Data processing method and system based on strip mine 5G base station migration

Abstract

The embodiment of the invention provides a data processing method and system based on strip mine 5G base station migration, and the method comprises the steps: obtaining a migration range matrix corresponding to a target base station migration region range associated with the strip mine 5G base station migration, and carrying out the region signaling processing of the signal migration data of a target base station based on the migration range matrix; and furthermore, signal coverage optimization is carried out, so that an optimized coverage map reflecting a signal continuous enhancement effect can be generated. In this way, the accuracy of base station migration is improved, meanwhile, signal coverage is more uniform, the use experience of a 5G network is greatly improved, the automation degree of the base station migration process is high, the labor cost and the error rate are reduced, and the working efficiency is improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of communication optimization technology, and specifically, to a data processing method and system based on 5G base station migration in open-pit mines. Background Art

[0002] In the existing technology, the layout and migration of 5G base stations is a complex process, especially in open-pit mines where the terrain is complex and the signal coverage requirements vary greatly. Traditional base station migration methods mainly rely on manual planning and adjustment, which is not only labor-intensive and inefficient, but also prone to errors. In addition, due to the complexity of the terrain and the influence of wireless signal propagation characteristics, the signal coverage effect after the base station migration is often difficult to achieve as expected, and there are problems such as uneven signal coverage and unstable signal strength.

[0003] On the other hand, although some existing technologies attempt to assist the layout and migration of base stations by collecting signal strength data and generating signal coverage maps, these technologies usually ignore the continuity and consistency of signal strength within the base station migration area, resulting in the generated signal coverage maps not accurately reflecting the actual situation. In addition, these technologies also fail to effectively optimize the signal coverage maps, so the effect in practical applications is not ideal. Summary of the invention

[0004] In order to at least overcome the above-mentioned deficiencies in the prior art, an embodiment of the present application aims to provide a data processing method and system based on open-pit mine 5G base station migration.

[0005] In a first aspect, an embodiment of the present application provides a data processing method based on 5G base station migration in an open-pit mine, the method comprising:

[0006] Obtain a migration range matrix corresponding to the target base station migration area range associated with the open-pit mine 5G base station migration; wherein the migration range matrix is ​​generated by a target signal coverage map corresponding to the target base station migration area range; the target signal coverage map includes multiple coverage areas indicated by the target base station migration area range, and the signal strength value in each coverage area is consistent;

[0007] Performing regional signalization processing on the target base station signal migration data based on the migration range matrix to generate a regional signal map corresponding to the target base station signal migration data; wherein the regional signal map is a map having signal distribution characteristics;

[0008] The signal coverage of the regional signal map is optimized to generate an optimized coverage map corresponding to the target base station signal migration data; wherein the optimized coverage map is used to represent a map having a continuous signal enhancement effect indicated by the target base station migration area range.

[0009] In a possible implementation of the first aspect, obtaining a migration range matrix corresponding to a target base station migration area range associated with an open-pit mine 5G base station migration includes:

[0010] Perform signal strength hierarchical processing on the target signal coverage map corresponding to the target base station migration area to generate a hierarchical signal map;

[0011] Loading the hierarchical signal map into a coverage area feature representation unit in a target machine learning network; wherein the coverage area feature representation unit includes a signal distribution mining subunit, a feature conversion subunit, an attention subunit, and an MLP subunit;

[0012] The signal distribution mining subunit and the feature conversion subunit are used to represent the hierarchical signal spectrum with feature vectors, and the target signal distribution map features are generated;

[0013] Using the attention sub-unit to perform attention processing on the target signal distribution map features to generate a target attention vector;

[0014] The MLP subunit is used to output the features of the target attention vector and generate a migration range matrix corresponding to the migration area range of the target base station.

[0015] In a possible implementation of the first aspect, the signal distribution mining subunit includes X signal clustering analysis layers; the feature conversion subunit includes Y signal strength vector analysis layers; X is a positive integer; Y is a positive integer not greater than X;

[0016] The hierarchical signal spectrum is represented by a feature vector by using a signal distribution mining subunit and a feature conversion subunit to generate target signal distribution map features, including: loading the hierarchical signal spectrum into the signal distribution mining subunit to generate signal clustering distribution vectors respectively obtained by X signal clustering analysis layers in the signal distribution mining subunit;

[0017] Obtaining signal cluster distribution vectors corresponding to Y signal strength vector analysis layers respectively from X signal cluster distribution vectors; wherein the Y signal cluster distribution vectors have different vector lengths;

[0018] Loading the Y signal cluster distribution vectors and the hierarchical signal spectrum into the feature conversion subunit, generating signal strength vectors respectively obtained by the Y signal strength vector analysis layers in the feature conversion subunit; wherein the Y signal strength vectors have different vector lengths;

[0019] The target signal strength vector in the Y signal strength vectors and the target signal cluster distribution vector in the Y signal cluster distribution vectors are aggregated to generate a target signal distribution map feature; wherein the target signal strength vector is the last signal strength vector in the Y signal strength vectors; and the target signal cluster distribution vector is the last signal cluster distribution vector in the Y signal cluster distribution vectors.

[0020] In a possible implementation of the first aspect, the X signal clustering analysis layers include a signal clustering analysis layer Za, where a is a positive integer not greater than X; loading the hierarchical signal spectrum into the signal distribution mining subunit, and generating signal clustering distribution vectors respectively obtained by the X signal clustering analysis layers in the signal distribution mining subunit, including:

[0021] Loading the hierarchical signal map into the signal distribution mining subunit;

[0022] If the signal clustering analysis layer Za is the first signal clustering analysis layer among the X signal clustering analysis layers, the signal clustering analysis layer Za is used to analyze the hierarchical signal spectrum to generate a signal clustering distribution vector obtained by the signal clustering analysis layer Za;

[0023] If the signal clustering analysis layer Za is not the first signal clustering analysis layer among the X signal clustering analysis layers, the signal clustering analysis layer Za is used to analyze the signal clustering distribution vector obtained by the signal clustering analysis layer Za-1 to generate the signal clustering distribution vector obtained by the signal clustering analysis layer Za; wherein the signal clustering analysis layer Za-1 is the previous signal clustering analysis layer of the signal clustering analysis layer Za.

[0024] In a possible implementation of the first aspect, the Y signal strength vector analysis layers include a signal strength vector analysis layer Kb, where b is a positive integer not greater than Y; the Y signal cluster distribution vectors and the hierarchical signal spectrum are loaded into a feature conversion subunit to generate signal strength vectors respectively obtained by the Y signal strength vector analysis layers in the feature conversion subunit, including:

[0025] Loading Y signal cluster distribution vectors and hierarchical signal maps into the feature conversion subunit;

[0026] If the signal strength vector analysis layer Kb is the first signal strength vector analysis layer among the Y signal strength vector analysis layers, the hierarchical signal spectrum is analyzed by using the signal strength vector analysis layer Kb to generate a signal strength vector obtained by the signal strength vector analysis layer Kb;

[0027] If the signal strength vector resolution layer Kb is not the first signal strength vector resolution layer among the Y signal strength vector resolution layers, then the signal cluster distribution vector corresponding to the signal strength vector resolution layer Kb-1 is obtained from the Y signal cluster distribution vectors, the signal cluster distribution vector corresponding to the signal strength vector resolution layer Kb-1 and the signal strength vector obtained by the signal strength vector resolution layer Kb-1 are aggregated to generate an aggregated signal feature vector corresponding to the signal strength vector resolution layer Kb-1, and the aggregated signal feature vector corresponding to the signal strength vector resolution layer Kb-1 is analyzed using the signal strength vector resolution layer Kb to generate a signal strength vector obtained by the signal strength vector resolution layer Kb; wherein, the signal strength vector resolution layer Kb-1 is the previous signal strength vector resolution layer of the signal strength vector resolution layer Kb; the signal cluster distribution vector corresponding to the signal strength vector resolution layer Kb-1 and the signal strength vector obtained by the signal strength vector resolution layer Kb-1 have a consistent vector length.

[0028] In a possible implementation of the first aspect, performing regional signalization processing on target base station signal migration data based on a migration range matrix to generate a regional signal map corresponding to the target base station signal migration data includes:

[0029] Loading the migration range matrix and the target base station signal migration data into a signal graph processing unit in the target machine learning network; wherein the signal graph processing unit includes a signal representation subunit, a migration range embedding subunit and a signal restoration subunit;

[0030] Using the signal representation subunit to extract features from the target base station signal migration data, and generate a base station signal migration vector;

[0031] The filter factor in the migration range embedding subunit is optimized by using the migration range matrix to generate an optimized filter factor, and a coverage area migration vector is generated according to the optimized filter factor and the base station signal migration vector;

[0032] The signal restoration subunit is used to perform signal restoration processing on the coverage area migration vector to generate a regional signal map corresponding to the target base station signal migration data.

[0033] In a possible implementation manner of the first aspect, the migration range embedding subunit includes G embedded filters, where G is a positive integer; the filter factors in the migration range embedding subunit are optimized by using the migration range matrix to generate the optimized filter factors, and the coverage area migration vector is generated according to the optimized filter factors and the base station signal migration vector, including:

[0034] Divide the migration range matrix into migration range units corresponding to G embedded filters respectively; wherein the G embedded filters include an embedded filter Pc, and c is a positive integer not greater than G;

[0035] Optimizing the filter factor in the embedded filter Pc by using the migration range unit corresponding to the embedded filter Pc, and generating the optimized filter factor corresponding to the embedded filter Pc;

[0036] If the embedded filter Pc is the first embedded filter among the G embedded filters, then in the embedded filter Pc, according to the optimized filter factor corresponding to the embedded filter Pc, the base station signal migration vector is parsed to generate an embedded filter vector obtained by the embedded filter Pc;

[0037] If the embedded filter Pc is not the first embedded filter among the G embedded filters, then in the embedded filter Pc, according to the optimized filter factor corresponding to the embedded filter Pc, the embedded filter vector obtained by the embedded filter Pc-1 is parsed to generate the embedded filter vector obtained by the embedded filter Pc; wherein the embedded filter Pc-1 is the previous embedded filter of the embedded filter Pc;

[0038] The embedded filter vector obtained by the last embedded filter among the G embedded filters is used as the coverage area migration vector.

[0039] In a possible implementation manner of the first aspect, the method further includes:

[0040] After detecting a signal restoration instruction for the optimized coverage map, loading the optimized coverage map into a target machine learning network; wherein the target machine learning network includes a migration range signal restoration unit;

[0041] The migration range signal restoration unit is used to perform migration range signal restoration on the optimized coverage map, and a signal restoration map is generated after a signal restoration map corresponding to the optimized coverage map is generated.

[0042] In a possible implementation manner of the first aspect, the method further includes:

[0043] In the basic machine learning network, a template migration range matrix corresponding to the template base station migration area range is obtained; wherein the template migration range matrix is ​​generated by a template target signal coverage map corresponding to the template base station migration area range; the template target signal coverage map includes multiple template coverage areas indicated by the template base station migration area range, and the signal strength value in each template coverage area is consistent;

[0044] Performing regional signalization processing on the template signal map based on the template migration range matrix to generate a template regional signal map corresponding to the template signal map; wherein the template regional signal map is a map having signal distribution characteristics;

[0045] Optimizing signal coverage of the template area signal map to generate a template optimized coverage map corresponding to the template signal map; wherein the template optimized coverage map is used to represent a map having a signal continuous enhancement effect indicated by the template base station migration area range;

[0046] Determine the network error parameters of the basic machine learning network based on the template target signal coverage map, the template signal map and the template optimized coverage map;

[0047] The network parameters in the basic machine learning network are optimized based on the network error parameters. When the optimized basic machine learning network meets the training termination requirements, the optimized basic machine learning network is output as the target machine learning network; wherein the target machine learning network is used to generate an optimized coverage map corresponding to the target base station signal migration data.

[0048] For example, in a possible implementation of the first aspect, the basic machine learning network includes an initial migration range signal restoration unit, a signal coverage optimization unit, a first observation unit, and a second observation unit;

[0049] Based on the template target signal coverage map, template signal map and template optimization coverage map, the network error parameters of the basic machine learning network are determined, including:

[0050] The template target signal coverage map is loaded into the initial migration range signal restoration unit, and the migration range signal is restored by the initial migration range signal restoration unit to the template target signal coverage map, so as to generate a template signal restoration map corresponding to the template target signal coverage map;

[0051] Loading the template target signal coverage map and the template optimized coverage map into the first observation unit, and using the first observation unit to generate a first signal observation confidence corresponding to the template target signal coverage map and a second signal observation confidence corresponding to the template optimized coverage map;

[0052] Loading the template signal map and the template signal restoration map into the second observation unit, and using the second observation unit to generate a third signal observation confidence corresponding to the template signal map and a fourth signal observation confidence corresponding to the template signal restoration map;

[0053] Generate Loss function values ​​of the initial migration range signal restoration unit, the signal coverage optimization unit, the first observation unit, and the second observation unit based on the first signal observation confidence, the second signal observation confidence, the third signal observation confidence, the fourth signal observation confidence, and the GAN Loss function;

[0054] The Loss function value is used as the network error parameter of the basic machine learning network.

[0055] In a possible implementation of the first aspect, the basic machine learning network includes an initial migration range signal restoration unit, a signal coverage optimization unit, and an initial signal map processing unit; determining a network error parameter of the basic machine learning network according to the template target signal coverage map, the template signal map, and the template optimized coverage map, including:

[0056] The template optimized coverage map is loaded into the initial migration range signal restoration unit, and the migration range signal of the template optimized coverage map is restored by using the initial migration range signal restoration unit to generate a reference restoration map corresponding to the template optimized coverage map;

[0057] The template target signal coverage map is loaded into the initial migration range signal restoration unit, and the migration range signal is restored by the initial migration range signal restoration unit to the template target signal coverage map, so as to generate a template signal restoration map corresponding to the template target signal coverage map;

[0058] The template signal restoration map is loaded into the initial signal map processing unit, and in the initial signal map processing unit, the template signal restoration map is subjected to regional signalization processing based on the template migration range matrix to generate a template reference signal map corresponding to the template signal restoration map; the template reference signal map is loaded into the signal coverage optimization unit, and the signal coverage optimization unit is used to optimize the signal coverage of the template reference signal map to generate an optimized signal map corresponding to the template signal restoration map;

[0059] Based on the template signal map and the reference restoration map, a first restoration knowledge vector is generated, and based on the template target signal coverage map and the optimized signal map, a second restoration knowledge vector is generated;

[0060] Based on the first restored knowledge vector and the second restored knowledge vector, a restored consistency error of the basic machine learning network is generated, and the restored consistency error is used as a network error parameter of the basic machine learning network.

[0061] In a possible implementation of the first aspect, the basic machine learning network includes an initial migration range signal restoration unit, a signal coverage optimization unit, and an initial signal graph processing unit;

[0062] Based on the template target signal coverage map, template signal map and template optimization coverage map, the network error parameters of the basic machine learning network are determined, including:

[0063] Loading the template signal map into the initial migration range signal restoration unit, performing migration range signal restoration on the template signal map using the initial migration range signal restoration unit, and generating a first reconstructed signal map corresponding to the template signal map;

[0064] The template target signal coverage map is loaded into the initial signal map processing unit, and in the initial signal map processing unit, the template target signal coverage map is subjected to regional signalization processing based on the template migration range matrix to generate a second reconstructed signal map corresponding to the template target signal coverage map;

[0065] Loading the second reconstructed signal map into the signal coverage optimization unit, using the signal coverage optimization unit to optimize the signal coverage of the second reconstructed signal map, and generating a third reconstructed signal map corresponding to the template target signal coverage spectrum;

[0066] Based on the template signal map and the first reconstructed signal map, a first reconstructed knowledge vector is generated; based on the template target signal coverage map and the third reconstructed signal map, a second reconstructed knowledge vector is generated;

[0067] Based on the first reconstructed knowledge vector and the second reconstructed knowledge vector, a reconstruction error of the basic machine learning network is generated, and the reconstruction error is used as a network error parameter of the basic machine learning network.

[0068] In a possible implementation of the first aspect, the basic machine learning network includes a first observation unit; determining a network error parameter of the basic machine learning network according to a template target signal coverage map, a template signal map, and a template optimization coverage map, including:

[0069] Loading the template target signal coverage map and the template optimization coverage map into the first observation unit, and using the first observation unit to generate a first signal coverage vector corresponding to the template target signal coverage map and a second signal coverage vector corresponding to the template optimization coverage map;

[0070] Performing attention processing on the first signal coverage vector and the second signal coverage vector respectively to generate a first attention knowledge vector corresponding to the first signal coverage vector and a second attention knowledge vector corresponding to the second signal coverage vector;

[0071] Analyze the first attention knowledge vector and the second attention knowledge vector respectively to generate a first weight feature corresponding to the first attention knowledge vector and a second weight feature corresponding to the second attention knowledge vector;

[0072] Acquire an initial key feature corresponding to the template base station migration area range, generate a first deviation error parameter based on the first weight feature and the initial key feature, and generate a second deviation error parameter based on the second weight feature and the initial key feature;

[0073] Based on the first deviation error parameter and the second deviation error parameter, a target deviation error parameter of the basic machine learning network is generated, and the target deviation error parameter is used as a network error parameter of the basic machine learning network.

[0074] In the second aspect, an embodiment of the present application also provides a data processing system based on the migration of 5G base stations in open-pit mines. The data processing system based on the migration of 5G base stations in open-pit mines includes a processor and a machine-readable storage medium. The machine-readable storage medium stores machine executable instructions. The machine executable instructions are loaded and executed by the processor to implement the data processing method based on the migration of 5G base stations in open-pit mines in any possible implementation manner in the first aspect.

[0075] In any of the above aspects, by obtaining the migration range matrix corresponding to the target base station migration area range associated with the open-pit mine 5G base station migration, the target base station signal migration data is processed into regional signals based on the migration range matrix, and the signal coverage is further optimized to generate an optimized coverage map reflecting the continuous signal enhancement effect. This not only improves the accuracy of base station migration, but also makes the signal coverage more uniform, greatly improves the user experience of the 5G network, makes the base station migration process highly automated, reduces labor costs and error rates, and improves work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required to be used in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be extracted in combination with these drawings without creative work.

[0077] Figure 1 A schematic flow chart of a data processing method based on 5G base station migration in an open-pit mine provided by an embodiment of the present invention;

[0078] Figure 2 A schematic block diagram of the structure of a data processing system based on 5G base station migration in an open-pit mine, which is used to implement the above-mentioned data processing method based on 5G base station migration in an open-pit mine, provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0079] The following description is intended to enable one of ordinary skill in the art to implement and incorporate the present invention, and the description is provided in the context of a specific application scenario and its requirements. It will be apparent to one of ordinary skill in the art that various changes may be made to the disclosed embodiments, and that the general principles defined in the present invention may be applied to other embodiments and application scenarios without departing from the principles and scope of the present invention. Therefore, the present invention is not limited to the described embodiments, but should be given the broadest scope consistent with the claims.

[0080] The present invention is described in detail below with reference to the accompanying drawings. The specific operation method in the method embodiment can also be applied to the device embodiment or the system embodiment.

[0081] Figure 1 It is a flow chart of a data processing method based on 5G base station migration in an open-pit mine provided by an embodiment of the present invention. The data processing method based on 5G base station migration in an open-pit mine is introduced in detail below.

[0082] Step S110, obtain the migration range matrix corresponding to the target base station migration area range associated with the open-pit mine 5G base station migration.

[0083] In this embodiment, the migration range matrix is ​​generated by a target signal coverage map corresponding to the target base station migration area range. The target signal coverage map includes multiple coverage areas indicated by the target base station migration area range, and the signal strength value in each coverage area is consistent.

[0084] Among them, suppose an open-pit mine needs to expand production and needs to migrate existing 5G base stations to new areas to support more equipment and personnel. First, the scope of the target base station migration area needs to be determined. This target base station migration area may include different parts of the mine area, such as mining areas, office areas, and residential areas. Then, a migration range matrix can be generated, which will contain information on all coverage areas within the scope of the target base station migration area.

[0085] For example, an open-pit mine refers to a large mine that mines mineral resources through open-pit methods such as digging and mining. 5G base stations are the wireless communication infrastructure of the fifth generation of mobile communication technology (5G), responsible for providing 5G network access services to end users. Compared with 4G base stations, 5G base stations have the advantages of higher data transmission rate, lower latency and larger connection capacity, and can support a large number of IoT devices to access at the same time.

[0086] The target base station relocation area refers to the area in the open-pit mine where the 5G base station needs to be relocated. This area may include different parts of the mine, such as the mining area, office area, and residential area.

[0087] The migration range matrix is ​​a matrix generated by the target signal coverage map corresponding to the target base station migration area range, which contains information about all coverage areas within the target base station migration area range, as well as the signal strength value in each coverage area. The target signal coverage map is a map showing the signal strength values ​​of all coverage areas within the target base station migration area range. The signal strength value in each coverage area is consistent. A coverage area refers to an area within the target base station migration area range, in which the signal strength value is consistent. For example, in an open-pit mine, the mining area, office area, and residential area can all be regarded as different coverage areas. The signal strength value refers to the strength of the wireless signal at a specific location, usually expressed in decibels (dB). The higher the signal strength value, the better the signal quality and the faster the transmission rate. In 5G networks, signal strength values ​​are critical to the stability and performance of the network.

[0088] Step S120 , performing regional signalization processing on the target base station signal migration data based on the migration range matrix, and generating a regional signal map corresponding to the target base station signal migration data.

[0089] After obtaining the migration range matrix, the target base station signal migration data needs to be processed to generate a regional signal map. The regional signal map is a map with signal distribution characteristics. This regional signal map will show the signal strength values ​​in different coverage areas. For example, if the signal strength value in the mining area is low, and the signal strength value in the office area is high, then in the regional signal map, the mining area will be shown as a low signal strength, and the office area will be shown as a high signal strength.

[0090] Exemplarily, the target base station signal migration data refers to signal data related to the target base station migration area, including signal strength value, signal coverage, base station location, etc. These data are important bases for technicians to perform base station migration and signal optimization.

[0091] Regional signal processing refers to dividing the target base station signal migration data according to the coverage area in the migration range matrix, generating the signal strength value in each coverage area, and thus obtaining a regional signal map. The regional signal map refers to a map with signal distribution characteristics, which shows the signal strength values ​​of all coverage areas within the migration area of ​​the target base station. The regional signal map can help technicians understand the distribution of signals within the migration area and provide a basis for subsequent signal optimization. Signal distribution characteristics refer to the strength distribution of wireless signals in a specific area. In the regional signal map, the signal distribution characteristics can use areas of different colors or different grayscales to represent signals of different strengths. For example, areas with higher signal strength values ​​can be represented by darker colors or higher grayscales, and areas with lower signal strength values ​​can be represented by lighter colors or lower grayscales.

[0092] In the regional signal map, the map can be a plane map or a three-dimensional map, which is used to display the terrain, buildings, roads and other information within the target base station migration area. By displaying the signal strength value on the map, technicians can more intuitively understand the distribution of the signal within the migration area.

[0093] Step S130, performing signal coverage optimization on the regional signal map to generate an optimized coverage map corresponding to the target base station signal migration data.

[0094] After the regional signal map is generated, it needs to be optimized for signal coverage to ensure that the signal continues to increase throughout the migration area. The optimized coverage map is used to represent a map with continuous signal enhancement as indicated by the migration area of ​​the target base station. This may involve adding new base stations in areas with weak signal strength or reorienting the signal of existing base stations. After the regional signal map is optimized, an optimized coverage map is generated, which shows the coverage areas with continuous signal enhancement. For example, if the signal strength value in the mining area is found to be low, a new base station may be added in the area, which will increase the signal strength value in the mining area on the optimized coverage map.

[0095] Exemplarily, signal coverage optimization refers to analyzing and adjusting the signal strength value in the regional signal map to improve the coverage and quality of the signal so that the signal is continuously enhanced throughout the migration area. Signal coverage optimization is an important part of base station migration and signal optimization, which can significantly improve the user's communication experience. The optimized coverage map refers to the map obtained after the regional signal map is optimized for signal coverage, showing a map with a continuous signal enhancement effect indicated by the target base station migration area. The optimized coverage map can help technicians understand the effect of signal coverage optimization and provide a basis for subsequent base station migration and signal optimization. The continuous signal enhancement effect means that the signal strength value gradually increases throughout the migration area, and there is no obvious signal blind spot or weak signal area. The continuous signal enhancement effect is the goal of base station migration and signal optimization, which can significantly improve the user's communication experience.

[0096] Based on the above steps, by obtaining the migration range matrix corresponding to the target base station migration area range associated with the open-pit mine 5G base station migration, the target base station signal migration data is processed into regional signals based on the migration range matrix, and the signal coverage is further optimized to generate an optimized coverage map reflecting the continuous signal enhancement effect. This not only improves the accuracy of base station migration, but also makes the signal coverage more uniform, greatly improves the user experience of the 5G network, makes the base station migration process highly automated, reduces labor costs and error rates, and improves work efficiency.

[0097] In a possible implementation, step S110 may include:

[0098] Step S111, performing signal strength hierarchical processing on the target signal coverage map corresponding to the target base station migration area to generate a hierarchical signal map.

[0099] For example, the signal strength values ​​in the target signal coverage map can be divided into different levels according to certain division criteria, such as low, medium, and high levels. By using the graded signal map, it is easier to understand the signal distribution within the migration area.

[0100] Step S112, loading the hierarchical signal spectrum into the coverage area feature representation unit in the target machine learning network. The coverage area feature representation unit includes a signal distribution mining subunit, a feature conversion subunit, an attention subunit and an MLP subunit.

[0101] After generating the hierarchical signal map, the map needs to be loaded into the coverage area feature representation unit in the target machine learning network. This coverage area feature representation unit includes a signal distribution mining subunit, a feature conversion subunit, an attention subunit, and an MLP subunit. The signal distribution mining subunit is responsible for mining the signal distribution information in the hierarchical signal map, the feature conversion subunit is responsible for converting the mined signal distribution information into a feature vector representation, the attention subunit is responsible for performing attention processing on the feature vector representation, and the MLP subunit is responsible for outputting the feature vector representation after attention processing.

[0102] Step S113, using the signal distribution mining subunit and the feature conversion subunit to represent the hierarchical signal spectrum as a feature vector, and generating target signal distribution map features.

[0103] For example, the signal distribution mining subunit mines signal distribution information from the hierarchical signal map, such as the signal strength value, the width of the signal coverage, etc. Then, the feature conversion subunit converts the signal distribution information into a feature vector representation to generate the target signal distribution map features.

[0104] Step S114, using the attention sub-unit to perform attention processing on the target signal distribution map features to generate a target attention vector.

[0105] For example, the attention subunit can be used to perform attention processing on the target signal distribution map features. Attention processing is a machine learning method that can extract the most important and representative features from the target signal distribution map features, thereby improving the accuracy of subsequent feature output. Through attention processing, a target attention vector is generated.

[0106] Step S115, using the MLP sub-unit to perform feature output on the target attention vector, and generate a migration range matrix corresponding to the migration area range of the target base station.

[0107] For example, the MLP subunit is used to perform feature output on the target attention vector to generate a migration range matrix corresponding to the migration area range of the target base station. The MLP subunit is a multi-layer perceptron that can map the target attention vector after attention processing to a high-dimensional space to generate a migration range matrix. The migration range matrix is ​​a matrix that contains information about all coverage areas within the migration area of ​​the target base station and the signal strength value of each coverage area, which can be used in subsequent base station migration and signal optimization steps.

[0108] For example, assuming that the target base station migration area of ​​the open-pit mine is A1-A10, the target signal coverage map of the area may first be processed by signal strength grading. Specifically, the signal strength value may be divided into three levels: low, medium, and high to generate a graded signal map. For example, the area with a signal strength value between 0 and 30 is a low level, the area with a signal strength value between 31 and 60 is a medium level, and the area with a signal strength value between 61 and 100 is a high level.

[0109] The generated hierarchical signal map is loaded into the coverage area feature representation unit in the target machine learning network. The coverage area feature representation unit includes a signal distribution mining subunit, a feature conversion subunit, an attention subunit, and an MLP subunit. The signal distribution mining subunit is responsible for mining the signal distribution information in the hierarchical signal map, the feature conversion subunit is responsible for converting the mined signal distribution information into a feature vector representation, the attention subunit is responsible for performing attention processing on the feature vector representation, and the MLP subunit is responsible for outputting the feature vector representation after attention processing.

[0110] The hierarchical signal map is represented by feature vectors using the signal distribution mining subunit and the feature conversion subunit. Specifically, the signal distribution mining subunit mines signal distribution information from the hierarchical signal map, such as the high and low signal strength values, the width of the signal coverage range, etc. Then, the feature conversion subunit converts these signal distribution information into feature vector representations to generate target signal distribution map features. For example, areas with low, medium, and high signal strength values ​​can be represented by feature vectors [0, 0, 1], [0, 1, 0], and [1, 0, 0], respectively, and areas with different widths of signal coverage ranges can be represented by feature vectors [1, 0] and [0, 1], respectively.

[0111] Next, the server can multiply each feature vector in the target signal distribution map feature by a weight coefficient, and then add all weighted feature vectors to generate a target attention vector. The value of the weight coefficient is usually obtained through network learning training, which can pay more attention to information such as the width of the signal coverage range and the signal strength value.

[0112] In a possible implementation, the signal distribution mining subunit includes X signal clustering analysis layers. The feature conversion subunit includes Y signal strength vector analysis layers. X is a positive integer. Y is a positive integer not greater than X.

[0113] Step S113 may include:

[0114] Step S1131, loading the hierarchical signal spectrum into the signal distribution mining subunit, generating signal cluster distribution vectors respectively obtained by X signal cluster analysis layers in the signal distribution mining subunit.

[0115] For example, suppose that the regional signal map contains 5 regions, each with a signal strength value. The server loads the regional signal map into the signal distribution mining subunit, which contains 3 signal clustering analysis layers. The first signal clustering analysis layer divides the regional signal map into 3 signal clusters and generates 3 signal cluster distribution vectors. The second signal clustering analysis layer further subdivides the 3 signal clusters and generates 6 signal cluster distribution vectors. The third signal clustering analysis layer further subdivides the 6 signal clusters and generates 12 signal cluster distribution vectors.

[0116] Step S1132: acquiring signal cluster distribution vectors corresponding to Y signal strength vector analysis layers respectively from the X signal cluster distribution vectors. The Y signal cluster distribution vectors have different vector lengths.

[0117] For example, the signal cluster distribution vectors corresponding to the three signal strength vector analysis layers are selected from the 12 signal cluster distribution vectors. For example, the first two vectors of the three signal cluster distribution vectors can be selected as the signal cluster distribution vectors corresponding to the three signal strength vector analysis layers.

[0118] Step S1133, Y signal cluster distribution vectors and hierarchical signal spectra are loaded into the feature conversion subunit to generate signal strength vectors respectively obtained by Y signal strength vector analysis layers in the feature conversion subunit. The Y signal strength vectors have different vector lengths.

[0119] For example, three signal cluster distribution vectors and a regional signal map can be loaded into a feature conversion subunit, which includes three signal strength vector parsing layers. Each signal strength vector parsing layer generates a signal strength vector based on the signal distribution in the three signal cluster distribution vectors and the regional signal map. For example, the first signal strength vector parsing layer generates a signal strength vector based on the first two signal cluster distribution vectors and the regional signal map. The second signal strength vector parsing layer generates a signal strength vector based on the last two signal cluster distribution vectors and the regional signal map.

[0120] Step S1134, aggregate the target signal strength vector in the Y signal strength vectors and the target signal cluster distribution vector in the Y signal cluster distribution vectors to generate a target signal distribution map feature. The target signal strength vector is the last signal strength vector in the Y signal strength vectors. The target signal cluster distribution vector is the last signal cluster distribution vector in the Y signal cluster distribution vectors.

[0121] For example, the target signal strength vectors in the Y signal strength vectors and the target signal cluster distribution vectors in the Y signal cluster distribution vectors are aggregated to generate a target signal distribution map feature. For example, if Y=2, the last signal strength vector in the 2 signal strength vectors and the last signal cluster distribution vector in the 2 signal cluster distribution vectors are aggregated to generate a target signal distribution map feature. The aggregation operation can be achieved by performing weighted summation on the target signal strength vector and the target signal cluster distribution vector.

[0122] That is, Y signal cluster distribution vectors and hierarchical signal spectra can be loaded into the feature conversion subunit, which contains Y signal strength vector parsing layers. Each signal strength vector parsing layer generates a signal strength vector based on the signal distribution in the Y signal cluster distribution vectors and hierarchical signal spectra. For example, if Y=2, the server loads 2 signal cluster distribution vectors and hierarchical signal spectra into 2 signal strength vector parsing layers in the feature conversion subunit to generate 2 signal strength vectors respectively. The signal strength vector can be represented as a two-dimensional vector, where each dimension represents the signal strength value in a signal cluster distribution vector.

[0123] On this basis, the Y signal strength vectors are aggregated to generate a target signal strength distribution map corresponding to the target base station migration area. The aggregation operation can merge the signal strength values ​​in the Y signal strength vectors to obtain a signal strength distribution map containing all signal strength values ​​within the target base station migration area. The target signal strength distribution map can be used in subsequent steps such as signal optimization and base station planning to achieve better signal coverage and quality. Specifically, the server can sum the signal strength values ​​at the same position in the Y signal strength vectors to obtain a signal strength value distribution map. Various clustering algorithms and deep learning models can also be used to aggregate signal strength values ​​and generate distribution maps.

[0124] In a possible implementation manner, the X signal clustering analysis layers include a signal clustering analysis layer Za, where a is a positive integer not greater than X.

[0125] Step S1131 may include: loading the hierarchical signal spectrum into the signal distribution mining subunit. If the signal clustering analysis layer Za is the first signal clustering analysis layer among the X signal clustering analysis layers, the signal clustering analysis layer Za is used to analyze the hierarchical signal spectrum to generate the signal clustering distribution vector obtained by the signal clustering analysis layer Za. If the signal clustering analysis layer Za is not the first signal clustering analysis layer among the X signal clustering analysis layers, the signal clustering analysis layer Za is used to analyze the signal clustering distribution vector obtained by the signal clustering analysis layer Za-1 to generate the signal clustering distribution vector obtained by the signal clustering analysis layer Za. The signal clustering analysis layer Za-1 is the previous signal clustering analysis layer of the signal clustering analysis layer Za.

[0126] In a possible implementation manner, the Y signal strength vector analysis layers include a signal strength vector analysis layer Kb, where b is a positive integer not greater than Y.

[0127] Step S1133 may include: loading Y signal cluster distribution vectors and hierarchical signal spectra into the feature conversion subunit. If the signal strength vector analysis layer Kb is the first signal strength vector analysis layer among the Y signal strength vector analysis layers, the hierarchical signal spectrum is analyzed by using the signal strength vector analysis layer Kb to generate a signal strength vector obtained by the signal strength vector analysis layer Kb. If the signal strength vector analysis layer Kb is not the first signal strength vector analysis layer among the Y signal strength vector analysis layers, the signal cluster distribution vector corresponding to the signal strength vector analysis layer Kb-1 is obtained from the Y signal cluster distribution vectors, the signal cluster distribution vector corresponding to the signal strength vector analysis layer Kb-1 and the signal strength vector obtained by the signal strength vector analysis layer Kb-1 are aggregated to generate an aggregated signal feature vector corresponding to the signal strength vector analysis layer Kb-1, and the aggregated signal feature vector corresponding to the signal strength vector analysis layer Kb-1 is analyzed by using the signal strength vector analysis layer Kb to generate a signal strength vector obtained by the signal strength vector analysis layer Kb. The signal strength vector analysis layer Kb-1 is the previous signal strength vector analysis layer of the signal strength vector analysis layer Kb. The signal cluster distribution vector corresponding to the signal strength vector analysis layer Kb-1 and the signal strength vector obtained by the signal strength vector analysis layer Kb-1 have the same vector length.

[0128] For example, if Kb is not the first signal strength vector analysis layer, then it is necessary to obtain the signal cluster distribution vector corresponding to Kb-1 from the Y signal cluster distribution vectors, and then aggregate the signal cluster distribution vector corresponding to the signal strength vector analysis layer Kb-1 and the signal strength vector obtained by the signal strength vector analysis layer Kb-1 to generate an aggregated signal feature vector corresponding to the signal strength vector analysis layer Kb-1. Specifically, a vector aggregation algorithm, such as an average value clustering algorithm, a K-means clustering algorithm, etc., can be used to aggregate the signal cluster distribution vector corresponding to the signal strength vector analysis layer Kb-1 and the signal strength vector obtained by the signal strength vector analysis layer Kb-1 to obtain an aggregated signal feature vector.

[0129] Next, the signal strength vector analysis layer Kb needs to be used to analyze the aggregate signal feature vector corresponding to the signal strength vector analysis layer Kb-1 to generate the signal strength vector obtained by the signal strength vector analysis layer Kb. Specifically, a feature analysis algorithm, such as a neural network algorithm, a decision tree algorithm, etc., can be used to input the aggregate signal feature vector into the signal strength vector analysis layer Kb, perform feature analysis, and obtain the signal strength vector obtained by the signal strength vector analysis layer Kb.

[0130] Through the above steps, feature mining of Y signal strength vector analysis layers can be completed to obtain the signal strength vectors corresponding to each signal strength vector analysis layer. These signal strength vectors can be used in subsequent steps such as signal optimization and base station planning to achieve better signal coverage and quality.

[0131] In a possible implementation, step S120 may include:

[0132] Step S121, loading the migration range matrix and the target base station signal migration data into the signal graph processing unit in the target machine learning network. The signal graph processing unit includes a signal representation subunit, a migration range embedding subunit and a signal restoration subunit.

[0133] Step S122: using the signal representation subunit to perform feature extraction on the target base station signal migration data to generate a base station signal migration vector.

[0134] For example, assuming that the target base station signal migration data includes multiple features such as signal strength, signal quality, and signal interference, these features need to be extracted to generate a base station signal migration vector.

[0135] First, the target base station signal migration data is input into the signal representation subunit. The signal representation subunit includes multiple neural network layers, each of which has multiple neurons. Each neuron receives the output of the neurons in the previous layer, weights them, and then outputs them to the next layer of neurons through an activation function.

[0136] In the signal representation subunit, the first neural network layer will perform one-hot encoding on each feature of the target base station signal migration data to generate a one-hot encoding vector. Then, these one-hot encoding vectors are input to the next neural network layer. The next neural network layer will perform a linear transformation on the output of the previous neural network layer to generate a new feature vector. This process will be repeated until the last neural network layer outputs the base station signal migration vector.

[0137] For example, assuming that the target base station signal migration data includes three features: signal strength, signal quality, and signal interference, the first neural network layer of the signal representation subunit will one-hot encode each feature to generate three one-hot encoded vectors. These one-hot encoded vectors are then input into the next neural network layer. The next neural network layer will linearly transform the output of the previous neural network layer to generate a new feature vector. This process will be repeated until the last neural network layer outputs the base station signal migration vector.

[0138] Finally, the base station signal migration vector output by the signal representation subunit is a vector containing multiple features, each of which represents an aspect of the target base station signal migration data. This vector can be used in subsequent steps such as signal optimization and base station planning to achieve better signal coverage and quality.

[0139] Step S123: Optimize the filter factors in the migration range embedding subunit using the migration range matrix to generate optimized filter factors, and generate coverage area migration vectors based on the optimized filter factors and the base station signal migration vectors.

[0140] For example, first, the migration range matrix and the base station signal migration vector can be input into the migration range embedding subunit. The migration range embedding subunit includes a plurality of filters, each of which has a filter factor. The filter factor is a vector used to filter the input data.

[0141] In the migration range embedding subunit, each filter factor can be optimized using the migration range matrix. Specifically, the migration range matrix and each filter factor can be dot-multiplied to generate a new vector. This new vector is then input into an activation function to generate the optimized filter factor.

[0142] For example, assuming that the migration range matrix is ​​a 3x3 matrix and the filter factor is a 3x1 vector, the server will perform a dot product of the migration range matrix and the filter factor to generate a new 3x1 vector. The server then inputs this new vector into an activation function to generate the optimized filter factor.

[0143] Then, the optimized filter factor and base station signal transition vector are input into the coverage area transition vector generator. The coverage area transition vector generator performs point multiplication on the optimized filter factor and the base station signal transition vector to generate a coverage area transition vector.

[0144] For example, assuming that the optimized filter factor is a 3x1 vector and the base station signal migration vector is a 3x1 vector, then the optimized filter factor and the base station signal migration vector are point-multiplied to generate the coverage area migration vector. Specifically, each element in the optimized filter factor is multiplied by the corresponding element in the base station signal migration vector, and then these products are added to generate an element in the coverage area migration vector.

[0145] For example, assuming that the optimized filter factors are [1, 2, 3] and the base station signal migration vector is [4, 5, 6], then the coverage area migration vector is (1x4)+(2x5)+(3x6)=32.

[0146] Finally, the coverage area migration vector is taken as the output result, and the coverage area migration vector is output. In this way, the migration range matrix can be effectively used to optimize the filter factor, and the coverage area migration vector is generated according to the optimized filter factor and the base station signal migration vector, thereby realizing the regional signalization processing of the base station signal migration data.

[0147] Step S124: Use the signal restoration subunit to perform signal restoration processing on the coverage area migration vector to generate a regional signal map corresponding to the target base station signal migration data.

[0148] For example, first, the coverage area migration vector can be input into the signal restoration subunit. The signal restoration subunit includes multiple neural network layers, each of which has multiple neurons. Each neuron receives the output of the neurons in the previous layer, weights them, and then outputs them to the next layer of neurons through an activation function.

[0149] In the signal restoration subunit, the first neural network layer deconvolves each element of the coverage area migration vector to generate a deconvolved image. These deconvolved images are then input into the next neural network layer. The next neural network layer upsamples the output of the previous neural network layer to generate a new image. This process is repeated until the last neural network layer outputs the regional signal map.

[0150] For example, assuming that the coverage area migration vector is a 32x32 matrix, the first neural network layer of the signal restoration subunit will deconvolve each element to generate a 32x32 deconvolution image. These deconvolution images are then input into the next neural network layer. The next neural network layer will upsample the output of the previous neural network layer to generate a new 32x32 image. This process is repeated until the last neural network layer outputs the regional signal map.

[0151] Finally, the regional signal map output by the signal restoration subunit is a two-dimensional image showing the distribution of the target base station signal migration data in the coverage area. This image can be used in subsequent steps such as signal optimization and base station planning to achieve better signal coverage and quality.

[0152] That is, in this embodiment, the migration range matrix and the target base station signal migration data can be loaded as input data into the signal graph processing unit in the target machine learning network. The signal graph processing unit includes a signal representation subunit, a migration range embedding subunit and a signal restoration subunit.

[0153] Next, the signal representation subunit is used to perform feature extraction on the target base station signal migration data to extract a base station signal migration vector. The base station signal migration vector is a vector that contains all feature information of the target base station signal migration data.

[0154] Then, the migration range matrix is ​​used to optimize the filter factors in the migration range embedding subunit to generate the optimized filter factors. Then, the coverage area migration vector is generated using the optimized filter factors and the base station signal migration vector. The coverage area migration vector is a vector that contains all the characteristic information of the target base station signal migration data in the coverage area.

[0155] Next, the signal restoration subunit performs signal restoration processing on the coverage area migration vector to generate a regional signal map corresponding to the target base station signal migration data. The regional signal map is a two-dimensional image that shows the distribution of the target base station signal migration data in the coverage area.

[0156] Through the above steps, the regional signalization processing of the target base station signal migration data is completed, and the regional signal map corresponding to the target base station signal migration data is generated. These regional signal maps can be used in subsequent steps such as signal optimization and base station planning to achieve better signal coverage and quality.

[0157] In a possible implementation manner, the migration range embedding subunit includes G embedding filters, where G is a positive integer.

[0158] Step S123 may include:

[0159] Step S1231: divide the migration range matrix into migration range units corresponding to G embedded filters respectively. The G embedded filters include an embedded filter Pc, where c is a positive integer not greater than G.

[0160] Step S1232: Optimize the filter factors in the embedded filter Pc by using the migration range unit corresponding to the embedded filter Pc, and generate optimized filter factors corresponding to the embedded filter Pc.

[0161] Step S1233: if the embedded filter Pc is the first embedded filter among the G embedded filters, then in the embedded filter Pc, the base station signal migration vector is parsed according to the optimized filter factor corresponding to the embedded filter Pc to generate an embedded filter vector obtained by the embedded filter Pc.

[0162] Step S1234: If the embedded filter Pc is not the first embedded filter among the G embedded filters, then in the embedded filter Pc, according to the optimized filter factor corresponding to the embedded filter Pc, the embedded filter vector obtained by the embedded filter Pc-1 is parsed to generate the embedded filter vector obtained by the embedded filter Pc. The embedded filter Pc-1 is the previous embedded filter of the embedded filter Pc.

[0163] Step S1235: Use the embedded filter vector obtained by the last embedded filter among the G embedded filters as the coverage area migration vector.

[0164] For example, suppose that the signal coverage of a certain area needs to be analyzed and optimized, and the area is divided into several cells, each of which uses an embedded filter for signal processing. G embedded filters are used, which are represented by Pc, where c is 1, 2, ..., G.

[0165] First, the transition range matrix is ​​divided into transition range units corresponding to G embedded filters respectively. Specifically, a row vector in the transition range matrix can be used as a transition range unit.

[0166] Next, the filter factors in the embedded filters are optimized using the migration range units corresponding to each embedded filter to generate optimized filter factors. Specifically, an optimization algorithm such as a gradient descent method can be used to update the values ​​of the filter factors according to the values ​​in the migration range units, so that the filter factors can better adapt to the distribution of the signal.

[0167] Then, the optimized filter factor is used to analyze the base station signal transition vector to generate an embedded filter vector. Specifically, a mathematical operation such as a convolution operation can be used to convolve the optimized filter factor and the base station signal transition vector to obtain the embedded filter vector.

[0168] If the embedded filter Pc is the first embedded filter among the G embedded filters, the generated embedded filter vector is used as the output of the embedded filter Pc. If the embedded filter Pc is not the first embedded filter, the embedded filter vector obtained by the previous embedded filter Pc-1 is used as input, and the same convolution operation is performed to obtain the embedded filter vector obtained by the embedded filter Pc.

[0169] Finally, the embedded filter vector obtained by the last embedded filter among the G embedded filters is used as the coverage area migration vector. Specifically, the embedded filter vector obtained by the last embedded filter can be used as a representation of the signal coverage of the area for subsequent steps such as signal optimization and base station planning. In this way, the distribution of signals can be better considered and the signal coverage effect can be improved.

[0170] In one possible implementation, after detecting a signal restoration instruction for the optimized coverage map, the optimized coverage map is loaded into the target machine learning network. The target machine learning network includes a migration range signal restoration unit. Then, the migration range signal restoration unit can be used to perform migration range signal restoration on the optimized coverage map, and a signal restoration map corresponding to the optimized coverage map is generated.

[0171] For example, upon detecting a signal restoration instruction for an optimized coverage map, the optimized coverage map is loaded into a target machine learning network. The target machine learning network includes a migration range signal restoration unit.

[0172] The migration range signal restoration unit can be used to restore the migration range signal of the optimized coverage map to generate a signal restoration map corresponding to the optimized coverage map. This process includes the following steps:

[0173] 1. Input the optimized coverage map into the target machine learning network.

[0174] 2. The target machine learning network will optimize each element of the coverage map for deconvolution operations to generate a deconvolved image.

[0175] 3. Input these deconvolution images into the migration range signal restoration unit.

[0176] 4. The migration range signal restoration unit will upsample the output of the previous neural network layer to generate a new image. This process will be repeated until the last neural network layer outputs a signal restoration map.

[0177] 5. Output the generated signal restoration map for subsequent signal optimization and base station planning steps.

[0178] Finally, a signal restoration map corresponding to the optimized coverage map is generated and output. This signal restoration map can be used in subsequent steps such as signal optimization and base station planning to achieve better signal coverage and quality.

[0179] In a possible implementation, the method further includes:

[0180] Step A110, in the basic machine learning network, obtain a template migration range matrix corresponding to the template base station migration area range. The template migration range matrix is ​​generated by a template target signal coverage map corresponding to the template base station migration area range. The template target signal coverage map includes multiple template coverage areas indicated by the template base station migration area range, and the signal strength value in each template coverage area is consistent.

[0181] Step A120: Perform regional signalization processing on the template signal map based on the template migration range matrix to generate a template regional signal map corresponding to the template signal map. The template regional signal map is a map with signal distribution characteristics.

[0182] Step A130, optimize the signal coverage of the template area signal map to generate a template optimized coverage map corresponding to the template signal map. The template optimized coverage map is used to represent a map with a signal continuous enhancement effect indicated by the template base station migration area range.

[0183] Step A140, determining the network error parameters of the basic machine learning network based on the template target signal coverage map, the template signal map and the template optimization coverage map.

[0184] Step A150, optimize the network parameters in the basic machine learning network based on the network error parameters, and when the optimized basic machine learning network meets the training termination requirements, output the optimized basic machine learning network as the target machine learning network. The target machine learning network is used to generate an optimized coverage map corresponding to the target base station signal migration data.

[0185] In one possible implementation, the basic machine learning network includes an initial migration range signal restoration unit, a signal coverage optimization unit, a first observation unit, and a second observation unit.

[0186] Step A140 may include: loading the template target signal coverage map into the initial migration range signal restoration unit, using the initial migration range signal restoration unit to perform migration range signal restoration on the template target signal coverage map, and generating a template signal restoration map corresponding to the template target signal coverage map. Loading the template target signal coverage map and the template optimization coverage map into the first observation unit, using the first observation unit to generate the first signal observation confidence corresponding to the template target signal coverage map, and the second signal observation confidence corresponding to the template optimization coverage map. Loading the template signal map and the template signal restoration map into the second observation unit, using the second observation unit to generate the third signal observation confidence corresponding to the template signal map, and the fourth signal observation confidence corresponding to the template signal restoration map. Based on the first signal observation confidence, the second signal observation confidence, the third signal observation confidence, the fourth signal observation confidence and the GAN Loss function, the Loss function values ​​of the initial migration range signal restoration unit, the signal coverage optimization unit, the first observation unit and the second observation unit are generated. The Loss function value is used as the network error parameter of the basic machine learning network.

[0187] In one possible implementation, the basic machine learning network includes an initial migration range signal restoration unit, a signal coverage optimization unit, and an initial signal graph processing unit.

[0188] Step A140 may include: loading the template optimized coverage map into the initial migration range signal restoration unit, performing migration range signal restoration on the template optimized coverage map using the initial migration range signal restoration unit, and generating a reference restoration map corresponding to the template optimized coverage map. Loading the template target signal coverage map spectrum into the initial migration range signal restoration unit, performing migration range signal restoration on the template target signal coverage map spectrum using the initial migration range signal restoration unit, and generating a template signal restoration map corresponding to the template target signal coverage map spectrum. Loading the template signal restoration map into the initial signal map processing unit, in which the template signal restoration map is subjected to regional signalization processing based on the template migration range matrix, and generating a template reference signal map corresponding to the template signal restoration map. Loading the template reference signal map into the signal coverage optimization unit, performing signal coverage optimization on the template reference signal map using the signal coverage optimization unit, and generating an optimized signal map corresponding to the template signal restoration map. Based on the template signal map and the reference restoration map, a first restoration knowledge vector is generated, and based on the template target signal coverage map spectrum and the optimized signal map, a second restoration knowledge vector is generated. Based on the first restored knowledge vector and the second restored knowledge vector, a restored consistency error of the basic machine learning network is generated, and the restored consistency error is used as a network error parameter of the basic machine learning network.

[0189] In one possible implementation, the basic machine learning network includes an initial migration range signal restoration unit, a signal coverage optimization unit, and an initial signal graph processing unit.

[0190] Step A140 may include: loading the template signal map into the initial migration range signal restoration unit, performing migration range signal restoration on the template signal map using the initial migration range signal restoration unit, and generating a first reconstructed signal map corresponding to the template signal map. Loading the template target signal coverage map into the initial signal map processing unit, performing regional signalization processing on the template target signal coverage map based on the template migration range matrix in the initial signal map processing unit, and generating a second reconstructed signal map corresponding to the template target signal coverage map. Loading the second reconstructed signal map into the signal coverage optimization unit, performing signal coverage optimization on the second reconstructed signal map using the signal coverage optimization unit, and generating a third reconstructed signal map corresponding to the template target signal coverage map. Based on the template signal map and the first reconstructed signal map, a first reconstructed knowledge vector is generated, and based on the template target signal coverage map and the third reconstructed signal map, a second reconstructed knowledge vector is generated. Based on the first reconstructed knowledge vector and the second reconstructed knowledge vector, a reconstruction error of the basic machine learning network is generated, and the reconstruction error is used as the network error parameter of the basic machine learning network.

[0191] In one possible implementation, the basic machine learning network includes a first observation unit.

[0192] Step A140 may include: loading the template target signal coverage map and the template optimization coverage map into the first observation unit, and using the first observation unit to generate the first signal coverage vector corresponding to the template target signal coverage map and the second signal coverage vector corresponding to the template optimization coverage map. Perform attention processing on the first signal coverage vector and the second signal coverage vector respectively to generate the first attention knowledge vector corresponding to the first signal coverage vector and the second attention knowledge vector corresponding to the second signal coverage vector. Analyze the first attention knowledge vector and the second attention knowledge vector respectively to generate the first weight feature corresponding to the first attention knowledge vector and the second weight feature corresponding to the second attention knowledge vector. Obtain the initial key feature corresponding to the template base station migration area range, generate the first deviation error parameter based on the first weight feature and the initial key feature, and generate the second deviation error parameter based on the second weight feature and the initial key feature. Generate the target deviation error parameter of the basic machine learning network based on the first deviation error parameter and the second deviation error parameter, and use the target deviation error parameter as the network error parameter of the basic machine learning network.

[0193] Figure 2 The hardware structure of the data processing system 100 based on 5G base station migration in an open-pit mine for implementing the above-mentioned data processing method based on 5G base station migration in an open-pit mine provided by an embodiment of the present invention is shown as follows: Figure 2 As shown, the data processing system 100 based on open-pit mine 5G base station migration may include a processor 110, a machine-readable storage medium 120, a bus 130 and a communication unit 140.

[0194] In an exemplary design concept, the data processing system 100 based on the migration of 5G base stations in open-pit mines can be a single server or a server group. The server group can be centralized or distributed (for example, the data processing system 100 based on the migration of 5G base stations in open-pit mines can be a distributed system). In an exemplary design concept, the data processing system 100 based on the migration of 5G base stations in open-pit mines can be local or remote. For example, the data processing system 100 based on the migration of 5G base stations in open-pit mines can access information and / or data stored in a machine-readable storage medium 120 via a network. For another example, the data processing system 100 based on the migration of 5G base stations in open-pit mines can be directly connected to the machine-readable storage medium 120 to access the stored information and / or data. In an exemplary design concept, the data processing system 100 based on the migration of 5G base stations in open-pit mines can be implemented on a cloud platform. As an example only, the cloud platform can include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an internal cloud, a multi-layer cloud, etc. or any combination thereof.

[0195] The machine-readable storage medium 120 can store data and / or instructions. In an exemplary design concept, the machine-readable storage medium 120 can store data obtained from an external terminal. In an exemplary design concept, the machine-readable storage medium 120 can store data and / or instructions that the data processing system 100 based on the migration of 5G base stations in an open-pit mine uses to execute or use to complete the exemplary method described in the present invention. In an exemplary design concept, the machine-readable storage medium 120 may include a large-capacity memory, a removable memory, a volatile read-write memory, a read-only memory (ROM), etc. or any combination thereof. Exemplary large-capacity memories may include disks, optical disks, solid-state disks, etc. Exemplary removable memories may include flash drives, floppy disks, optical disks, memory cards, compressed disks, tapes, etc. Exemplary volatile read-write memories may include random access memory (RAM). Exemplary RAM may include active random access memory (DRAM), double data rate synchronous active random access memory (DDR SDRAM), passive random access memory (SRAM), thyristor random access memory (T-RAM), and zero capacitance random access memory (Z-RAM), etc. Exemplary read-only memories may include mask read-only memories (MROMs), programmable read-only memories (PROMs), erasable programmable read-only memories (PEROMs), electrically erasable programmable read-only memories (EEPROMs), compact disk read-only memories (CD-ROMs), and digital versatile disk read-only memories, etc. In an exemplary design concept, the machine-readable storage medium 120 may be implemented on a cloud platform. As an example only, a cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an internal cloud, a multi-layer cloud, etc., or any combination thereof.

[0196] During the specific implementation process, at least one processor 110 executes computer executable instructions stored in the machine-readable storage medium 120, so that the processor 110 can execute the data processing method based on open-pit mine 5G base station migration in the above method embodiment. The processor 110, the machine-readable storage medium 120 and the communication unit 140 are connected through the bus 130, and the processor 110 can be used to control the sending and receiving actions of the communication unit 140.

[0197] The specific implementation process of the processor 110 can be found in the various method embodiments executed by the above-mentioned data processing system 100 based on open-pit mine 5G base station migration. The implementation principles and technical effects are similar, and this embodiment will not be repeated here.

[0198] In addition, an embodiment of the present invention also provides a readable storage medium, in which computer executable instructions are preset. When the processor executes the computer executable instructions, the above-mentioned data processing method based on open-pit mine 5G base station migration is implemented.

[0199] It should be understood that the above description is for illustrative purposes only and is not intended to limit the scope of the present invention. For those of ordinary skill in the art, various modifications and variations can be made in conjunction with the description of the present invention. However, these modifications and variations will not deviate from the scope of the present invention.

[0200] The basic concepts have been described above. Obviously, for those of ordinary skill in the art who have read this application, the above invention disclosure is only for example and does not constitute a limitation of the present invention. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements and corrections to the present invention. Such modifications, improvements and corrections are suggested in the present invention, so such modifications, improvements and corrections still belong to the spirit and scope of the exemplary embodiments of the present invention.

[0201] At the same time, the present invention uses specific words to describe the embodiments of the present invention. For example, "one embodiment", "an embodiment", and / or "some embodiments" refer to a certain feature, structure or characteristic related to at least one embodiment of the present invention. Therefore, it should be emphasized and noted that "one embodiment" or "an embodiment" or "an alternative embodiment" mentioned twice or more in different positions in this specification does not necessarily refer to the same embodiment. In addition, some features, structures or characteristics in one or more embodiments of the present invention can be appropriately combined.

[0202] In addition, it will be appreciated by those skilled in the art that various aspects of the present invention may be illustrated and described in conjunction with a plurality of patentable categories or situations, including any new and useful process, machine, product, or combination of substances, or any new and useful improvements thereto. Accordingly, various aspects of the present invention may be performed entirely by hardware, entirely by software (including firmware, resident software, microcode, etc.), or by a combination of hardware and software. The above hardware or software may all be referred to as "units," "modules," or "systems." In addition, various aspects disclosed in the present invention may take the form of a computer program product embodied in one or more computer-readable media, wherein computer-readable program code is contained therein.

[0203] A computer readable signal medium may include a propagated data signal containing computer program code, such as in baseband or as part of a carrier wave. Such propagated signals may be in a variety of forms, including electromagnetic, optical, etc., or any suitable combination. A computer readable signal medium may be any computer readable medium other than a computer readable storage medium, which may be coupled to an instruction execution system, device, or apparatus to communicate, propagate, or transmit a program for use. The program code on the computer readable signal medium may be transmitted in conjunction with any suitable medium, including radio, cable, fiber optic cable, RF, or the like, or any combination thereof.

[0204] The computer program codes required for the operation of each part of the present invention can be written in any one or more programming languages, including subject-oriented programming languages ​​such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB.NET, Python, etc., conventional procedural programming languages ​​such as C language, Visual Basic, Fortran 2003, Perl, COBOL 2002, PHP, ABAP, active programming languages ​​such as Python, Ruby and Groovy, or other programming languages. The program code can be run entirely on the power operation partition computer, or as an independent software package on the power operation partition computer, or partly on the power operation partition computer and partly on the remote computer, or completely on the remote computer or server. In the latter case, the remote computer can be connected to the power operation partition computer in combination with any network form, such as a local area network (LAN) or a wide area network (WAN), or connected to an external computer (for example, in combination with the Internet), or in a cloud computing environment, or as a service such as software as a service (SaaS).

[0205] In addition, unless expressly stated in the claims, the order of processing elements and sequences of the present invention, the use of alphanumeric characters, or the use of other names are not intended to limit the order of the processes and methods of the present invention. Although the above disclosure discusses some invention embodiments currently considered useful in conjunction with various examples, it should be understood that such details are only for illustrative purposes, and the attached claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that match the essence and scope of the embodiments of the present invention. For example, although the system components described above can be implemented in conjunction with hardware devices, they can also be implemented in conjunction with software solutions alone, such as installing the described system on an existing server or mobile device.

[0206] Similarly, it should be noted that in order to simplify the description of the present invention disclosure and thus help understand one or more embodiments of the invention, in the foregoing description of the embodiments of the invention, sometimes multiple features are grouped into one embodiment, drawings, or descriptions thereof. Similarly, it should be noted that in order to simplify the description of the present invention disclosure and thus help understand one or more embodiments of the invention, in the foregoing description of the embodiments of the invention, sometimes multiple features are grouped into one embodiment, drawings, or descriptions thereof.

Claims

1. A data processing method based on 5G base station migration in an open-pit mine, characterized in that: The method comprises: Obtain a migration range matrix corresponding to a target base station migration area range associated with an open-pit mine 5G base station migration; wherein the migration range matrix is ​​generated by a target signal coverage map corresponding to the target base station migration area range; the target signal coverage map includes multiple coverage areas indicated by the target base station migration area range, and the signal strength value in each coverage area is consistent; Performing regional signalization processing on the target base station signal migration data based on the migration range matrix to generate a regional signal map corresponding to the target base station signal migration data; wherein the regional signal map is a map having signal distribution characteristics; The signal coverage of the regional signal map is optimized to generate an optimized coverage map corresponding to the target base station signal migration data; wherein the optimized coverage map is used to represent a map having a continuous signal enhancement effect indicated by the target base station migration area range.

2. The data processing method based on open-pit mine 5G base station migration according to claim 1 is characterized in that: Obtain the migration range matrix corresponding to the target base station migration area range associated with the open-pit mine 5G base station migration, including: Performing signal strength hierarchical processing on the target signal coverage map corresponding to the target base station migration area to generate a hierarchical signal map; Loading the hierarchical signal map into a coverage area feature representation unit in a target machine learning network; wherein the coverage area feature representation unit includes a signal distribution mining subunit, a feature conversion subunit, an attention subunit, and an MLP subunit; Using the signal distribution mining subunit and the feature conversion subunit to represent the hierarchical signal spectrum with a feature vector, and generating target signal distribution map features; Using the attention subunit to perform attention processing on the target signal distribution map features to generate a target attention vector; The MLP subunit is used to perform feature output on the target attention vector to generate the migration range matrix corresponding to the migration area range of the target base station.

3. The data processing method based on open-pit mine 5G base station migration according to claim 2 is characterized in that: The signal distribution mining subunit includes X signal clustering analysis layers; the feature conversion subunit includes Y signal strength vector analysis layers; X is a positive integer; Y is a positive integer not greater than X; the signal distribution mining subunit and the feature conversion subunit are used to represent the hierarchical signal spectrum by feature vectors to generate target signal distribution map features, including: Loading the hierarchical signal spectrum into the signal distribution mining subunit, generating signal clustering distribution vectors respectively obtained by the X signal clustering analysis layers in the signal distribution mining subunit; Obtaining signal cluster distribution vectors corresponding to the Y signal strength vector analysis layers respectively from the X signal cluster distribution vectors; wherein the Y signal cluster distribution vectors have different vector lengths; Loading the Y signal cluster distribution vectors and the hierarchical signal spectrum into the feature conversion subunit, generating signal strength vectors respectively obtained by the Y signal strength vector analysis layers in the feature conversion subunit; wherein the Y signal strength vectors have different vector lengths; Aggregate the target signal strength vector among the Y signal strength vectors and the target signal cluster distribution vector among the Y signal cluster distribution vectors to generate a target signal distribution graph feature; wherein the target signal strength vector is the last signal strength vector among the Y signal strength vectors; and the target signal cluster distribution vector is the last signal cluster distribution vector among the Y signal cluster distribution vectors.

4. The data processing method based on open-pit mine 5G base station migration according to claim 3 is characterized in that: The X signal clustering analysis layers include a signal clustering analysis layer Za, where a is a positive integer not greater than X; the step of loading the hierarchical signal spectrum into the signal distribution mining subunit to generate signal clustering distribution vectors respectively obtained by the X signal clustering analysis layers in the signal distribution mining subunit includes: Loading the hierarchical signal map into the signal distribution mining subunit; If the signal clustering analysis layer Za is the first signal clustering analysis layer among the X signal clustering analysis layers, the hierarchical signal spectrum is analyzed by using the signal clustering analysis layer Za to generate a signal clustering distribution vector obtained by the signal clustering analysis layer Za; If the signal clustering analysis layer Za is not the first signal clustering analysis layer among the X signal clustering analysis layers, the signal clustering analysis layer Za is used to analyze the signal clustering distribution vector obtained by the signal clustering analysis layer Za-1 to generate the signal clustering distribution vector obtained by the signal clustering analysis layer Za; wherein the signal clustering analysis layer Za-1 is the previous signal clustering analysis layer of the signal clustering analysis layer Za.

5. The data processing method based on open-pit mine 5G base station migration according to claim 3 is characterized in that: The Y signal strength vector analysis layers include a signal strength vector analysis layer Kb, where b is a positive integer not greater than Y; loading the Y signal cluster distribution vectors and the hierarchical signal spectrum into the feature conversion subunit to generate signal strength vectors respectively obtained by the Y signal strength vector analysis layers in the feature conversion subunit, including: Loading the Y signal cluster distribution vectors and the hierarchical signal spectrum into the feature conversion subunit; If the signal strength vector parsing layer Kb is the first signal strength vector parsing layer among the Y signal strength vector parsing layers, the hierarchical signal spectrum is parsed using the signal strength vector parsing layer Kb to generate a signal strength vector obtained by the signal strength vector parsing layer Kb; If the signal strength vector resolution layer Kb is not the first signal strength vector resolution layer among the Y signal strength vector resolution layers, then the signal cluster distribution vector corresponding to the signal strength vector resolution layer Kb-1 is obtained from the Y signal cluster distribution vectors, the signal cluster distribution vector corresponding to the signal strength vector resolution layer Kb-1 and the signal strength vector obtained by the signal strength vector resolution layer Kb-1 are aggregated to generate an aggregated signal feature vector corresponding to the signal strength vector resolution layer Kb-1, and the aggregated signal feature vector corresponding to the signal strength vector resolution layer Kb-1 is analyzed using the signal strength vector resolution layer Kb to generate the signal strength vector obtained by the signal strength vector resolution layer Kb; wherein the signal strength vector resolution layer Kb-1 is the previous signal strength vector resolution layer of the signal strength vector resolution layer Kb; the signal cluster distribution vector corresponding to the signal strength vector resolution layer Kb-1 and the signal strength vector obtained by the signal strength vector resolution layer Kb-1 have a consistent vector length.

6. The data processing method based on open-pit mine 5G base station migration according to claim 1 is characterized in that: The performing regional signalization processing on the target base station signal migration data based on the migration range matrix to generate a regional signal map corresponding to the target base station signal migration data includes: Loading the migration range matrix and the target base station signal migration data into a signal graph processing unit in the target machine learning network; the signal graph processing unit includes a signal representation subunit, a migration range embedding subunit and a signal restoration subunit; Using the signal representation subunit to extract features from the target base station signal migration data to generate a base station signal migration vector; Optimizing the filter factor in the migration range embedding subunit using the migration range matrix to generate an optimized filter factor, and generating a coverage area migration vector according to the optimized filter factor and the base station signal migration vector; The signal restoration subunit is used to perform signal restoration processing on the coverage area migration vector to generate a regional signal map corresponding to the target base station signal migration data.

7. The data processing method based on open-pit mine 5G base station migration according to claim 6 is characterized in that: The migration range embedding subunit includes G embedding filters, where G is a positive integer; optimizing the filter factors in the migration range embedding subunit using the migration range matrix to generate optimized filter factors, and generating a coverage area migration vector based on the optimized filter factors and the base station signal migration vector, including: Dividing the migration range matrix into migration range units corresponding to the G embedded filters respectively; wherein the G embedded filters include an embedded filter Pc, and c is a positive integer not greater than G; Optimizing the filter factors in the embedded filter Pc by using the migration range unit corresponding to the embedded filter Pc to generate optimized filter factors corresponding to the embedded filter Pc; If the embedded filter Pc is the first embedded filter among the G embedded filters, then in the embedded filter Pc, according to the optimized filter factor corresponding to the embedded filter Pc, the base station signal migration vector is parsed to generate an embedded filter vector obtained by the embedded filter Pc; If the embedded filter Pc is not the first embedded filter among the G embedded filters, then in the embedded filter Pc, according to the optimized filter factor corresponding to the embedded filter Pc, the embedded filter vector obtained by the embedded filter Pc-1 is parsed to generate the embedded filter vector obtained by the embedded filter Pc; wherein the embedded filter Pc-1 is the previous embedded filter of the embedded filter Pc; The embedded filter vector obtained by the last embedded filter among the G embedded filters is used as the coverage area migration vector.

8. The data processing method based on open-pit mine 5G base station migration according to claim 1 is characterized in that: The method further comprises: After detecting a signal restoration instruction for the optimized coverage map, loading the optimized coverage map into a target machine learning network; wherein the target machine learning network includes a migration range signal restoration unit; The migration range signal restoration unit is used to perform migration range signal restoration on the optimized coverage map, and the signal restoration map is generated after a signal restoration map corresponding to the optimized coverage map is generated.

9. The data processing method based on open-pit mine 5G base station migration according to claim 1 is characterized in that: The method further comprises: In a basic machine learning network, a template migration range matrix corresponding to a template base station migration area range is obtained; wherein the template migration range matrix is ​​generated by a template target signal coverage map corresponding to the template base station migration area range; the template target signal coverage map includes a plurality of template coverage areas indicated by the template base station migration area range, and the signal strength value in each template coverage area is consistent; Performing regional signalization processing on the template signal map based on the template migration range matrix to generate a template regional signal map corresponding to the template signal map; wherein the template regional signal map is a map having signal distribution characteristics; Optimizing signal coverage of the template area signal map to generate a template optimized coverage map corresponding to the template signal map; wherein the template optimized coverage map is used to represent a map having a signal continuous enhancement effect indicated by the template base station migration area range; Determining a network error parameter of the basic machine learning network according to the template target signal coverage map, the template signal map and the template optimized coverage map; The network parameters in the basic machine learning network are optimized based on the network error parameters. When the optimized basic machine learning network meets the training termination requirements, the optimized basic machine learning network is output as a target machine learning network; wherein the target machine learning network is used to generate an optimized coverage map corresponding to the target base station signal migration data.

10. A data processing system based on 5G base station migration in open-pit mines, characterized in that: The data processing system based on open-pit mine 5G base station migration includes a processor and a machine-readable storage medium, wherein the machine-readable storage medium stores machine executable instructions, and the machine executable instructions are loaded and executed by the processor to implement the data processing method based on open-pit mine 5G base station migration according to any one of claims 1 to 9.

Images