CN118803931A - Method, device, electronic equipment, product and storage medium for distinguishing remote sectors - Google Patents

Abstract

The application relates to the technical field of communication, and provides a method and a device for distinguishing a remote sector, electronic equipment, a product and a storage medium. The method comprises the following steps: according to the application scene of the base station sector, selecting a plurality of initial parameters of the base station sector as characteristic parameters, wherein the initial parameters are determined based on the relation characteristics, the optical fiber length and the base station coverage distance characteristics; and inputting the characteristic parameters into a remote sector discrimination model, and obtaining a remote sector discrimination result of the base station sector output by the remote sector discrimination model. According to the application, the remote base station is judged through the Bayesian network, the dependency relationship among the characteristic parameters is further considered, and the accuracy of judging the remote sector is improved. And a field Jing Goujian Bayesian network is adopted, so that the accuracy of distinguishing the remote sector is improved. The application comprehensively considers the relation characteristic of the base station sector and the peripheral sectors, the optical fiber length and the base station coverage distance characteristic of the base station sector, further determines the initial parameters, and is beneficial to improving the accuracy of distinguishing the remote sector.

Description

Method, device, electronic equipment, product and storage medium for distinguishing remote sectors

Technical Field

The present application relates to the field of communication technology, and in particular to a method, device, electronic equipment, product and storage medium for determining a remote sector.

Background Art

A remote sector is a sector that achieves signal coverage by pulling the radio frequency equipment of a base station to a distant end through optical fiber. Since this deployment method does not require the addition of equipment such as baseband units (BBU) at the radio frequency equipment installation point, it can effectively reduce network construction costs. This deployment method also allows operators to provide wireless communication services in locations where traditional base stations cannot be installed (such as old rooftops and hillsides where computer rooms cannot be built), making network coverage more flexible. The urgent need to save investment in 4G and 5G and the separate deployment method of the central unit (CU) and distributed unit (DU) of 5G have led to the existence of a large number of remote sectors in the current network.

Currently, whether a base station sector in the network is a remote sector is identified based on the manually recorded fields in the engineering parameter table. Record distortion may occur due to problems such as recording errors and failure to update network changes in a timely manner. Inaccurate recording of remote sectors can lead to the following problems: (1) Degraded network performance: If an ordinary base station is mistakenly identified as a remote sector, it may lead to inaccurate network planning and optimization, thereby affecting network performance. (2) Difficult maintenance: Incorrect identification may cause maintenance personnel to be unable to accurately judge the actual situation and location of the base station, increasing the difficulty of troubleshooting and repairing. (3) Impact on investment benefits: When planning a new base station, if a remote sector is mistakenly identified as an ordinary base station, the benefits of the base station will be misjudged, which may lead to unnecessary investment. Conversely, it may lead to insufficient network coverage and affect user perception.

In summary, the existing method for determining the distant base station is inaccurate.

Summary of the invention

The embodiments of the present application provide a method, device, electronic device, product and storage medium for determining a remote sector, so as to solve the technical problem of inaccurately determining a remote base station.

In a first aspect, the present application provides a method for distinguishing a distant sector, comprising: selecting multiple initial parameters of a base station sector as feature parameters according to an application scenario of the base station sector, the initial parameters being determined based on the relationship characteristics between the base station sector and the surrounding sectors of the base station sector, the optical fiber length from the baseband unit of the base station sector to the signal processing unit or the signal transmission unit, and the base station coverage distance characteristics of the base station sector; inputting the feature parameters into a distant sector distinction model to obtain a distant sector distinction result of the base station sector output by the distant sector distinction model; wherein the distant sector distinction model is obtained based on a preset Bayesian network under multiple application scenarios, based on multiple scene sample initial parameters of each application scenario and label training of the distant sector distinction result, the scene sample initial parameters are determined based on the sample initial parameters, and the preset Bayesian network under each application scenario is constructed based on multiple scene sample initial parameters and the dependency relationship between the multiple scene sample initial parameters.

In one embodiment, the base station coverage distance characteristics include the minimum coverage distance of the MDT maximum coverage distance, the MDT minimum coverage distance, the propagation model coverage distance, the over-the-top service OTT maximum coverage distance, the OTT minimum coverage distance, the timing advance TA maximum coverage distance, and the TA minimum coverage distance; the relationship characteristics include the switching coefficient between non-co-station sectors and co-station sectors of the base station sector, the sampling point quantity coefficient between non-co-station sectors and co-station sectors of the base station sector, and the common user coefficient between non-co-station sectors and co-station sectors of the base station sector.

In one embodiment, multiple initial parameters are determined, including: taking the optical fiber length from the baseband unit to the signal processing unit or the signal transmission unit, the first ratio of the MDT maximum coverage distance and the propagation model coverage distance, the second ratio of the OTT maximum coverage distance and the propagation model coverage distance, the MDT minimum coverage distance, the OTT minimum coverage distance, the third ratio of the MDT maximum coverage distance and the TA maximum coverage distance, the fourth ratio of the MDT minimum coverage distance and the TA minimum coverage distance, the fifth ratio of the OTT maximum coverage distance and the TA maximum coverage distance, the sixth ratio of the OTT minimum coverage distance and the TA minimum coverage distance, the switching coefficient, the sampling point quantity coefficient and the common user coefficient as multiple initial parameters.

In one embodiment, the remote sector discrimination model is determined based on the following steps: based on the label of the remote sector discrimination result of the sample base station sector and the scene sample initial parameters, the scene sample data of each application scenario carrying the label is obtained, and the preset Bayesian network of the application scenario is trained according to the scene sample data; when the F1 score of the trained preset Bayesian network is greater than or equal to the set threshold, the training is terminated to obtain the remote sector discrimination model; when the F1 score of the trained preset Bayesian network is less than the set threshold, at least one remaining sample initial parameter is added to the scene sample initial parameter to update the preset Bayesian network of the application scenario and the scene sample data of the application scenario, and the preset Bayesian network of the updated application scenario is trained based on the updated scene sample data of the application scenario until the F1 score of the preset Bayesian network of the updated application scenario is greater than or equal to the set threshold to obtain the remote sector discrimination model, and the remaining sample initial parameters are other sample initial parameters except the scene sample initial parameters.

In one embodiment, the preset Bayesian network for the application scenario is constructed based on the following steps: using the initial parameters of each scenario sample as the network nodes of the preset Bayesian network; determining the connection relationship of the network nodes based on the dependency relationship of the initial parameters of the scenario samples; determining the conditional probability distribution value of each network node based on the initial parameters of the scenario samples and the dependency relationship; and obtaining the preset Bayesian network for the application scenario based on the network nodes, the connection relationship of the network nodes, and the conditional probability distribution values of the network nodes.

In one embodiment, determining the scene sample initial parameters includes: obtaining key influencing parameters in the application scenario according to expert experience; and obtaining the scene sample initial parameters according to the intersection of the key influencing parameters and the sample initial parameters.

In the second aspect, an embodiment of the present application provides a remote sector identification device, including: a characteristic parameter determination module, which is used to select multiple initial parameters of a base station sector as characteristic parameters according to the application scenario of the base station sector, and the initial parameters are determined based on the relationship characteristics between the base station sector and the surrounding sectors of the base station sector, the optical fiber length from the baseband unit of the base station sector to the signal processing unit or the signal transmission unit, and the base station coverage distance characteristics of the base station sector; a remote sector identification module, which is used to input the characteristic parameters into a remote sector identification model to obtain a remote sector identification result of the base station sector output by the remote sector identification model; wherein the remote sector identification model is obtained on the basis of a preset Bayesian network under multiple application scenarios, based on multiple scene sample initial parameters of each application scenario and label training of the remote sector identification result, the scene sample initial parameters are determined based on the sample initial parameters, and the preset Bayesian network under each application scenario is constructed based on multiple scene sample initial parameters and the dependency relationship between multiple scene sample initial parameters.

In a third aspect, an embodiment of the present application provides an electronic device, including a processor and a memory storing a computer program, and when the processor executes the program, the method for determining the remote sector of the first aspect is implemented.

In a fourth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the method for determining the remote sector of the first aspect.

In a fifth aspect, an embodiment of the present application provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for determining a distant sector according to the first aspect.

The method, device, electronic device, product and storage medium for distinguishing a remote sector provided by the embodiment of the present application select multiple initial parameters of the base station sector as characteristic parameters according to the application scenario of the base station sector, and the initial parameters are determined based on the relationship characteristics between the base station sector and the surrounding sectors of the base station sector, the optical fiber length from the baseband unit of the base station sector to the signal processing unit or the signal transmission unit, and the base station coverage distance characteristics of the base station sector; the characteristic parameters are input into the remote sector distinguishing model to obtain the remote sector distinguishing result of the base station sector output by the remote sector distinguishing model; wherein the remote sector distinguishing model is obtained based on the preset Bayesian network in multiple application scenarios, based on the initial parameters of multiple scene samples of each application scenario and the label training of the remote sector distinguishing result, the scene sample initial parameters are determined based on the sample initial parameters, and the preset Bayesian network in each application scenario is constructed based on the dependency relationship between the initial parameters of multiple scene samples and the initial parameters of multiple scene samples. The present application distinguishes the remote base station through the Bayesian network, further considers the dependency relationship between the characteristic parameters, and improves the accuracy of distinguishing the remote sector. The use of scenario-based Bayesian networks is conducive to improving the accuracy of identifying distant sectors. This application comprehensively considers the relationship characteristics between base station sectors and surrounding sectors, optical fiber length, and base station coverage distance characteristics of base station sectors, and then determines the initial parameters, which is conducive to improving the accuracy of identifying distant sectors.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the technical solutions in the present application or the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

FIG1 is a flow chart of a method for determining a remote sector according to an embodiment of the present application;

FIG2 is a schematic diagram of a training process of a remote sector discrimination model provided in an embodiment of the present application;

FIG3 is a schematic diagram of the structure of a preset Bayesian network provided in an embodiment of the present application;

FIG4 is a schematic diagram of the structure of a device for distinguishing a remote sector provided in an embodiment of the present application;

FIG. 5 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

DETAILED DESCRIPTION

In order to make the purpose, technical solutions and advantages of this application clearer, the technical solutions in this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

FIG1 is a flow chart of a method for determining a remote sector provided by an embodiment of the present application. Referring to FIG1 , an embodiment of the present application provides a method for determining a remote sector, including steps S100 to S200, each of which is as follows:

S100: According to an application scenario of the base station sector, multiple initial parameters of the base station sector are selected as characteristic parameters.

The initial parameters are determined based on the relationship characteristics between the base station sector and its surrounding sectors, the optical fiber length from the baseband unit of the base station sector to the signal processing unit or signal transmission unit, and the base station coverage distance characteristics of the base station sector.

Obtain the base station data of the base station, which includes the base station engineering parameter table, the Minimization of Drive Test (MDT) data table (including the base station identifier to which the sampling point belongs, the latitude and longitude information of the sampling point, the reference signal receiving power, etc.), the OTT (Over The Top, over-the-top service) data table, the Operation and Maintenance Center (OMC) measurement table and the user information table. The OMC measurement table includes the timing advance (Timing Advance) distribution of the base station sector, the neighboring cell and other information. Clean and pre-process the MDT data and OTT data in the base station data to remove missing values or outliers.

Based on the above embodiments, the base station coverage distance characteristics include the maximum coverage distance of minimized drive test MDT, the minimum coverage distance of MDT, the propagation model coverage distance, the maximum coverage distance of over-the-top service OTT, the minimum coverage distance of OTT, the maximum coverage distance of timing advance TA, and the minimum coverage distance of TA; the relationship characteristics include the switching coefficient between non-co-station sectors and co-station sectors of the base station sector, the sampling point quantity coefficient between non-co-station sectors and co-station sectors of the base station sector, and the common user coefficient between non-co-station sectors and co-station sectors of the base station sector.

The optical fiber length from the baseband unit (BBU) of the base station sector to the signal processing unit or signal transmission unit is obtained according to the base station data. The signal processing unit or signal transmission unit includes an active antenna unit (AAU) and a remote radio unit (RRU). The BBU is connected to the AAU/RRU through an optical fiber. The distance between the BBU and the AAU/RRU is far, and the optical fiber length is also large. However, if the optical fiber between the BBU and the AAU/RRU is large, the base station is not necessarily a remote base station. It may also be caused by reasons such as optical fiber redundancy or optical fiber transmission through a transmission loop.

The base station coverage distance characteristics of the base station sector are obtained based on the base station data. Since the AAU and BBU of the remote base station are not in the same location, there is a large gap between the MDT and OTT coverage distances, propagation model distances, and TA distances of the remote base station. The MDT data contains information such as level, GPS longitude and latitude. Since GPS only has signals outdoors, the MDT data mainly reflects outdoor coverage and location information.

Calculate the maximum coverage distance of MDT. Calculate the distance from each measurement point in the MDT data table to the center point of the base station, and take the average of the largest top k distances (TOPK) as the maximum coverage distance of MDT. k is manually selected, for example, the number of sampling points corresponding to 1% of the total sampling points can be selected.

Calculate the propagation model coverage distance. The coverage distance estimated using the 3GPP propagation model is called the model coverage distance. According to the propagation model calculation, the distance corresponding to the minimum access level is the propagation model coverage distance. Manually select the minimum access level, which is generally a unified value for the entire network and can be queried in the engineering parameters. It is generally set to -110dbm. Calculate the maximum access path loss. The maximum access path loss is the path loss when the level is the minimum access level. The calculation formula for the maximum access path loss is:

PL _max = T _v −R _v −RC;

Where T _v is the base station transmit power, R _v is the minimum access level, RC is the building penetration loss, and PL _max is the maximum access path loss.

According to the maximum access path loss, the propagation model coverage distance is calculated by the 3gpp channel model. In the present invention, the distance from the mobile phone to the base station corresponding to the maximum access path loss is the model coverage distance. The calculation formula of the model coverage distance is:

Where Des is the model coverage distance, PL _max is the maximum access path loss, and f _c is the sector frequency of the base station sector.

Calculate the minimum MDT coverage distance. The larger the minimum MDT coverage distance, the more likely the base station sector is a remote sector. The minimum MDT coverage distance for non-remote sectors is generally between 0 and 10 meters. Calculate the distance from each measurement point in the MDT data table to the center point of the base station, and take the average of the first k smallest distances as the minimum MDT coverage distance.

Calculate the maximum coverage distance of OTT. Calculate the distance from the building to which each measurement point in the OTT data table belongs to the center point of the base station, and take the average of the largest first k distances as the maximum coverage distance of OTT. Specifically, extract the longitude and latitude information of each building in MDT, calculate the geodesic distance between the longitude and latitude of the base station in the base station engineering parameters (the longitude and latitude are the distance of the base station BBU), and the geodesic distance is the distance from the building to the base station. Select the average of the largest first k geodesic distances as the maximum coverage distance of OTT.

Calculate the minimum OTT coverage distance. The larger the minimum OTT coverage distance, the more likely the base station sector is a remote sector. The minimum OTT coverage distance for non-remote sectors is generally between 0 and 10 meters. Calculate the distance from each building in the OTT data table to the center point of the base station, and take the average of the first k smallest distances as the minimum OTT coverage distance.

Calculate the maximum coverage distance of TA. Extract the distribution table of TA through the OMC measurement table. Obtain the sector coverage plan (SCS) configuration data of each base station sector by querying the OMC engineering parameter table. According to the mapping relationship table of SCS configuration data (for example, subcarrier spacing) and TA distance, match the TA distance of the base station sector. The mapping relationship table of subcarrier spacing and TA distance is shown in Table 1.

Table 1 Mapping relationship between subcarrier spacing and TA

Subcarrier spacing (Hz) TA distance (m) 15 78.125 30 39.0625 60 19.53125 120 9.765625

Get the first k sampling points with the largest TA distance and at least one TA maximum distance corresponding to the k sampling points, perform weighted summation on the TA distance distribution values, and obtain the TA maximum coverage distance. The calculation formula for the TA maximum distance is:

Wherein, k is the number of sampling points with the largest TA distance selected, _n1 is the number of sampling points corresponding to the TA maximum distance 1, _n2 is the number of sampling points corresponding to the TA maximum distance 2, and _n1 + _n2 =k.

Calculate the TA minimum coverage distance. Obtain the first k sampling points with the smallest TA distance and at least one TA minimum distance corresponding to the k sampling points, perform weighted summation on the TA minimum distances, and obtain the TA minimum coverage distance. The calculation formula for the TA minimum distance is:

Wherein, k is the number of sampling points with the minimum TA distance selected, n ₃ is the number of sampling points corresponding to the minimum TA distance 3, n ₄ is the number of sampling points corresponding to the minimum TA distance 4, and n ₃ +n ₄ =k.

Determine the relationship characteristics between the base station sector and the surrounding sectors of the base station sector. The surrounding sectors of the base station sector are the surrounding sectors where the AAU/RRU is located, which are different from the sectors around the BBU. The AAU/RRU of the ordinary base station sector is in the same location as the BBU, and the AAU/RRU is the same as the surrounding sectors of the BBU. There are fewer common users between the remote sector and the sectors around the BBU, and there are more common users between the ordinary base station sector and the sectors around the BBU.

Because the AAU/RRU of the remote sector is far from the base station, there are fewer handovers with other sectors in the same station, while there are more handovers with other sectors in the same station for non-remote sectors. The calculation formula of the handover coefficient is:

Where r is the switching coefficient, _h1 is the maximum number of switching between the base station sector and the same-station sector, and _h2 is the number of switching between the base station sector and the non-same-station sector with the largest number of switching.

Calculate the sampling point quantity coefficient of the non-co-site sector and the co-site sector of the base station sector in the non-primary coverage sector of the measurement report data of the base station sector. The calculation formula of the sampling point quantity coefficient is:

Among them, p is the sampling point quantity coefficient, _m1 is the sampling point quantity of other base station sectors (including non-co-station sectors and co-station sectors) with the most sampling points in the measurement report data except the sector of the current base station, and _m2 is the sampling point quantity of the co-station sector with the most sampling points in the measurement report data.

Calculate the common user coefficient between non-co-site sectors and co-site sectors of the base station sector. Because the location of the AAU/RRU of the remote sector is relatively far away from the station, there are fewer common users with other sectors in the same station, and more common users with other sectors in the same station for non-remote sectors. Extract all other base station sectors within M meters from the base station sector to form a set Ω, where M is a manually specified distance as large as possible, such as 10km. From the user information table, extract the user's unique ID (generally the user's international mobile user identity code, which can also be a unique identifier for other users for information security considerations) in each sector of the set Ω within time T, where T is a manually specified time period, such as 1 day, 1 week, etc. The calculation formula for the common user coefficient between non-co-site sectors and co-site sectors of the base station sector is:

Among them, _U2 is the maximum number of common users between the base station sector and all non-co-station sectors in the set Ω, _U1 is the maximum number of common users between the base station sector and the co-station sector, common users refer to users with the same user unique ID, and q is the common user coefficient.

Based on the above embodiments, multiple initial parameters are determined. Specifically, the optical fiber length from the baseband unit to the signal processing unit or the signal transmission unit, the first ratio of the MDT maximum coverage distance and the propagation model coverage distance, the second ratio of the OTT maximum coverage distance and the propagation model coverage distance, the MDT minimum coverage distance, the OTT minimum coverage distance, the third ratio of the MDT maximum coverage distance and the TA maximum coverage distance, the fourth ratio of the MDT minimum coverage distance and the TA minimum coverage distance, the fifth ratio of the OTT maximum coverage distance and the TA maximum coverage distance, the sixth ratio of the OTT minimum coverage distance and the TA minimum coverage distance, the switching coefficient, the sampling point quantity coefficient and the common user coefficient are used as multiple initial parameters.

The initial parameters include df (the optical fiber length from the baseband unit to the signal processing unit or the signal transmission unit), λ _mdt (the first ratio), λ _OTT (the second ratio), D _minMDT (the minimum coverage distance of MDT), D _minOTT (the minimum coverage distance of OTT), λ _dtmax (the third ratio), λ _dtmin (the fourth ratio), λ _OTmax (the fifth ratio), λ _OTmin (the sixth ratio), r (the switching coefficient), p (the sampling point quantity coefficient) and q (the common user coefficient). According to the above initial parameters, the characteristic matrix corresponding to each base station sector is constructed.

According to the application scenario of the base station sector, corresponding initial parameters are selected from the characteristics and matrix of the base station sector as characteristic parameters of the base station sector.

Compared with the traditional method of using industrial parameter recording data, this application constructs more comprehensive initial parameters. This method uses 5 types of data sources (including base station industrial parameter table, MDT data table, OTT data table, OMC measurement table and user information table) to generate 12 types of initial parameters, which can more comprehensively characterize the information of base station sectors and lay the foundation for the next step of remote sector identification.

S200: Inputting characteristic parameters into a remote sector discrimination model, and obtaining a remote sector discrimination result of the base station sector output by the remote sector discrimination model.

Among them, the distant sector discrimination model is obtained based on the preset Bayesian network in multiple application scenarios, based on multiple scene sample initial parameters of each application scenario and label training of the distant sector discrimination results. The scene sample initial parameters are determined based on the sample initial parameters, and the preset Bayesian network in each application scenario is constructed based on multiple scene sample initial parameters and the dependency relationship between multiple scene sample initial parameters.

The distant sector discrimination model is determined based on the following steps: based on the label of the distant sector discrimination result of the sample base station sector and the scene sample initial parameters, the scene sample data of each application scenario carrying the label is obtained, and the preset Bayesian network of the application scenario is trained according to the scene sample data; when the F1 score of the trained preset Bayesian network is greater than or equal to the set threshold, the training is terminated to obtain the distant sector discrimination model; when the F1 score of the trained preset Bayesian network is less than the set threshold, at least one remaining sample initial parameter is added to the scene sample initial parameter to update the preset Bayesian network of the application scenario and the scene sample data of the application scenario, and the preset Bayesian network of the updated application scenario is trained based on the updated scene sample data of the application scenario until the F1 score of the preset Bayesian network of the updated application scenario is greater than or equal to the set threshold to obtain the distant sector discrimination model, and the remaining sample initial parameters are other sample initial parameters except the scene sample initial parameters.

Base station sectors with accurate data in the network are selected as sample base station sectors. There are a number of key base stations that are maintained in the network, and the working parameter information of these base stations is very accurate. Key base station sectors are selected and the labels (lables) of the remote sector discrimination results are marked, where the remote sector is marked as 1 and the non-remote sector is marked as 0, forming a sample base station sector set. According to the sample initial parameters of the sample base station sectors (including the above-mentioned df, λ _mdt , λ _OTT , D _minMDT , D _minOTT , λ _dtmax , λ _dtmin , λ _OTmax , λ _OTmin , r, p and q), labels and application scenarios, the feature matrix of all sample base station sectors is obtained.

Determine the initial parameters of the scene sample. Specifically, obtain the key influencing parameters in the application scenario based on expert experience; and obtain the initial parameters of the scene sample based on the intersection of the key influencing parameters and the sample initial parameters.

For different application scenarios, it is necessary to use the corresponding scene sample initial parameters to build the corresponding preset Bayesian network.

The preset Bayesian network for the application scenario is constructed based on the following steps: using the initial parameters of each scenario sample as the network nodes of the preset Bayesian network; determining the connection relationship of the network nodes based on the dependency relationship of the initial parameters of the scenario samples; determining the conditional probability distribution value of each network node based on the initial parameters of the scenario samples and the dependency relationship; obtaining the preset Bayesian network for the application scenario based on the network nodes, the connection relationship of the network nodes and the conditional probability distribution values of the network nodes.

For example, according to expert experience, in mountainous scenarios, since the coverage of high mountain sites is far away, the discrimination of coverage distance features is poor. The key influencing parameters for determining whether a base station sector is a remote sector are the fiber distance and the common sectors and common users of the surrounding sectors. The intersection of the key influencing parameters and the sample initial parameters is df, r, p and label. df, r, p and label are the initial parameters of the scene samples. df, r, p and label constitute four network nodes. The connection relationship between network nodes is constructed according to the dependency relationship between the initial parameters of the scene samples. For example, according to expert experience or manual specification, the dependency relationship between df, r, p and label is that both r and p depend on df, and label depends on p. Through expert experience, the conditional probability distribution value (CPD) of each network node is determined. As shown in Figure 3, the preset Bayesian network corresponding to the application scenario is finally obtained according to the connection relationship of the network nodes in the application scenario and the conditional probability distribution value (CPD) of each network node.

As shown in Figure 2, obtain the sample base station sector in the mountain scene, the scene sample initial parameters (including df, r and p) and the label (lable) of the distant sector discrimination result. According to the label, df, r and p are marked to obtain the scene sample data in the application scenario (mountain scene). According to the preset ratio, the scene sample data is divided into a training set and a verification set. The preset Bayesian network in the mountain scene is trained according to the training set. The preset Bayesian network in the mountain scene is verified according to the verification set. During the verification process, the F1 score of the preset Bayesian network is calculated. If the F1 score is greater than or equal to the set threshold, it is determined that the preset Bayesian network training is completed, and the corresponding distant sector discrimination model in the application scenario is obtained. If the F1 score is less than the set threshold, a sample initial parameter is selected from the remaining sample initial parameters and added to the scene sample initial parameters. For example, the scene sample initial parameters in the mountain scene are df, r and p, and the remaining sample initial parameters are λ _mdt , λ _OTT , D _minMDT , D _minOTT , λ _dtmax , λ _dtmin , λ _OTmax , λ _OTmin and q. According to expert experience, one remaining sample initial parameter is selected and added to the scene sample initial parameter to update the scene sample initial parameter. The preset Bayesian network and scene sample data in the application scenario are updated according to the updated scene sample initial parameters. According to the above method, the updated preset Bayesian network is trained and verified (incremental learning training Bayesian network) until the F1 score of the preset Bayesian network is greater than the set threshold. The calculation formula of the F1 score is:

Among them, P is the precision, R is the recall rate, F1 is the F1 score, TP is the prediction result and label output by the preset Bayesian network, both of which are the number of distant sectors, FP is the label is not the number of distant sectors and the prediction result is the number of distant sectors, and FN is the prediction result is not the number of distant sectors and label is the number of distant sectors.

This application adopts the method of constructing Bayesian networks by scenario, divides the application scenarios and establishes a targeted model for distinguishing distant sectors. According to the F1 score of the Bayesian network, the network nodes are dynamically added to improve the accuracy of distinguishing distant sectors and reduce the amount of calculation.

Update the preset Bayesian network for the application scenario according to the updated initial parameters of the scene samples. Add network nodes of the preset Bayesian network according to the updated initial parameters of the scene samples. According to the updated scene sample data, manually select a suitable algorithm (e.g., K2, hill climbing algorithm, etc.), train the updated preset Bayesian network, and update the dependency (connection relationship) of all network nodes. Calculate the edge probability distribution of all network nodes using the kernel density estimation method according to the updated scene sample data, and then calculate the CPD of the updated network nodes using the Bayesian formula according to the dependency.

The method for distinguishing a distant sector provided in the embodiment of the present application selects multiple initial parameters of the base station sector as characteristic parameters according to the application scenario of the base station sector, and the initial parameters are determined based on the relationship characteristics between the base station sector and the surrounding sectors of the base station sector, the optical fiber length from the baseband unit of the base station sector to the signal processing unit or the signal transmission unit, and the base station coverage distance characteristics of the base station sector; the characteristic parameters are input into the distant sector distinguishing model to obtain the distant sector distinguishing result of the base station sector output by the distant sector distinguishing model; wherein, the distant sector distinguishing model is obtained on the basis of the preset Bayesian network under multiple application scenarios, based on the initial parameters of multiple scene samples of each application scenario and the label training of the distant sector distinguishing result, the scene sample initial parameters are determined based on the sample initial parameters, and the preset Bayesian network under each application scenario is constructed based on the dependency relationship between the initial parameters of multiple scene samples and the initial parameters of multiple scene samples. The present application distinguishes the distant base station through the Bayesian network, further considers the dependency relationship between the characteristic parameters, and improves the accuracy of distinguishing the distant sector. The use of scene-by-scene construction of the Bayesian network is conducive to improving the accuracy of distinguishing the distant sector. The present application comprehensively considers the relationship characteristics between the base station sector and the surrounding sectors, the optical fiber length and the base station coverage distance characteristics of the base station sector, and then determines the initial parameters, which is conducive to improving the accuracy of distinguishing the distant sectors.

The following is a description of the remote sector identification device provided in the embodiment of the present application. The remote sector identification device described below and the remote sector identification method described above can be referred to each other. Referring to FIG. 4, FIG. 4 is a schematic diagram of the structure of the remote sector identification device provided in the embodiment of the present application. A remote sector identification device includes:

The characteristic parameter determination module 401 is used to select multiple initial parameters of the base station sector as characteristic parameters according to the application scenario of the base station sector. The initial parameters are determined based on the relationship characteristics between the base station sector and the surrounding sectors of the base station sector, the optical fiber length from the baseband unit of the base station sector to the signal processing unit or the signal transmission unit, and the base station coverage distance characteristics of the base station sector.

The distant sector discrimination module 402 is used to input feature parameters into the distant sector discrimination model to obtain the distant sector discrimination result of the base station sector output by the distant sector discrimination model; wherein the distant sector discrimination model is obtained based on the preset Bayesian network under multiple application scenarios, based on multiple scene sample initial parameters of each application scenario and the label training of the distant sector discrimination result, the scene sample initial parameters are determined based on the sample initial parameters, and the preset Bayesian network under each application scenario is constructed based on multiple scene sample initial parameters and the dependency relationship between the multiple scene sample initial parameters.

The remote sector discrimination device provided in the embodiment of the present application selects multiple initial parameters of the base station sector as characteristic parameters according to the application scenario of the base station sector, and the initial parameters are determined based on the relationship characteristics between the base station sector and the surrounding sectors of the base station sector, the optical fiber length from the baseband unit of the base station sector to the signal processing unit or the signal transmission unit, and the base station coverage distance characteristics of the base station sector; the characteristic parameters are input into the remote sector discrimination model to obtain the remote sector discrimination result of the base station sector output by the remote sector discrimination model; wherein, the remote sector discrimination model is obtained on the basis of the preset Bayesian network under multiple application scenarios, based on the initial parameters of multiple scene samples of each application scenario and the label training of the remote sector discrimination result, the scene sample initial parameters are determined based on the sample initial parameters, and the preset Bayesian network under each application scenario is constructed based on the dependency relationship between the initial parameters of multiple scene samples and the initial parameters of multiple scene samples. The present application discriminates the remote base station through the Bayesian network, further considers the dependency relationship between the characteristic parameters, and improves the accuracy of discriminating the remote sector. The use of scene-by-scene construction of the Bayesian network is conducive to improving the accuracy of discriminating the remote sector. The present application comprehensively considers the relationship characteristics between the base station sector and the surrounding sectors, the optical fiber length and the base station coverage distance characteristics of the base station sector, and then determines the initial parameters, which is conducive to improving the accuracy of distinguishing the distant sectors.

In one embodiment, the base station coverage distance characteristics include the minimum coverage distance of the MDT maximum coverage distance, the MDT minimum coverage distance, the propagation model coverage distance, the over-the-top service OTT maximum coverage distance, the OTT minimum coverage distance, the timing advance TA maximum coverage distance, and the TA minimum coverage distance; the relationship characteristics include the switching coefficient between non-co-station sectors and co-station sectors of the base station sector, the sampling point quantity coefficient between non-co-station sectors and co-station sectors of the base station sector, and the common user coefficient between non-co-station sectors and co-station sectors of the base station sector.

In one embodiment, the characteristic parameter determination module 401 is used to: use the optical fiber length from the baseband unit to the signal processing unit or the signal transmission unit, the first ratio of the MDT maximum coverage distance and the propagation model coverage distance, the second ratio of the OTT maximum coverage distance and the propagation model coverage distance, the MDT minimum coverage distance, the OTT minimum coverage distance, the third ratio of the MDT maximum coverage distance and the TA maximum coverage distance, the fourth ratio of the MDT minimum coverage distance and the TA minimum coverage distance, the fifth ratio of the OTT maximum coverage distance and the TA maximum coverage distance, the sixth ratio of the OTT minimum coverage distance and the TA minimum coverage distance, the switching coefficient, the sampling point quantity coefficient and the common user coefficient as multiple initial parameters.

In one embodiment, the remote sector identification module 402 is used to: obtain scene sample data of each application scenario carrying the label based on the label of the remote sector identification result of the sample base station sector and the scene sample initial parameters, and train the preset Bayesian network of the application scenario according to the scene sample data; when the F1 score of the trained preset Bayesian network is greater than or equal to the set threshold, end the training to obtain the remote sector identification model; when the F1 score of the trained preset Bayesian network is less than the set threshold, add at least one remaining sample initial parameter to the scene sample initial parameter to update the preset Bayesian network of the application scenario and the scene sample data of the application scenario, and train the updated preset Bayesian network of the application scenario based on the updated scene sample data of the application scenario until the F1 score of the updated preset Bayesian network of the application scenario is greater than or equal to the set threshold to obtain the remote sector identification model, and the remaining sample initial parameters are other sample initial parameters except the scene sample initial parameters.

In one embodiment, the distant sector identification module 402 is used to: use the initial parameters of each scene sample as a network node of a preset Bayesian network; determine the connection relationship of the network nodes based on the dependency relationship of the scene sample initial parameters; determine the conditional probability distribution value of each network node based on the scene sample initial parameters and the dependency relationship; obtain the preset Bayesian network of the application scenario based on the network nodes, the connection relationship of the network nodes and the conditional probability distribution value of the network nodes.

In one embodiment, the remote sector determination module 402 is used to: obtain key influencing parameters in the application scenario according to expert experience; and obtain the scene sample initial parameters according to the intersection of the key influencing parameters and the sample initial parameters.

FIG5 illustrates a schematic diagram of the physical structure of an electronic device. As shown in FIG5, the electronic device may include: a processor 510, a communication interface 520, a memory 530, and a communication bus 540, wherein the processor 510, the communication interface 520, and the memory 530 communicate with each other through the communication bus 540. The processor 510 may call a computer program in the memory 530 to execute a method for distinguishing a remote sector, for example, including:

According to the application scenario of the base station sector, multiple initial parameters of the base station sector are selected as feature parameters, and the initial parameters are determined based on the relationship characteristics between the base station sector and the surrounding sectors of the base station sector, the optical fiber length from the baseband unit of the base station sector to the signal processing unit or the signal transmission unit, and the base station coverage distance characteristics of the base station sector; the feature parameters are input into a distant sector discrimination model to obtain a distant sector discrimination result of the base station sector output by the distant sector discrimination model; wherein the distant sector discrimination model is obtained on the basis of a preset Bayesian network under multiple application scenarios, based on multiple scene sample initial parameters of each application scenario and label training of the distant sector discrimination result, the scene sample initial parameters are determined based on the sample initial parameters, and the preset Bayesian network under each application scenario is constructed based on multiple scene sample initial parameters and the dependency relationship between the multiple scene sample initial parameters.

In addition, the logic instructions in the above-mentioned memory 530 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art, and the computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

On the other hand, an embodiment of the present application further provides a computer program product, the computer program product including a computer program, the computer program can be stored on a non-transitory computer-readable storage medium, and when the computer program is executed by a processor, the computer can execute the method for determining the remote sector provided in the above embodiments, for example, including:

According to the application scenario of the base station sector, multiple initial parameters of the base station sector are selected as feature parameters, and the initial parameters are determined based on the relationship characteristics between the base station sector and the surrounding sectors of the base station sector, the optical fiber length from the baseband unit of the base station sector to the signal processing unit or the signal transmission unit, and the base station coverage distance characteristics of the base station sector; the feature parameters are input into a distant sector discrimination model to obtain a distant sector discrimination result of the base station sector output by the distant sector discrimination model; wherein the distant sector discrimination model is obtained on the basis of a preset Bayesian network under multiple application scenarios, based on multiple scene sample initial parameters of each application scenario and label training of the distant sector discrimination result, the scene sample initial parameters are determined based on the sample initial parameters, and the preset Bayesian network under each application scenario is constructed based on multiple scene sample initial parameters and the dependency relationship between the multiple scene sample initial parameters.

On the other hand, an embodiment of the present application further provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores a computer program, wherein the computer program is used to enable a processor to execute the method for determining a remote sector provided in the above embodiments, for example, including:

According to the application scenario of the base station sector, multiple initial parameters of the base station sector are selected as feature parameters, and the initial parameters are determined based on the relationship characteristics between the base station sector and the surrounding sectors of the base station sector, the optical fiber length from the baseband unit of the base station sector to the signal processing unit or the signal transmission unit, and the base station coverage distance characteristics of the base station sector; the feature parameters are input into a distant sector discrimination model to obtain a distant sector discrimination result of the base station sector output by the distant sector discrimination model; wherein the distant sector discrimination model is obtained on the basis of a preset Bayesian network under multiple application scenarios, based on multiple scene sample initial parameters of each application scenario and label training of the distant sector discrimination result, the scene sample initial parameters are determined based on the sample initial parameters, and the preset Bayesian network under each application scenario is constructed based on multiple scene sample initial parameters and the dependency relationship between the multiple scene sample initial parameters.

The non-transitory computer-readable storage medium can be any available medium or data storage device that can be accessed by the processor, including but not limited to magnetic storage (such as floppy disks, hard disks, magnetic tapes, magneto-optical disks (MO), etc.), optical storage (such as CD, DVD, BD, HVD, etc.), and semiconductor storage (such as ROM, EPROM, EEPROM, non-volatile memory (NAND FLASH), solid-state drive (SSD)), etc.

The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM/RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims (10)

1. A method for determining a remote sector, comprising: According to the application scenario of the base station sector, multiple initial parameters of the base station sector are selected as characteristic parameters, where the initial parameters are determined based on the relationship characteristics between the base station sector and the surrounding sectors of the base station sector, the optical fiber length from the baseband unit of the base station sector to the signal processing unit or the signal transmission unit, and the base station coverage distance characteristics of the base station sector; Inputting the characteristic parameter into a remote sector discrimination model, and obtaining a remote sector discrimination result of the base station sector output by the remote sector discrimination model; Among them, the distant sector discrimination model is obtained based on the preset Bayesian network in the multiple application scenarios, based on the multiple scene sample initial parameters of each application scenario and the label training of the distant sector discrimination results, the scene sample initial parameters are determined based on the sample initial parameters, and the preset Bayesian network in each application scenario is constructed based on the multiple scene sample initial parameters and the dependency relationship between the multiple scene sample initial parameters. 2. The method for determining the remote sector according to claim 1 is characterized in that the base station coverage distance characteristics include the maximum coverage distance of minimized drive test MDT, the minimum coverage distance of MDT, the coverage distance of the propagation model, the maximum coverage distance of the over-the-top service OTT, the minimum coverage distance of OTT, the maximum coverage distance of the timing advance TA, and the minimum coverage distance of TA; the relationship characteristics include the switching coefficient between the non-co-station sectors and the co-station sectors of the base station sector, the sampling point quantity coefficient between the non-co-station sectors and the co-station sectors of the base station sector, and the common user coefficient between the non-co-station sectors and the co-station sectors of the base station sector. 3. The method for determining a remote sector according to claim 2, wherein determining a plurality of the initial parameters comprises: The optical fiber length from the baseband unit to the signal processing unit or the signal transmission unit, the first ratio of the MDT maximum coverage distance and the propagation model coverage distance, the second ratio of the OTT maximum coverage distance and the propagation model coverage distance, the MDT minimum coverage distance, the OTT minimum coverage distance, the third ratio of the MDT maximum coverage distance and the TA maximum coverage distance, the fourth ratio of the MDT minimum coverage distance and the TA minimum coverage distance, the fifth ratio of the OTT maximum coverage distance and the TA maximum coverage distance, the sixth ratio of the OTT minimum coverage distance and the TA minimum coverage distance, the switching coefficient, the sampling point quantity coefficient and the common user coefficient are used as multiple initial parameters. 4. The method for distinguishing a remote sector according to claim 1, wherein the remote sector distinguishing model is determined based on the following steps: Based on the label of the remote sector discrimination result of the sample base station sector and the initial parameter of the scene sample, scene sample data of each application scene carrying the label is obtained, The preset Bayesian network of the application scenario is trained according to the scenario sample data; when the F1 score of the trained preset Bayesian network is greater than or equal to the set threshold, the training is terminated to obtain the remote sector discrimination model; When the F1 score of the trained preset Bayesian network is less than the set threshold, at least one remaining sample initial parameter is added to the scene sample initial parameter to update the preset Bayesian network of the application scenario and the scene sample data of the application scenario, and the preset Bayesian network of the updated application scenario is trained based on the scene sample data of the updated application scenario until the F1 score of the preset Bayesian network of the updated application scenario is greater than or equal to the set threshold, so as to obtain the distant sector discrimination model, and the remaining sample initial parameters are other sample initial parameters except the scene sample initial parameters. 5. The method for determining a remote sector according to claim 1, wherein the preset Bayesian network of the application scenario is constructed based on the following steps: Using the initial parameters of each of the scene samples as a network node of the preset Bayesian network; Determining a connection relationship between network nodes based on a dependency relationship between initial parameters of the scene samples; Determining a conditional probability distribution value of each network node based on the scene sample initial parameters and the dependency relationship; A preset Bayesian network for the application scenario is obtained based on the network nodes, the connection relationships between the network nodes and the conditional probability distribution values of the network nodes. 6. The method for determining a remote sector according to claim 1, wherein determining the initial parameters of the scene sample comprises: Obtain key influencing parameters in the application scenario based on expert experience; The scene sample initial parameters are obtained according to the intersection of the key influencing parameters and the sample initial parameters. 7. A device for determining a remote sector, comprising: A characteristic parameter determination module, used for selecting a plurality of initial parameters of a base station sector as characteristic parameters according to an application scenario of the base station sector, wherein the initial parameters are determined based on the relationship characteristics between the base station sector and the surrounding sectors of the base station sector, the optical fiber length from the baseband unit of the base station sector to the signal processing unit or the signal transmission unit, and the base station coverage distance characteristics of the base station sector; A remote sector discrimination module is used to input the characteristic parameters into a remote sector discrimination model to obtain the remote sector discrimination result of the base station sector output by the remote sector discrimination model; wherein the remote sector discrimination model is obtained on the basis of a preset Bayesian network under a plurality of the application scenarios, based on a plurality of scene sample initial parameters of each of the application scenarios and label training of the remote sector discrimination result, the scene sample initial parameters are determined based on sample initial parameters, and the preset Bayesian network under each of the application scenarios is constructed based on a plurality of the scene sample initial parameters and a dependency relationship between the plurality of the scene sample initial parameters. 8. An electronic device, comprising a processor and a memory storing a computer program, wherein the processor implements the method for determining a remote sector as claimed in any one of claims 1 to 6 when executing the computer program. 9. A computer program product, comprising a computer program, characterized in that when the computer program is executed by a processor, the method for determining a remote sector according to any one of claims 1 to 6 is implemented. 10. A non-transitory computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the method for determining a remote sector according to any one of claims 1 to 6 is implemented.