the input data to obtain preprocessed data; then, it extracts physical resource block map feature data from the preprocessed data; and finally, it maps the feature data of each dimension to a preset numerical range to generate target training data suitable for model training and cluster analysis.

[0033] It should be noted that the standardized mapping can adopt normalization, interval mapping, encoding conversion or other processing methods that can unify the data distribution, and is not limited to a certain fixed implementation method.

[0034] In this embodiment, by uniformly cleaning, extracting features, and standardizing the mapping of interfering data, the impact of abnormal data and dimensional differences on subsequent model analysis can be reduced, providing effective input data for adaptive clustering parameter generation and clustering processing.

[0035] Step S3: Based on the target training data, generate adaptive clustering parameters using a recurrent neural network; In one example, the clustering parameter generation module acts as the receiver, receiving the target training data sent by the data preprocessing module. The target training data is then input into a recurrent neural network. The recurrent neural network learns from the full-area interference data and generates adaptive clustering parameters that fit the current target area. These adaptive clustering parameters may include a spatial distance threshold, a physical resource block temporal volatility threshold, a physical resource block temporal correlation threshold, a physical resource block frequency volatility threshold, a physical resource block frequency correlation threshold, and the number of interference clusters.

[0036] It should be noted that the recurrent neural network can directly generate the adaptive clustering parameters based on the target training data of the current target region, or it can update the parameters based on the historical model and the current data.

[0037] In this embodiment, clustering parameters are adaptively generated based on the target training data using a recurrent neural network. These clustering parameters can constrain the subsequent clustering process, making the subsequent clustering results more consistent with the actual distribution characteristics of the disturbed cells in the current area.

[0038] Step S4: Based on the adaptive clustering parameters, cluster the target training data using a clustering algorithm to obtain the interference cluster information corresponding to each interference cluster; In one example, the clustering module receives adaptive clustering parameters sent by the clustering parameter generation module, and performs segmentation and clustering analysis on the target training data based on the adaptive clustering parameters and the target training data to obtain interference cluster information corresponding to multiple interference clusters. The interference cluster information may include regional clustering statistics and interference cluster member information.

[0039] It should be noted that the clustering algorithm can be the K-Means clustering algorithm or other clustering algorithms that can achieve clustering processing; this invention does not limit the specific clustering algorithm.

[0040] It is understood that the results output by the clustering module are mainly information at the interference cluster level, rather than the final location result of the main cell. The main cell can be further determined in subsequent steps based on the interference cluster information.

[0041] In this embodiment, clustering is performed based on multidimensional similarity and adaptive clustering parameters, which can improve the accuracy of clustering interference clusters.

[0042] Step S5: Based on the information of each interference cluster, determine the area range of each interference cluster and identify the main cell affected by interference in each interference cluster to locate the interference source.

[0043] In one example, the interference source localization module receives interference cluster information sent by the clustering module, and classifies the interfered cells with the same cluster number according to the cluster number in the member information of each interference cluster to determine the regional range of each interference cluster. Subsequently, for each interference cluster, based on the interference degree information of each interfered cell in the interference cluster, the interfered cell with the highest interference degree is selected as the main interfered cell of the corresponding interference cluster, and the interference source localization result is generated based on the regional range and the main interfered cell.

[0044] It should be noted that the interference source location result can be sent by the interference source location module to the result output module or the GIS display module to output the interference cluster investigation plan, the interference cell investigation plan, or the regional GIS display result.

[0045] It should be noted that the interfered main cell can be the cell with the highest degree of interference within the cluster, or it can be a representative cell that meets the preset typicality conditions on the basis of a high degree of interference; the interference source location result can include the area range, the interfered main cell, the coordinates of the investigation location, and the list of cells within the cluster.

[0046] In this embodiment, by determining the area range based on interference cluster information and filtering the main cells affected by interference, the efficiency and accuracy of interference source location are improved, the scope of on-site investigation is narrowed, and more targeted investigation information is provided to frontline personnel.

[0047] In another preferred embodiment, the preprocessing of the interference data to obtain the target training data includes: The interference data is cleaned, and the data format and units are standardized to obtain preprocessed data. The data cleaning includes removing outliers, filling in blanks, and removing duplicates. Physical resource block map feature data is extracted from the preprocessed data, and the physical resource block map feature data includes at least time-domain fluctuation dimension feature data and frequency-domain correlation dimension feature data. The physical resource block map feature data is standardized and mapped to a preset numerical range to obtain the target training data.

[0048] In one example, after receiving raw interference data from several affected cells within the target area, the data interface module sends the raw interference data to the data preprocessing module. The raw interference data may include fields such as province, city, date, cell number, cell CGI, longitude, latitude, operating frequency band, frequency point number, bandwidth, average interference value, hourly granular average interference value, and PRB-level average interference value.

[0049] Subsequently, the data preprocessing module first cleans the raw interference data, removing obviously abnormal field records, filling in missing values ​​for missing fields, and deduplicating duplicate reports. Further, it unifies the format and units of the cleaned data, converting latitude / longitude, frequency band fields, hourly granular interference mean fields, and PRB-level interference mean fields into a unified and computable data format. Then, the data preprocessing module constructs a time-domain interference change sequence based on the hourly granular interference mean, extracting the corresponding time-domain volatility dimension feature data. Simultaneously, it constructs a physical resource block map based on the interference mean values ​​corresponding to multiple PRBs, extracting the corresponding frequency domain correlation dimension feature data. (See [link to relevant documentation]). Figure 2 This diagram illustrates the preprocessing of PRB frequency domain features and hourly time domain features of an interfering cell, representing another preferred embodiment of the interference source localization method based on adaptive clustering provided in the first aspect of this invention. The left-hand graph shows the distribution of interference waveforms of several interfering cells within an interference cluster on different physical resource blocks, while the right-hand graph shows the distribution of interference waveforms of the same interfering cells at different hourly levels. By analyzing the frequency domain and time domain waveforms, corresponding physical resource block spectral feature data can be extracted. This physical resource block spectral feature data includes at least time-domain volatility dimension feature data and frequency-domain correlation dimension feature data.

[0050] Furthermore, the data preprocessing module performs standardized mapping on the extracted feature data of each dimension, mapping features with different dimensions and different value ranges to a unified preset numerical range, generating the target training data, and sending the target training data to the subsequent clustering parameter generation module.

[0051] It should be noted that the original interference data can originate from either 4G or 5G cells.

[0052] It should be noted that the time-domain volatility dimension feature data is used to characterize the interference changes of the interfered cell in different time periods, and the frequency-domain correlation dimension feature data is used to characterize the interference distribution relationship of the interfered cell on different physical resource blocks.

[0053] It is understood that the standardized mapping is not limited to a fixed implementation method, as long as it can convert multidimensional interference features into a unified input form suitable for model training and cluster analysis.

[0054] In this embodiment, by performing data cleaning, feature extraction, and standardized mapping on the original interference data, the impact of abnormal data and dimensional differences on subsequent analysis can be reduced, providing a consistent and reliable data foundation for the recurrent neural network to generate adaptive clustering parameters and for subsequent clustering analysis.

[0055] In another preferred embodiment, the adaptive clustering parameters include a spatial distance threshold, a physical resource block temporal volatility threshold, a physical resource block temporal correlation threshold, a physical resource block frequency domain volatility threshold, a physical resource block frequency domain correlation threshold, and the number of interfering clusters. The step of generating adaptive clustering parameters using a recurrent neural network based on the target training data includes: The target training data is input into the recurrent neural network for learning, and the spatial distance threshold, the temporal volatility threshold of the physical resource block, the temporal correlation threshold of the physical resource block, the frequency volatility threshold of the physical resource block, the frequency correlation threshold of the physical resource block, and the number of interference clusters are obtained respectively. Wherein, the spatial distance threshold is the farthest spatial distance between each of the interfered cells within the same interference cluster; the physical resource block temporal volatility threshold and the physical resource block frequency volatility threshold are respectively the variance thresholds of the temporal and frequency characteristics of the physical resource blocks within the same interference cluster; the physical resource block temporal correlation threshold and the physical resource block frequency correlation threshold are respectively the similarity thresholds of the temporal and frequency characteristics of the physical resource blocks within the same interference cluster; and the number of interference clusters is the number of target interference clusters corresponding to the data to be clustered.

[0056] In one example, after receiving the target training data, the clustering parameter generation module inputs the target training data into a recurrent neural network model. The recurrent neural network model learns the spatial distribution characteristics, temporal variation characteristics, and frequency domain distribution characteristics of the interfering cells within the target area based on multidimensional feature data of the interfering cells, and outputs adaptive clustering parameters to constrain the subsequent clustering process. Specifically, the recurrent neural network model can output spatial distance threshold, physical resource block temporal volatility threshold, physical resource block temporal correlation threshold, physical resource block frequency volatility threshold, physical resource block frequency correlation threshold, and the number of interfering clusters. The spatial distance threshold can be used to constrain the maximum spatial distance between cells within the same interfering cluster; the physical resource block temporal volatility threshold and physical resource block frequency volatility threshold can be used to constrain the degree of volatility of the corresponding features of each cell within the same interfering cluster; the physical resource block temporal correlation threshold and physical resource block frequency correlation threshold can be used to constrain the similarity of the corresponding features of each cell within the same interfering cluster; and the number of interfering clusters can be used to characterize the number of target interfering clusters that should be obtained from the current data to be clustered. The clustering parameter generation module then sends the adaptive clustering parameters to the clustering module for use in subsequent clustering processing.

[0057] See Figure 3This is a schematic diagram of interference source clustering parameter configuration, which is another preferred embodiment of the interference source localization method based on adaptive clustering provided in the first aspect of the present invention. By configuring the above parameters, parameter constraints are provided for subsequent segmentation processing and clustering analysis based on spatial, temporal and frequency domain similarity.

[0058] It should be noted that the adaptive clustering parameters are not fixed preset values, but can be dynamically adjusted as the distribution of data from disturbed cells within the target area changes.

[0059] In this embodiment, a recurrent neural network is used to automatically generate clustering constraint parameters based on the data characteristics of the current target area, thereby improving the matching degree between the subsequent interference cluster division and the actual regional interference distribution.

[0060] In yet another preferred embodiment, the step of clustering the target training data using a clustering algorithm based on the adaptive clustering parameters to obtain interference cluster information corresponding to each interference cluster includes: Based on spatial, temporal, and frequency domain similarity, the target training data is segmented using the clustering algorithm. Then, based on the spatial distance threshold, the temporal volatility threshold, the temporal correlation threshold, the frequency volatility threshold, and the number of interference clusters, clustering constraints are applied to the segmented data to obtain the interference cluster information for each target interference cluster.

[0061] In one example, after receiving the target training data and adaptive clustering parameters, the clustering module first segments the target training data based on spatial, temporal, and frequency similarity. Specifically, the clustering module can prioritize dividing spatially close cells into several candidate data segments according to the spatial proximity relationship between interfered cells within the target area. Then, it further filters the candidate data segments based on their similarity by combining the temporal volatility characteristics and frequency correlation characteristics of the cells within each candidate data segment. Afterward, the clustering module uses the K-Means clustering algorithm to perform clustering processing on the segmented data, introducing adaptive clustering parameters such as spatial distance threshold, physical resource block temporal volatility threshold, physical resource block temporal correlation threshold, physical resource block frequency volatility threshold, physical resource block frequency correlation threshold, and the number of interference clusters to constrain the clustering process. Data points that meet the constraints are grouped into the same target interference cluster. Finally, the clustering module outputs interference cluster information corresponding to multiple target interference clusters and sends this interference cluster information to the interference source localization module.

[0062] See Figure 4 This is a schematic diagram of the interference source segmentation clustering results of another preferred embodiment of the interference source localization method based on adaptive clustering provided in the first aspect of the present invention. Figure 4 The results of the target training data after sharding and clustering analysis are shown in scatter form. Data points of different colors and shapes represent different target interference clusters. Data points within the same target interference cluster are relatively concentrated in the feature space, and there are obvious distinguishing boundaries between different target interference clusters.

[0063] It should be noted that the clustering algorithm described in this embodiment is K-Means clustering algorithm, but it is not limited to K-Means. Any clustering algorithm that can combine the adaptive clustering parameters to complete the clustering of disturbed cells can be applied.

[0064] It is understood that the interference cluster information in this embodiment is mainly cluster-level clustering results, which is used to support the subsequent determination of the area range and the selection of the main cell affected by interference.

[0065] In this embodiment, by combining adaptive clustering parameters with multidimensional similarity in the spatial, temporal, and frequency domains to perform clustering processing, the accuracy of classifying cells affected by the same interference source into the same interference cluster can be improved, thereby enhancing the effectiveness of regional interference identification.

[0066] In yet another preferred embodiment, the interference cluster information includes at least regional clustering statistics and interference cluster member information; The regional clustering statistics include at least the province, city, frequency band, number of interfering cells, and number of clusters; The interference cluster member information includes at least the cell identifier, location information, operating frequency band, and cluster number.

[0067] In one example, after completing the clustering processing of the target training data, the clustering module structures the output results and generates interference cluster information. This interference cluster information can be divided into two categories: regional clustering statistics and interference cluster member information. The regional clustering statistics characterize the overall clustering results for a specific region, city, or frequency band, and may include information such as province, city, frequency band, number of interfering cells, and number of clusters. The interference cluster member information characterizes the cell composition within each target interference cluster, and may include cell identifier, location information, operating frequency band, and cluster number, and may further include detailed member information such as date, frequency point number, and bandwidth. Further, the clustering module organizes the regional clustering statistics into a regional clustering statistics table (see Table 3) and the interference cluster member information into a cluster member detail table (see Table 4), and sends these to the subsequent interference source localization module.

[0068] Table 3: Regional Clustering Statistics

[0069] Table 4: Detailed List of Cluster Member Cells in Interference Regions

[0070] It should be noted that the regional clustering statistics are used to reflect the interference clustering results of a certain region as a whole, and the interference cluster member information is used to reflect the cell composition within a specific interference cluster.

[0071] It should be noted that the cell identifier can be a cell number, CGI, NCGI, or other identifiers that can uniquely identify a cell; the location information can be longitude and latitude, or other information that can characterize spatial location.

[0072] It is understood that, in addition to the aforementioned fields, the interference cluster information may also include other fields related to the clustering results, depending on the actual application needs, without affecting the implementation of this invention.

[0073] In this embodiment, by dividing the clustering results into regional clustering statistics and interference cluster member information, the clustering results can be characterized from both the regional statistical level and the intra-cluster member level, which facilitates subsequent regional range identification and main cell selection.

[0074] In another preferred embodiment, the step of determining the regional range of each interference cluster based on the information of each interference cluster, and identifying the main cell affected by interference within each interference cluster to locate the interference source, includes: Based on the cluster number corresponding to the member information of each interference cluster, the interfered cells with the same cluster number are classified to obtain the cell details information corresponding to each interference cluster; and based on the location information in the cell details information corresponding to each interference cluster, the area range corresponding to each interference cluster is determined. For each interference cluster, based on the interference severity information of each interfered cell within the interference cluster, the interfered cell with the highest interference severity is selected as the primary interfered cell of the corresponding interference cluster. Based on the area and the affected primary cell, the source of interference is located.

[0075] In one example, after receiving the interference cluster member information corresponding to each target interference cluster, the interference source localization module first classifies the cells with the same cluster number according to the cluster number corresponding to each interfered cell, thus obtaining the cell details information corresponding to each target interference cluster. Subsequently, the interference source localization module reads the longitude and latitude information in the cell details information to determine the spatial distribution range of the corresponding target interference cluster and form the corresponding area range.

[0076] Subsequently, for each target interference cluster, the interference source localization module further reads the interference severity information of each affected cell within that cluster, such as the average interference value, hourly granular interference intensity, or other indicators used to characterize the interference severity. From this, it filters out the affected cell with the highest interference severity and identifies it as the primary affected cell of the corresponding target interference cluster. Furthermore, the interference source localization module can jointly organize the area range and the primary affected cell into an interference source localization result, and can further generate an interference cluster investigation plan (see Table 5), an interference cell investigation plan (see Table 6), and GIS display results. Figure 5 This is a GIS schematic diagram of the interference source localization area, which is another preferred embodiment of the interference source localization method based on adaptive clustering provided in the first aspect of the present invention.

[0077] Table 5: Interference Cluster Investigation Plan

[0078] Table 6: Interference Cell Investigation Plan

[0079] It should be noted that the area range can be the area boundary, area outline or area coverage determined based on the location information of cells within the cluster. This invention does not limit the specific form of the area range.

[0080] It should be noted that when multiple affected cells have similar or identical levels of interference, the typicality, representativeness, or centrality of the cells can be further considered to determine the final affected primary cell.

[0081] It is understood that the interference source location result may include only the area range and the main cell affected by the interference, or it may further include the location coordinates, the central cell within the cluster, and a list of interfering cells within the cluster.

[0082] In this embodiment, by first determining the regional range of each interference cluster and then selecting the main cell with the highest degree of interference from each interference cluster, the scope of interference investigation can be narrowed, and a more targeted investigation entry point can be provided for on-site personnel, thereby improving the efficiency of interference source location.

[0083] In yet another preferred embodiment, it further includes: Obtain expert annotation information corresponding to historical interference data to construct the training set of the recurrent neural network, wherein the expert annotation information includes interference categories; The expert annotation information is input into the recurrent neural network, and the model parameters of the recurrent neural network are optimized and updated using the backpropagation algorithm based on the training error.

[0084] In one example, when training a recurrent neural network, the model training module first obtains historical interference data and corresponding expert annotation information from a historical case database, an expert experience database, or other historical interference data storage units. The expert annotation information may include interference categories and expert experience tags related to those categories. The model training module then preprocesses and standardizes the historical interference data to form training input data, and encodes the expert annotation information as supervision information, which, together with the training input data, constitutes the training set of the recurrent neural network. Next, the model training module inputs the training set into the recurrent neural network, obtains the model output through forward computation, and then, based on the training error between the model output and the expert annotation information, uses the backpropagation algorithm to iteratively optimize and update the model parameters of the recurrent neural network to obtain an optimized recurrent neural network model. The optimized model can be deployed to an online clustering parameter generation module for subsequent adaptive clustering parameter generation of the target region.

[0085] It should be noted that the expert annotation information can be generated from manual investigation results, or it can be obtained from existing network optimization experience, historical interference classification records, or expert knowledge rules.

[0086] It should be noted that the training set can be used for offline centralized training, as well as for periodic updates or incremental training after the model is deployed online.

[0087] It is understood that the expert annotation information in this embodiment is mainly used to assist the recurrent neural network in learning different types of interference and their corresponding data distribution patterns, thereby improving the model's adaptability to complex interference scenarios.

[0088] In this embodiment, a training set is constructed by introducing expert annotation information corresponding to historical interference data, and the recurrent neural network is optimized and updated using the backpropagation algorithm. This combines expert experience with the model's self-learning process, improving the model's ability to identify complex interference features and the accuracy of adaptive clustering parameter generation.

[0089] A second aspect of the present invention provides an interference source localization device based on adaptive clustering, used to implement the interference source localization method based on adaptive clustering described in any of the embodiments of the first aspect above. See also... Figure 6 The diagram shown is a structural block diagram of a preferred embodiment of an interference source localization device based on adaptive clustering provided in the second aspect of the present invention. The device includes: Data acquisition module 11 is used to acquire interference data of several interfered cells within the target area. The interference data includes at least cell physical resource block data and cell location information. Data preprocessing module 12 is used to preprocess the interference data to obtain target training data; Clustering parameter generation module 13 is used to generate adaptive clustering parameters based on the target training data using a recurrent neural network. Clustering module 14 is used to cluster the target training data based on the adaptive clustering parameters using a clustering algorithm to obtain the interference cluster information corresponding to each interference cluster; The interference source location module 15 is used to determine the area range of each interference cluster based on the information of each interference cluster, and to determine the main cell affected by interference in each interference cluster, so as to locate the interference source.

[0090] Preferably, the data preprocessing module 12 is further configured to preprocess the interference data through the following steps to obtain the target training data: The interference data is cleaned, and the data format and units are standardized to obtain preprocessed data. The data cleaning includes removing outliers, filling in blanks, and removing duplicates. Physical resource block map feature data is extracted from the preprocessed data, and the physical resource block map feature data includes at least time-domain fluctuation dimension feature data and frequency-domain correlation dimension feature data. The physical resource block map feature data is standardized and mapped to a preset numerical range to obtain the target training data.

[0091] Preferably, the adaptive clustering parameters include a spatial distance threshold, a physical resource block temporal volatility threshold, a physical resource block temporal correlation threshold, a physical resource block frequency domain volatility threshold, a physical resource block frequency domain correlation threshold, and the number of interfering clusters. The clustering parameter generation module 13 is further configured to generate adaptive clustering parameters based on the target training data using a recurrent neural network through the following steps: The target training data is input into the recurrent neural network for learning, and the spatial distance threshold, the temporal volatility threshold of the physical resource block, the temporal correlation threshold of the physical resource block, the frequency volatility threshold of the physical resource block, the frequency correlation threshold of the physical resource block, and the number of interference clusters are obtained respectively. Wherein, the spatial distance threshold is the farthest spatial distance between each of the interfered cells within the same interference cluster; the physical resource block temporal volatility threshold and the physical resource block frequency volatility threshold are respectively the variance thresholds of the temporal and frequency characteristics of the physical resource blocks within the same interference cluster; the physical resource block temporal correlation threshold and the physical resource block frequency correlation threshold are respectively the similarity thresholds of the temporal and frequency characteristics of the physical resource blocks within the same interference cluster; and the number of interference clusters is the number of target interference clusters corresponding to the data to be clustered.

[0092] Preferably, the clustering module 14 is further configured to cluster the target training data based on the adaptive clustering parameters using a clustering algorithm to obtain interference cluster information corresponding to each interference cluster: Based on spatial, temporal, and frequency domain similarity, the clustering algorithm is used to segment the target training data. Then, based on the spatial distance threshold, the temporal volatility threshold, the temporal correlation threshold, the frequency volatility threshold, and the number of interference clusters, clustering constraints are applied to the segmented data to obtain the interference cluster information for each target interference cluster.

[0093] Preferably, the interference cluster information includes at least regional clustering statistics and interference cluster member information; The regional clustering statistics include at least the province, city, frequency band, number of interfering cells, and number of clusters; The interference cluster member information includes at least the cell identifier, location information, operating frequency band, and cluster number.

[0094] Preferably, the interference source localization module 14 is further configured to determine the regional range of each interference cluster based on the information of each interference cluster through the following steps, and determine the main cell affected by interference in each interference cluster, so as to achieve the localization of the interference source: Based on the cluster number corresponding to the member information of each interference cluster, the interfered cells with the same cluster number are classified to obtain the cell details information corresponding to each interference cluster; and based on the location information in the cell details information corresponding to each interference cluster, the area range corresponding to each interference cluster is determined. For each interference cluster, based on the interference severity information of each interfered cell within the interference cluster, the interfered cell with the highest interference severity is selected as the primary interfered cell of the corresponding interference cluster. Based on the area and the affected primary cell, the source of interference is located.

[0095] Preferably, the device further includes an expert pre-training module, specifically comprising: The first expert pre-training module unit is used to obtain expert annotation information corresponding to historical interference data in order to construct the training set of the recurrent neural network. The expert annotation information includes interference categories. The second expert pre-training module unit is used to input the expert annotation information into the recurrent neural network, and to optimize and update the model parameters of the recurrent neural network using the backpropagation algorithm based on the training error.

[0096] It should be noted that the interference source localization device based on adaptive clustering provided in the second aspect embodiment of the present invention can realize all the processes of the interference source localization method based on adaptive clustering described in the first aspect. The functions and technical effects of each module and unit in the device are the same as those of the interference source localization method based on adaptive clustering described in the first aspect embodiment, and will not be repeated here.

[0097] A third aspect of the present invention also provides a computer-readable storage medium, the computer-readable storage medium including a stored computer program; wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform an interference source localization method based on adaptive clustering as described in any of the first aspects above.

[0098] The fourth aspect of the present invention also provides a terminal device, see [link to documentation]. Figure 7 The diagram shown is a structural block diagram of a preferred embodiment of a terminal device provided in the fourth aspect of the present invention. The terminal device includes a processor 10, a memory 20, and a computer program stored in the memory 20 and configured to be executed by the processor 10. When the processor 10 executes the computer program, it implements an interference source localization method based on adaptive clustering as described in any of the above embodiments.

[0099] Preferably, the computer program can be divided into one or more modules / units (such as computer program 1, computer program 2, ...), and the one or more modules / units are stored in the memory 20 and executed by the processor 10 to complete the present invention. The one or more modules / units can be a series of computer program instruction segments capable of performing specific functions, and the instruction segments are used to describe the execution process of the computer program in the terminal device.

[0100] The processor 10 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor, or the processor 10 may be any conventional processor. The processor 10 is the control center of the terminal device, connecting various parts of the terminal device through various interfaces and lines.

[0101] The memory 20 mainly includes a program storage area and a data storage area. The program storage area can store the operating system, applications required for at least one function, etc., while the data storage area can store related data, etc. Furthermore, the memory 20 can be a high-speed random access memory, or a non-volatile memory, such as a plug-in hard drive, a smart media card (SMC), a secure digital card (SD), and a flash card, or other volatile solid-state storage devices.

[0102] It should be noted that the aforementioned terminal device may include, but is not limited to, processors and memory. Those skilled in the art will understand that the above content is merely an example describing the structure of the terminal device and does not constitute a limitation on the structure of the aforementioned terminal device. The aforementioned terminal device may include more or fewer components than those described above, or combine certain components, or different components.

[0103] Through the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus necessary hardware platforms, and of course, it can also be implemented entirely by hardware. Based on this understanding, all or part of the technical solution of the present invention that contributes to the background technology can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM (Read-Only Memory) / RAM (Random Access Memory), magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of the present invention.

[0104] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for locating interference sources based on adaptive clustering, characterized in that, include: Obtain interference data from several affected cells within the target area, wherein the interference data includes at least cell physical resource block data and cell location information; The interference data is preprocessed to obtain the target training data; Based on the target training data, adaptive clustering parameters are generated using a recurrent neural network. Based on the adaptive clustering parameters, the target training data is clustered using a clustering algorithm to obtain the interference cluster information corresponding to each interference cluster. Based on the information of each interference cluster, the regional range of each interference cluster is determined, and the main cell affected by the interference is identified in each interference cluster, so as to locate the source of interference.

2. The interference source localization method based on adaptive clustering as described in claim 1, characterized in that, The preprocessing of the interference data to obtain the target training data includes: The interference data is cleaned, and the data format and units are standardized to obtain preprocessed data. The data cleaning includes removing outliers, filling in blanks, and removing duplicates. Physical resource block map feature data is extracted from the preprocessed data, and the physical resource block map feature data includes at least time-domain fluctuation dimension feature data and frequency-domain correlation dimension feature data. The physical resource block map feature data is standardized and mapped to a preset numerical range to obtain the target training data.

3. The interference source localization method based on adaptive clustering as described in claim 1, characterized in that, The adaptive clustering parameters include a spatial distance threshold, a physical resource block temporal volatility threshold, a physical resource block temporal correlation threshold, a physical resource block frequency domain volatility threshold, a physical resource block frequency domain correlation threshold, and the number of interference clusters. The step of generating adaptive clustering parameters using a recurrent neural network based on the target training data includes: The target training data is input into the recurrent neural network for learning, and the spatial distance threshold, the temporal volatility threshold of the physical resource block, the temporal correlation threshold of the physical resource block, the frequency volatility threshold of the physical resource block, the frequency correlation threshold of the physical resource block, and the number of interference clusters are obtained respectively. Wherein, the spatial distance threshold is the farthest spatial distance between each of the interfered cells within the same interference cluster; the physical resource block temporal volatility threshold and the physical resource block frequency volatility threshold are respectively the variance thresholds of the temporal and frequency characteristics of the physical resource blocks within the same interference cluster; the physical resource block temporal correlation threshold and the physical resource block frequency correlation threshold are respectively the similarity thresholds of the temporal and frequency characteristics of the physical resource blocks within the same interference cluster; and the number of interference clusters is the number of target interference clusters corresponding to the data to be clustered.

4. The interference source localization method based on adaptive clustering as described in claim 3, characterized in that, The step of clustering the target training data using a clustering algorithm based on the adaptive clustering parameters to obtain interference cluster information corresponding to each interference cluster includes: Based on spatial, temporal, and frequency domain similarity, the clustering algorithm is used to segment the target training data. Then, based on the spatial distance threshold, the temporal volatility threshold, the temporal correlation threshold, the frequency volatility threshold, and the number of interference clusters, clustering constraints are applied to the segmented data to obtain the interference cluster information for each target interference cluster.

5. The interference source localization method based on adaptive clustering as described in claim 4, characterized in that, The interference cluster information includes at least regional clustering statistics and interference cluster member information; The regional clustering statistics include at least the province, city, frequency band, number of interfering cells, and number of clusters; The interference cluster member information includes at least the cell identifier, location information, operating frequency band, and cluster number.

6. The interference source localization method based on adaptive clustering as described in claim 5, characterized in that, The step of determining the regional range of each interference cluster based on the information of each interference cluster, and identifying the main cell affected by interference within each interference cluster, in order to locate the source of interference, includes: Based on the cluster number corresponding to the member information of each interference cluster, the interfered cells with the same cluster number are classified to obtain the cell details information corresponding to each interference cluster; and based on the location information in the cell details information corresponding to each interference cluster, the area range corresponding to each interference cluster is determined. For each interference cluster, based on the interference severity information of each interfered cell within the interference cluster, the interfered cell with the highest interference severity is selected as the primary interfered cell of the corresponding interference cluster. Based on the area and the affected primary cell, the source of interference is located.

7. The interference source localization method based on adaptive clustering as described in claim 1, characterized in that, Also includes: Obtain expert annotation information corresponding to historical interference data to construct the training set of the recurrent neural network, wherein the expert annotation information includes interference categories; The expert annotation information is input into the recurrent neural network, and the model parameters of the recurrent neural network are optimized and updated using the backpropagation algorithm based on the training error.

8. An interference source localization device based on adaptive clustering, used to implement the interference source localization method based on adaptive clustering as described in any one of claims 1 to 7, the device comprising: The data acquisition module is used to acquire interference data of several interfered cells within the target area. The interference data includes at least cell physical resource block data and cell location information. The data preprocessing module is used to preprocess the interference data to obtain the target training data; The clustering parameter generation module is used to generate adaptive clustering parameters based on the target training data using a recurrent neural network. The clustering module is used to cluster the target training data based on the adaptive clustering parameters and using a clustering algorithm to obtain the interference cluster information corresponding to each interference cluster. The interference source localization module is used to determine the area range of each interference cluster based on the information of each interference cluster, and to identify the main cell affected by interference in each interference cluster, so as to locate the interference source.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program; wherein, when the computer program is executed, it controls the device on which the computer-readable storage medium is located to perform an interference source localization method based on adaptive clustering as described in any one of claims 1 to 7.

10. A terminal device, characterized in that, The system includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements an interference source localization method based on adaptive clustering as described in any one of claims 1 to 7.

Images