Beam configuration method and device, electronic device and computer storage medium

Abstract

A beam configuration method and device, an electronic device and a computer storage medium are provided. The beam configuration method includes: forming at least one beam cluster from each beam of a base station; determining the configuration of the at least one beam cluster; and configuring the beams in the at least one beam cluster accordingly according to the determined configuration.

Description

Technical Field

[0001] The present disclosure relates to the field of wireless communication technology. More specifically, the present disclosure relates to a beam configuration method and device, an electronic device, and a computer storage medium. Background Art

[0002] In 5G networks, base stations are deployed more densely and in greater numbers, and the energy consumption of base stations is a major expense for operators. Therefore, base station energy-saving technology is particularly important in 5G networks. For base station energy saving, existing technologies are mainly divided into two categories:

[0003] (1) Energy saving based on cells: Taking the cell as the control unit, the switching of the radio unit (RU), digital unit (DU) and control unit (CU) of the entire cell is controlled to achieve energy saving of the base station to which the cell belongs. When there are no users accessing the cell, the cell is turned off, that is, the RU, DU and CU of the cell are turned off.

[0004] (2) Energy saving based on carriers: With carriers as control units, the carriers are switched on and off as a whole to achieve energy saving for the base stations to which the carriers belong. When a carrier has no users accessing it, the carrier is turned off. The method of turning off a carrier can be achieved by changing the configuration information of the base station.

[0005] The problems of the prior art mainly include the following three aspects:

[0006] 1) Spatially, traffic in different areas within a cell or carrier is unevenly distributed, resulting in poor energy-saving effects.

[0007] Generally, the coverage of a cell or a carrier is large, and it is rare that there is no user access at all. If the existing technical solution is used, the coverage of the cell or carrier needs to be always open. However, the distribution of users within the coverage of the cell or carrier is often uneven, that is, the distribution of users in different areas of the cell or carrier may be different. If the existing technical solution is used for the switch control at the cell or carrier level, since there are users accessing the cell or carrier, the wireless service of the cell or carrier cannot be turned off, but there are no service users in some areas of the cell. For these areas, the wireless service turned on will cause energy waste (also called energy waste or resource waste) of the base station.

[0008] 2) In terms of space, the same network parameter configuration is used for different areas with different service types within a cell or carrier, resulting in poor energy saving effect.

[0009] In the prior art, when the energy-saving function is not enabled for a cell or carrier, the configuration of network parameters within all ranges of a cell or carrier, including the synchronization signal block (SSB) period, power size and bandwidth part (BWP), etc. are the same. However, from a spatial perspective, the service types corresponding to different areas under the coverage of a cell may be different. If the same network parameter configuration is used for each area with different service types, it will cause a waste of resources.

[0010] 3) In terms of time, the number of users accessing a cell or carrier varies over time, but fixed network parameter configuration will result in poor energy saving effects.

[0011] The number of users in each area of ​​a cell is not fixed. The number of users changes with people's daily activities and schedules. If the configuration of network parameters in the area remains unchanged at different times, for example, the same signal transmission power is always used, it will cause a waste of resources. In addition, the existing technology uses fixed network parameter configurations and cannot adapt to dynamic user needs. Users cannot obtain the best quality of service (QoS) experience.

[0012] In summary, according to the solutions of the prior art, the energy saving methods at the cell or carrier level cannot solve the problems of uneven user distribution, different types of services and time-varying number of users in the cell, and the energy saving effect of the existing solutions is poor. Summary of the invention

[0013] The exemplary embodiments of the present disclosure provide a beam configuration method and apparatus, an electronic device, and a computer storage medium to solve the problem of base station energy waste caused by uneven user distribution in a cell, different types of services, and a time-varying number of users.

[0014] According to an exemplary embodiment of the present disclosure, a beam configuration method is provided, including: forming at least one beam cluster from individual beams of a base station; determining the configuration of the at least one beam cluster; and correspondingly configuring the beams within the at least one beam cluster according to the determined configuration.

[0015] Optionally, the beam configuration method may further include: collecting beam data of the beam of the base station, wherein the beam data may at least include data used for communication, wherein the step of forming at least one beam cluster from each beam of the base station may include: forming at least one beam cluster from each beam of the base station according to the beam data.

[0016] Optionally, the data used for communication includes at least one of the following: user location information, current time information, service type information of the beam, service traffic information of the beam, service distribution information of the beam, beam synchronization signal period information, beam transmission power information, signal quality information, and beam number information.

[0017] Optionally, the service type information includes at least one of the following: a statistical value of a user's moving speed, a signal's service quality level identifier, and an uplink / downlink data ratio.

[0018] Optionally, the service traffic includes at least one of the following: throughput, resource block utilization, number of IP data packets, and number of access users.

[0019] Optionally, the step of forming at least one beam cluster from each beam of the base station according to the beam data includes: clustering beams with similar characteristics into the same beam cluster according to the beam data.

[0020] Optionally, the step of clustering beams with similar characteristics into the same beam cluster according to the beam data may include: extracting service characteristic information of each beam from the data used for communication in the beam data, the service characteristic information including at least one of service traffic information, service type information, and service distribution information; and clustering at least one beam with similar characteristics into a beam cluster according to the extracted service characteristic information.

[0021] Optionally, the step of clustering at least one beam having similar characteristics into a beam cluster according to the extracted service characteristic information includes: clustering at least one beam having a characteristic similarity exceeding a first threshold into a beam cluster according to the extracted service characteristic information.

[0022] Optionally, the step of clustering beams with similar characteristics into the same beam cluster according to the beam data includes: clustering beams with similar characteristics into the same beam cluster based on a mean shift aggregation method according to the beam data.

[0023] Optionally, according to the beam data, based on the average displacement aggregation method, the step of clustering beams with similar characteristics into the same beam cluster includes: obtaining sample points corresponding to the beam data of each beam of the base station; randomly selecting a sample point from all sample points as the center, and finding all sample points within a preset radius of the center; calculating the average displacement vector of the displacement vectors from all sample points within the preset radius to the center; clustering the beams corresponding to all sample points within the preset radius according to the average displacement vector; when there are unclustered beams, taking the sample points of all unclustered beams as all sample points, and jumping to the step of randomly selecting a sample point from all sample points as the center.

[0024] Optionally, the step of forming at least one beam cluster from each beam of the base station may include: obtaining a service intensity heat map corresponding to the base station; and clustering at least one beam of the base station into a beam cluster based on color similarity in the service intensity heat map.

[0025] Optionally, the step of forming at least one beam cluster from each beam of the base station may include: determining the switching correlation between each beam according to the beam switching relationship corresponding to the base station; and clustering at least one beam of the base station into a beam cluster based on the switching correlation between each beam.

[0026] Optionally, the step of determining the configuration of at least one beam cluster may include: calculating a traffic prediction value of the beam cluster; determining the configuration of the beam cluster based on the traffic prediction value of the beam cluster, wherein the configuration of the beam cluster includes at least one of a period of the synchronization signal SSB, a power level, a beam cluster bandwidth, a direction of the beam cluster, and a switch of the beam cluster.

[0027] Optionally, the step of calculating the traffic prediction value of a beam cluster may include: determining the parameters of each time period kernel in the prediction model based on the data used for communication in the beam data of each beam in the beam cluster, and obtaining a prediction model based on a time series, wherein the prediction model includes a periodic kernel based on an hour as a periodic granularity; using the prediction model based on a time series to determine the traffic prediction value of each beam cluster at a predetermined time.

[0028] Optionally, after configuring the beams within the beam cluster, the beam configuration method may further include: calculating the prediction deviation of the beam cluster based on the traffic prediction value and the actual traffic value of the beam cluster; determining whether to re-cluster the beams within the beam cluster and the beams associated with the beam cluster based on the calculated prediction deviation; determining the configuration of the re-clustered beam cluster; and configuring the beams within the re-clustered beam cluster accordingly according to the determined configuration.

[0029] Optionally, the beam configuration method also includes: collecting beam data of the beam of the base station, wherein the beam data at least includes data for predicting traffic demand change events, wherein, after configuring the beams in the beam cluster, the beam configuration method may also include: predicting traffic demand change events based on the data for predicting traffic demand change events included in the beam data; when a traffic demand change event is predicted, for at least one beam cluster associated with the predicted traffic demand change event, re-clustering the beams in the at least one beam cluster and the beams associated with the at least one beam cluster; determining the configuration of the re-clustered beam cluster; and correspondingly configuring the beams in the re-clustered beam cluster according to the determined configuration.

[0030] Optionally, the data used to predict traffic demand change events includes at least one of the following: climate data, network application data, multimedia data, traffic data, public utility data, and holiday data.

[0031] According to an exemplary embodiment of the present disclosure, a beam configuration device is provided, including: a beam cluster forming unit, configured to form at least one beam cluster from various beams of a base station; a configuration determination unit, configured to determine the configuration of the at least one beam cluster; and a beam configuration unit, configured to configure the beams within the at least one beam cluster accordingly according to the determined configuration.

[0032] Optionally, the beam configuration device may further include: a beam data collection unit configured to collect beam data of the beam of the base station, wherein the beam cluster forming unit may be configured to: form at least one beam cluster from each beam of the base station according to the beam data.

[0033] Optionally, the data used for communication includes at least one of the following: user location information, current time information, service type information of the beam, service traffic information of the beam, service distribution information, beam synchronization signal period information, transmission power information, signal quality information, and beam number information.

[0034] Optionally, the service type information includes at least one of the following: a statistical value of a user's moving speed, a signal's service quality level identifier, and an uplink / downlink data ratio.

[0035] Optionally, the service traffic includes at least one of the following: throughput, resource block utilization, number of IP data packets, and number of access users.

[0036] Optionally, the beam cluster forming unit is configured to: cluster beams with similar characteristics into the same beam cluster according to the beam data.

[0037] Optionally, the beam cluster forming unit can also be configured to: extract service characteristic information of each beam from the data used for communication in the beam data, the service characteristic information including at least one of service traffic information, service type information, and service distribution information; and cluster at least one beam with similar characteristics into a beam cluster according to the extracted service characteristic information.

[0038] Optionally, the beam cluster forming unit is further configured to: cluster at least one beam whose characteristic similarity exceeds a first threshold into a beam cluster according to the extracted service characteristic information.

[0039] Optionally, the beam cluster forming unit is further configured to: cluster beams with similar characteristics into the same beam cluster based on the mean shift aggregation method according to the beam data.

[0040] Optionally, the beam cluster forming unit can also be configured to: obtain sample points corresponding to the beam data of each beam of the base station; randomly select a sample point from all sample points as the center, and find all sample points within a preset radius of the center; calculate the average displacement vector of the displacement vectors from all sample points within the preset radius to the center; cluster the beams corresponding to all sample points within the preset radius according to the average displacement vector; when there are unclustered beams, take the sample points of all unclustered beams as all sample points, and jump to the step of randomly selecting a sample point from all sample points as the center.

[0041] Optionally, the beam cluster forming unit may be configured to: obtain a service intensity heat map corresponding to the base station; and cluster at least one beam of the base station into a beam cluster based on color similarity in the service intensity heat map.

[0042] Optionally, the beam cluster forming unit may be configured to: determine a switching correlation between the beams according to a beam switching relationship corresponding to the base station; and cluster at least one beam of the base station into a beam cluster based on the switching correlation between the beams.

[0043] Optionally, the configuration determination unit may be configured to: calculate a traffic prediction value of a beam cluster; and determine a configuration of the beam cluster based on the traffic prediction value of the beam cluster, wherein the configuration of the beam cluster includes at least one of a period of a synchronization signal SSB, a power level, a beam cluster bandwidth, a direction of the beam cluster, and a switch of the beam cluster.

[0044] Optionally, the configuration determination unit can also be configured to: determine the parameters of each time period kernel in the prediction model based on the data used for communication in the beam data of each beam in the beam cluster, and obtain a prediction model based on time series, wherein the prediction model includes a period kernel based on the hour as the period granularity; use the prediction model based on time series to determine the traffic prediction value of each beam cluster at a predetermined time.

[0045] Optionally, the beam configuration device may also include: a deviation calculation unit, configured to calculate the prediction deviation of the beam cluster based on the traffic prediction value and the actual traffic value of the beam cluster; a first re-clustering unit, configured to determine whether to re-cluster the beams within the beam cluster and the beams associated with the beam cluster based on the calculated prediction deviation; a configuration redetermination unit, configured to determine the configuration of the beam cluster obtained by re-clustering; and a reconfiguration unit, configured to configure the beams within the beam cluster obtained by re-clustering accordingly according to the determined configuration.

[0046] Optionally, the beam configuration device may further include: a beam data collection unit, configured to collect beam data of the beam of the base station, wherein the beam data may include data for predicting traffic demand change events, wherein the beam configuration device may further include: an event prediction unit, configured to predict a traffic demand change event based on the data for predicting a traffic demand change event; a second re-clustering unit, configured to, when a traffic demand change event is predicted, re-cluster the beams in the at least one beam cluster and the beams associated with the at least one beam cluster associated with the predicted traffic demand change event; a configuration re-determination unit, configured to determine the configuration of the beam cluster obtained by re-clustering; and a reconfiguration unit, configured to configure the beams within the re-clustered beam cluster accordingly according to the determined configuration.

[0047] Optionally, the data used to predict traffic demand change events includes at least one of the following: climate data, network application data, multimedia data, traffic data, public utility data, and holiday data.

[0048] According to an exemplary embodiment of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the beam configuration method according to the exemplary embodiment of the present disclosure is implemented.

[0049] According to an exemplary embodiment of the present disclosure, an electronic device is provided, including: a processor; and a memory storing a computer program, wherein when the computer program is executed by the processor, a beam configuration method according to an exemplary embodiment of the present disclosure is implemented.

[0050] According to the beam configuration method and apparatus, electronic device and computer storage medium of the exemplary embodiments of the present disclosure, at least one beam cluster is first formed from the various beams of the base station, and then the configuration of the at least one beam cluster is determined. Finally, the beams in the at least one beam cluster are configured accordingly according to the determined configuration, so that beams with different beam characteristics are clustered into different beam clusters, and different configuration parameters are configured for different beam clusters, thereby achieving a smaller-scale and more accurate regional division of the space within the cell coverage, realizing fine-grained control, and saving the energy consumption of the base station.

[0051] Additional aspects and / or advantages of the present general inventive concept will be set forth in part in the following description and in part will be apparent from the description or may be learned through practice of the present general inventive concept. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The above and other objects and features of the exemplary embodiments of the present disclosure will become more apparent through the following description in conjunction with the accompanying drawings which exemplarily illustrate the embodiments, in which:

[0053] Figure 1a A schematic diagram of a cell-level energy-saving solution in the prior art is shown;

[0054] Figure 1b A schematic diagram of a carrier-level energy-saving solution in the prior art is shown;

[0055] Figure 1c A schematic diagram showing energy waste caused by uneven user distribution in a cell in the prior art;

[0056] Figure 1d A schematic diagram showing energy waste caused by using the same configuration for different service types in the prior art;

[0057] Figure 1e A schematic diagram showing energy waste caused by the time-varying number of users in the prior art;

[0058] Figure 2 A flowchart showing a beam configuration method according to an exemplary embodiment of the present disclosure is shown;

[0059] Figure 3 A business flow diagram according to an exemplary embodiment of the present disclosure is shown;

[0060] Figure 4 An example of clustering results of beam clustering according to an exemplary embodiment of the present disclosure is shown;

[0061] Figure 5 Shows the beam switching relationship of the base station;

[0062] Figure 6a and Figure 6b A schematic diagram showing an example of a beam cluster;

[0063] Figure 7 A schematic diagram showing the configuration of a beam cluster at different times;

[0064] Figure 8a An example of a multi-kernel Gaussian prediction model according to an exemplary embodiment of the present disclosure is shown;

[0065] Figure 8b shows an example structure of a Gaussian kernel according to an exemplary embodiment of the present disclosure;

[0066] Figures 9a to 9f A flowchart showing a beam configuration method according to an exemplary embodiment of the present disclosure is shown;

[0067] Fig.10A schematic diagram showing beams used when regenerating a beam cluster;

[0068] Fig.11 An example of forming a beam cluster for a business type in a commercial area is shown;

[0069] Fig.12 An example of forming a beam cluster for a time-varying number of user accesses in a certain area is shown;

[0070] Fig.13 An example of adaptive change of beam clusters in a scenario with uneven user distribution is shown;

[0071] Fig.14 An example of forming a new beam cluster when a traffic demand change event is detected is shown;

[0072] Fig.15 Shows time flow graphs for four land use types unique to cities;

[0073] Fig.16 An example of energy saving effect based on the present disclosure is shown;

[0074] Fig.17 Shows the occupancy ratios of office area, public facilities area, residential area and green area in the example city land planning;

[0075] Fig.18 A block diagram showing a beam configuration device according to an exemplary embodiment of the present disclosure;

[0076] Fig.19 A block diagram showing a beam configuration device according to an exemplary embodiment of the present disclosure;

[0077] Fig. 20 A schematic diagram showing an electronic device according to an exemplary embodiment of the present disclosure. DETAILED DESCRIPTION

[0078] Reference will now be made in detail to exemplary embodiments of the present disclosure, examples of which are shown in the accompanying drawings, wherein like reference numerals refer to like parts throughout. The embodiments will be described below with reference to the drawings in order to explain the present disclosure.

[0079] like Figure 1a As shown in the figure, it is a schematic diagram of the energy-saving scheme for the cell level in the prior art. Given a base station and three cells, when there are users accessing the three cells, the three cells are in the ON state, that is, RU, DU and CU are all in the ON state. When there is no user access in a cell, the cell is turned off and the cell is in the OFF state, that is, the RU, DU and CU of the cell are turned off.

[0080] like Figure 1bAs shown in the figure, it is a schematic diagram of the energy-saving scheme for the carrier level in the prior art. Given two carriers, carrier component (CC) 1 and CC2, when the number of users is relatively large, both carriers are in the on state to provide sufficient frequency resources for users. When the number of users is relatively small, some carriers have no user access, and the carrier can be turned off, for example, CC2. The method of turning off the carrier is actually to turn off the frequency resources of the base station, that is, the resource block (RB) corresponding to the carrier, to save the energy of the base station.

[0081] With respect to the above-mentioned prior art solutions, the inventors of the present application have found the following technical problems:

[0082] When a cell or carrier is closed in an area, users cannot use wireless services or obtain good wireless QoS. In order to ensure the user experience, in the existing base station energy-saving solution, the energy-saving function of the base station will be triggered only when there is no user access to an area. So in general, the existing cell-based and carrier-based energy control are based on large areas and coarse granularity. The energy-saving function will be triggered only when there is no service user in a large area. Therefore, the energy-saving granularity of the traditional method is too large, not fine enough, and the energy-saving effect is not good. The actual energy-saving efficiency is low, and fine energy control cannot be achieved.

[0083] In addition, from the time dimension, most base stations are in a relatively idle state for a long time, and the number of connected users is small. For example, 10% of the first-class base stations may carry 50% of the traffic, and these base stations are heavily loaded. 40% of the second-class base stations may carry 45% of the traffic, and the other 50% of the third-class base stations only carry 5% of the traffic and are relatively idle. Therefore, the traffic within the coverage area of ​​each base station is unevenly distributed. However, in the prior art, even if the number of users connected to the base station is small, the base station must always be in the on state, which is another important reason for energy waste.

[0084] Specifically, the problems of the prior art mainly include the following three aspects:

[0085] 1) Spatially, traffic in different areas within a cell or carrier is unevenly distributed, resulting in poor energy-saving effects.

[0086] For example, if only a certain area of ​​a cell has very few users accessing it, then this cell cannot be shut down. In this case, the base station will still send broadcast signals to other areas of the cell (areas without user access), thus wasting energy. Similarly, for energy saving based on carriers, for example, if a carrier only serves very few users in a certain area, then the energy of this carrier for other areas is wasted.

[0087] like Figure 1c As shown in FIG. 1 , in a cell, users are unevenly distributed, that is, some local areas have users accessing, while some local areas have no users accessing. If wireless services are always enabled in areas without users accessing, the base station still needs to send broadcast signals, which is actually a waste of energy.

[0088] Therefore, in this scenario, the local uneven distribution of users within a cell causes energy waste.

[0089] 2) In terms of space, the same network parameter configuration is used for different areas with different service types within a cell or carrier, resulting in poor energy saving effect.

[0090] like Figure 1d As shown, in a cell, some areas are dedicated to pedestrians (pedestrian streets, open squares, etc.), and the access users are pedestrians (slow-moving users), and some areas are dedicated to vehicle users (viaducts, highways, etc.), and the access users are vehicle users (fast-moving users). These two types of users have different requirements for the synchronization signal period of the network due to their different moving speeds. Usually, high-speed moving users have higher requirements for synchronization signals and need short-period SSB configuration. Low-speed moving users have relatively low requirements for synchronization signals, and long-period SSB configuration can meet the requirements, but short-period SSB configuration is also possible. In order to meet the needs of all users, the network usually configures the network as a short-period SSB to meet both pedestrian users and vehicle users. This is because the existing technology does not classify the service types in a cell, so it usually configures the cell according to the shortest SSB, that is, a cell cannot achieve different configurations, that is, the configuration parameters of pedestrian users and vehicle users are the same, and they all use short-period SSB, which means that the base station frequently sends SSB. For pedestrian users, since they have low requirements for latency, long-period SSB can meet their needs. Frequent sending of SSB signals by base stations causes energy waste.

[0091] Therefore, in this scenario, the use of indifferent network configurations for different service types in a cell results in a waste of energy.

[0092] 3) In terms of time, the number of users accessing a cell or carrier varies over time, but fixed network parameter configuration will result in poor energy saving effects.

[0093] like Figure 1eAs shown, within a cell, the number of users is time-varying, but in the prior art, the network parameter configuration within the cell remains unchanged, that is, the same network parameter configuration is used at different times. When the number of users is small, the base station signal transmission power in the corresponding area can be reduced or turned off accordingly, otherwise it will cause waste of energy. For example, at 08:00, there are a large number of users in a certain cell, and the base station uses power p1 to cover a large area. However, at 14:00, the number of users in the cell is greatly reduced, but in the prior art, the base station uses the same configuration and the power is still p1, resulting in waste of energy. If at 14:00, the base station can adaptively change the power to use power p2, and p2 < p1, then the purpose of energy saving can be achieved to a certain extent. Therefore, the adaptive network parameter configuration can achieve the effect of energy saving.

[0094] Therefore, in this scenario, using the same network parameter configuration in a cell at different times causes waste of energy.

[0095] For the above technical problems, the embodiments of the present application propose a base station intelligent energy-saving method with beam clusters as the control unit in the 5G network. There are many small beams in the base station of the 5G network. The embodiments of the present application propose to collect the beam data corresponding to the beams of the base station, cluster (which can also be called aggregation) the small beams into each beam cluster based on the service characteristics of each beam, predict the traffic of each beam cluster, configure network parameters for each beam cluster based on the predicted traffic, and perform corresponding operations according to the configured network parameters.

[0096] The solution proposed by the embodiments of the present application enables the base station to adapt to scenarios with different spatial and temporal changes, and then can use beam clusters as units to more accurately configure the beams covering different ranges in the cell, adapt to the time-varying user volume, unevenly distributed services, and different service types, and achieve the purpose of high-precision energy saving. The embodiments of the present application perform a more small-scale and accurate regional division of the space within the cell coverage, through learning and prediction of data, and also implement an intelligent granularity control method in terms of time, saving more base station energy consumption.

[0097] Figure 2 The flowchart of the beam configuration method according to an exemplary embodiment of the present disclosure is shown.

[0098] Referring to Figure 2 , in step S201, at least one beam cluster is formed from the respective beams of the base station.

[0099] In 5G systems, beamforming technology will be used in frequency bands above 6 GHz, which means that a cell will be composed of multiple beams, each with its own independent coverage space. Due to the geographical space environment, the uneven distribution of user access numbers, and the time-varying characteristics, the wireless service requirements within each beam coverage also change accordingly.

[0100] There will be many beams under the 5G base station, each beam can independently provide wireless services to users, and the data about the service users of the beam will be used to cluster beam clusters. The service characteristics of each beam can be first extracted from the database (in the embodiment of the present application, the service characteristics of the beam can also be referred to as the characteristics of the beam or beam characteristics). These service characteristics include the distribution of services in time and space (such as the statistical values ​​of beam number, time and number of user access), service type (such as the statistical values ​​of beam number and user movement speed, or beam number and signal service quality level identifier (QoS Class Identifier, referred to as QCI), or beam number and uplink / downlink data ratio, etc.), and service traffic (such as beam number and throughput, resource block (Resource block, referred to as RB) utilization rate, number of access users or time statistics, etc.), and then compare the service characteristics of these beams, and cluster beams with similar service characteristics into a beam cluster. It should be noted that the number of beams in a beam cluster can be one or more. And the beams in the same beam cluster can be adjacent or non-adjacent in space.

[0101] In an exemplary embodiment of the present disclosure, before at least one beam cluster is formed from each beam of the base station, the beam data of the base station may be collected. The base station or the server for energy saving collects the beam data in real time, and stores the collected beam data in a database. Here, the beam data may include at least data for communication. The data for communication may include at least one of user location information, current time information, service type information of the beam, service flow information of the beam, service distribution information of the beam, beam synchronization signal period information, transmission power information of the beam, signal quality information, beam number information, etc. Here, the location information includes but is not limited to the geographic location. Taking the geographic location as an example: in addition to the longitude and latitude information, it may also be the corresponding administrative area information, or the mapping index of the processed geographic location. Alternatively, the beam index id also corresponds to a small area. In an exemplary embodiment of the present disclosure, when at least one beam cluster is formed from each beam of the base station, at least one beam cluster may be formed from each beam of the base station according to the beam data. In an exemplary embodiment of the present disclosure, when at least one beam cluster is formed from the respective beams of the base station according to the beam data, beams having similar characteristics may be clustered into the same beam cluster according to the beam data.

[0102] Specifically, beam data may include wireless data from a base station or terminal, that is, conventional data in a communication system, which may also be referred to as data used for communication; these data belong to traditional wireless data and can be used as the most basic input. These basic wireless data include user location, time, beam service type, service distribution, beam service flow, beam synchronization signal period, transmission power, signal quality, beam number, etc. The measurement indicators of service flow (i.e., the amount of service) may be base station throughput, RB utilization rate, number of IP packets, number of access users, etc.; the classification of service types may be based on mobile speed (for example, it may be distinguished according to different levels of speed such as 30 kilometers per hour, 60 kilometers per hour, 120 kilometers per hour, etc.), or QCI (such as video, voice, data, etc.), or classification according to the ratio of uplink and downlink data (for example, virtual reality (VR) service (more uplink data than downlink data) and video viewing service (more downlink data than uplink data)); or other classification methods.

[0103] In an exemplary embodiment of the present disclosure, when clustering beams with similar characteristics into the same beam cluster according to the beam data, service characteristic information of each beam can be first extracted from the data used for communication in the beam data, the service characteristic information including at least one of service traffic information, service type information, and service distribution information. Then, based on the extracted service characteristic information, at least one beam with similar characteristics can be clustered into a beam cluster. Specifically, at least one beam whose characteristic similarity exceeds a first threshold can be clustered into a beam cluster.

[0104] Specifically, the service characteristics of each beam can be extracted from the historical data of each beam in the database (for example, the data of the previous few days, such as day X), and at least one of the service flow, service type, and service distribution of each beam can be obtained. Here, service flow refers to: time series data, and the specific measurement index can be at least one of base station throughput, RB utilization rate, number of IP data packets, number of access users, etc.; service type refers to: classification of services, which can be classified according to mobile speed; or classified according to QCI; or classified according to the ratio of uplink and downlink data; or other classification methods; service distribution refers to: the distribution of services in space, that is, the statistical value between the beam number (or the geographical location information of the beam) and at least one of the number of user access, RB utilization rate, QCI, user mobile speed, channel quality, etc.

[0105] In an exemplary embodiment of the present disclosure, when clustering at least one beam with similar characteristics into a beam cluster according to the extracted service feature information, at least one beam whose feature similarity exceeds a first threshold can be clustered into a beam cluster according to the extracted service feature information.

[0106] In an exemplary embodiment of the present disclosure, when beams with similar characteristics are clustered into the same beam cluster according to the beam data, the beams with similar characteristics can be clustered into the same beam cluster based on the average shift aggregation method according to the beam data. Since no prior knowledge is required when clustering beam clusters using the average shift aggregation method, such as the number of beams in the beam cluster does not need to be known in advance, the efficiency is high, the flexibility is high, the computational complexity is low, and the method can be applied in various uneven traffic distribution scenarios.

[0107] In an exemplary embodiment of the present disclosure, when beams with similar characteristics are clustered into the same beam cluster based on the beam data and the average displacement aggregation method, the sample points corresponding to the beam data of each beam of the base station can be first obtained, and then a sample point is randomly selected from all the sample points as the center, and all sample points within a range of a preset radius with the selected center as the center are found, and the average displacement vector corresponding to the displacement vectors from all sample points within the preset radius to the center is calculated, and the beams corresponding to all sample points within the preset radius are clustered according to the average displacement vector. When there are beams that are not clustered, the sample points of all beams that are not clustered are taken as all sample points, and the step of randomly selecting a sample point from all sample points as the center is jumped. Here, the sample point represents a vector of a characteristic quantitative index value of a beam, including one or more of service type information, service flow information, and time information. The radius refers to the distance value between vectors, and the distance can be but is not limited to Euclidean distance, Manhattan distance, Chebyshev distance, Mahalanobis distance, Hamming distance, or correlation distance.

[0108] Specifically, the specific steps for clustering input samples can be:

[0109] (1) Randomly select a sample point from the input sample points (the sample points corresponding to the beam data of each of the above beams) as the center;

[0110] (2) Then find all sample points within a radius of r from the center (the size of the parameter r can be determined according to the domain value of all training samples), marked as c_i, where i represents the identifier of the beam cluster to be created;

[0111] (3) Calculate the displacement vectors from these sample points c_i to the center center to obtain their average displacement vector shift;

[0112] (4) Update the new center to center = center + shift, which is the sum of the center center in step (1) and the shift shift in step (3);

[0113] (5) Repeat steps (2), (3), and (4) until the average displacement vector shift is very small, such as less than a certain threshold ε, and then execute step (6);

[0114] (6) Compare the new sample point c_i with the old sample point c_(i-1). If they are close, cluster them. Otherwise, create a new beam cluster i=i+1. The distance between the new sample point c_i and the old sample point c_(i-1) can be calculated by comparing the distance between the sample point cluster to which the new sample point c_i belongs and the sample point cluster to which the old sample point c_(i-1) belongs.

[0115] (7) Repeat steps (1)(2)(3)(4)(5)(6) until all sample points have been visited.

[0116] For example, Figure 3 A schematic diagram showing the traffic of each beam of a base station changing over time, wherein beams 1 to 4 mainly serve pedestrian users (i.e., the service type is pedestrian user service or pedestrian service), and beam 5 mainly serves vehicle users (i.e., the service type is vehicle service). The service feature information of each beam can be firstly extracted (including at least one of service traffic information, service type information, and service distribution information), and then beams with similar features can be clustered into a beam cluster according to the extracted service feature information. For example, according to Figure 3 The business traffic in the traffic flow is extracted to extract the business type information. Then, through the average clustering method, beams 1 to 4 serving pedestrian users are clustered into a beam cluster, and beams 5 serving vehicle users are clustered into a beam cluster.

[0117] Figure 4 An example of a clustering result of beam clustering according to an exemplary embodiment of the present disclosure is shown. Figure 4 As shown, the beams of the base station include beams 1 to 18 in the figure. Through the clustering method proposed in the embodiment of the present application, the above 18 beams can be clustered into three categories. The beams in the stadium are clustered into beam cluster 1 (including beams 7 to 9 and beams 13 to 15), and the beam cluster 2 for pedestrians (including beams 1 to 4, beam 10, and beam 16) and the beam cluster 3 for vehicles (including beams 5 to 6, beams 11 to 12, and beams 17 to 18) are clustered outside the stadium.

[0118] In addition to the above-mentioned average displacement method, another exemplary embodiment of the present disclosure proposes that beam clusters can also be clustered through a service intensity heat map. When at least one beam cluster is formed from each beam of a base station, a service intensity heat map of the base station can be first obtained. The service intensity heat map can reflect the service intensity of each beam under the base station, and image recognition is performed on the obtained service intensity heat map. Then, based on the color similarity in the service intensity heat map, at least one beam of the base station is clustered into a beam cluster, and beams with similar colors in the service intensity heat map are clustered into a beam cluster. For example, at least one beam in an area where the color similarity in the service intensity heat map exceeds a second threshold is clustered into a beam cluster. Here, the service intensity heat map represents the service intensity of an area by color, and the service intensity is represented by one or more of the following: the number of users, the amount of traffic, and the resource block (RB) utilization rate.

[0119] In addition to the above-mentioned average displacement method and the service intensity heat map method, another exemplary embodiment of the present disclosure proposes that beam clusters can also be clustered according to the beam switching relationship corresponding to the base station. Figure 5 When at least one beam cluster is formed from each beam of the base station, the beam switching relationship of the base station (such as the beam switching relationship of the base station) can be shown. Figure 5 As shown in FIG. 1 , the switching correlation between each beam is calculated, and then based on the switching correlation between each beam, at least one beam of the base station is clustered into a beam cluster. For example, at least one beam whose switching correlation exceeds a third threshold value can be clustered into a beam cluster. Since the frequently switched beams are the trajectories where the user often moves, there is a strong correlation between such beams. Specifically, as Figure 5 As shown, firstly, text recognition is performed on the beam switching table of the base station to identify the frequently switched beams and cluster the frequently switched beams together. The beam switching table of the base station is the correspondence between the beam number and the adjacent beam number to be switched. The beam switching table records the user equipment switching from one beam to another.

[0120] It should be noted that beam clusters are not limited to spatially adjacent beam areas, but can also be spatially non-adjacent beam areas. If two beams that are far apart have similar service characteristics, they can also be clustered into a beam cluster. The number of beams in a beam cluster can be one or more. Figure 6a A schematic diagram showing an example of aggregating two beams into a beam cluster, Figure 6b FIG. 4 is a schematic diagram showing an example of aggregating multiple beams into three beam clusters. Figure 6b As shown, beam cluster 1 includes three beams, beam cluster 2 includes two beams (beam 1 and beam 2 respectively), and beam cluster 3 includes two beams. Figure 6bOnly the beams contained in the beam cluster are shown to be spatially adjacent beams, but beam clusters composed of spatially non-adjacent beams are also within the protection scope of the present invention. The purpose of forming a beam cluster is to reduce complexity and facilitate fast calculation. Since the service characteristics of beams in a beam cluster are extremely similar, they can be controlled uniformly to avoid controlling each beam separately. Among them, the beam cluster is a virtual concept, and it is not required that the beams must be spatially adjacent.

[0121] The embodiments of the present application propose that, in addition to the above-mentioned clustering method using artificial intelligence to obtain beam clusters, non-artificial intelligence methods can also be used to obtain beam clusters. For example, the Euclidean distance between the vectors corresponding to the beam data can be compared to determine whether two beams are similar, thereby clustering similar beams into a beam cluster.

[0122] In step S202, the configuration of the at least one beam cluster is determined.

[0123] In an exemplary embodiment of the present disclosure, when determining the configuration of the at least one beam cluster, the traffic prediction value of each beam cluster may be calculated first, and then the configuration of each beam cluster may be determined based on the traffic prediction value of each beam cluster. Here, the configuration of the beam cluster includes at least one of the period of the synchronization signal SSB, the power size of the beam cluster, the bandwidth of the beam cluster (such as BWP), the direction of the beam cluster, and the switch of the beam cluster.

[0124] In an exemplary embodiment of the present disclosure, when predicting the traffic (also referred to as the number of user access) of each beam cluster, the prediction can be made based on an artificial intelligence method, and a prediction model can be constructed to predict the traffic prediction value based on the prediction model.

[0125] When constructing a prediction model, you can first analyze the change pattern of the traffic in each beam cluster based on the data used for communication collected in real time or within a preset time period (as training data for constructing the prediction model), and construct a prediction model for the traffic in each beam cluster based on the change pattern of the traffic in each beam cluster. You can build a prediction model offline based on the data used for communication within a preset time period, or you can build a prediction model online based on the data collected in real time. Then use the constructed prediction model to calculate the traffic prediction value of the beam cluster. Here, the constructed prediction model can be updated in real time based on the changes in the training data.

[0126] The above prediction model may be a prediction model based on time series, such as NGP (multiple kernel Gaussian process), which may also be called an NGP prediction model or a multi-kernel Gaussian prediction model.

[0127] In an exemplary embodiment of the present disclosure, when calculating the traffic prediction value of each beam cluster, the parameters of each time period kernel in a pre-constructed prediction model can be first determined based on the data used for communication in the beam data of each beam in the beam cluster to obtain a prediction model based on a time series. Then, the prediction model based on a time series is used to determine the traffic prediction value of the beam cluster at a predetermined time, such as the traffic prediction value at the current time, or the traffic prediction value at a certain time after a set time period, such as the traffic prediction value after 1 minute, etc.

[0128] The multi-core Gaussian prediction model includes at least one time period kernel, and may include at least a fine-grained time period kernel, such as a period kernel based on an hour as a period granularity.

[0129] Specifically, after the beam cluster is formed, it is necessary to perform periodic prediction of the traffic within the beam cluster. The period can be T1, and the determination of period T1 can be determined according to the speed of business changes in each area. For example, if the number of user accesses in an area changes frequently, then the number of user accesses in this area needs to be predicted frequently, and the traffic prediction period of the beam cluster in this area will be shorter. Otherwise, if the number of user accesses in an area changes slowly, the prediction period can be extended accordingly, and the traffic prediction period of the beam cluster in this area will be longer. The length of the specific period can be determined by the speed of change of the number of user accesses. Periodic traffic prediction and configuration for beam clusters can adapt to dynamic business changes and achieve precise energy saving.

[0130] It should be noted that it is not necessary to determine the configuration of each beam cluster. If there is an abnormal beam cluster, there is no need to reconfigure such a beam cluster. Among them, the historical configuration of the beam cluster can be judged whether it is normal by comparing the traffic prediction value of the beam cluster with the actual traffic value of the beam cluster. If an abnormality is found, such as the configured beam cluster cannot meet the actual traffic demand, the beam cluster is considered abnormal.

[0131] It should be noted that each beam cluster is independent of each other. This is because the similarity between beams has been taken into account when clustering beam clusters. Only beams with very close feature similarity can be aggregated into the same beam cluster. In other words, the similarity between beams from different beam clusters is very low, and it can be considered that different beam clusters are very different and independent of each other. The reason for periodic traffic prediction for beam clusters is that although the beams in a beam cluster have similar features, from the time dimension, the amount of traffic in each time period may be very different, so the corresponding parameters of the beam cluster can be periodically reconfigured to achieve the ultimate energy-saving effect. Of course, if there is no difference in the service traffic between the two time periods, the configuration of the beam cluster can remain unchanged. Figure 7 Schematic diagram showing the configuration of the beam cluster at different times. Figure 7 As shown, at time 1, the configurations of the three beam clusters are configuration 1, configuration 2, and configuration 3, respectively. At time 2, the configurations of the three beam clusters are configuration 4, configuration 2 (remain unchanged), and configuration 5, respectively. That is, at different times, the configuration of the beam cluster can change or remain unchanged. Since the beams in a beam cluster are highly similar, they can use the same prediction model to predict traffic.

[0132] The configuration period of each beam cluster (which is the same as the traffic prediction period of the beam cluster) can be different, and the configuration period of each beam cluster can be variable. Assuming that the minimum configuration time unit is Tk, the actual configuration period of each beam cluster is an integer multiple of Tk. Assuming that there are three beam clusters, the configuration periods of these three beam clusters are n1*Tk, n2*Tk, and n3*Tk, respectively, where n1, n2, and n3 are integers greater than zero. If the beams of beam cluster 1 mainly serve pedestrian users, the traffic flow of pedestrian user services can be predicted for beam cluster 1 to obtain the configuration of beam cluster 1 in different time periods. If the beams of beam clusters 2 and 3 mainly serve vehicle users, beam cluster 2 covers vehicle area 1, and beam cluster 3 covers vehicle area 2, the traffic flow of vehicle area 1 is predicted for beam cluster 2 to obtain the configuration of beam cluster 2 in different time periods. The traffic flow of vehicle area 2 is predicted for beam cluster 3 to obtain the configuration of beam cluster 3 in different time periods.

[0133] by Figure 4 For example, one area is a football field, corresponding to beam cluster 1; one area is a pedestrian walkway, corresponding to beam cluster 2; and one area is a car road, corresponding to beam cluster 3. The football field may only have a large number of users accessing it on weekends; the pedestrian walkway may have people during the day but no one at night; the car road may have vehicles 24 hours a day, and the number of vehicles changes slowly. Then a possible configuration of Tk, n1, n2, and n3 is Tk = 1 hour, n1 = 1, n2 = 1, and n3 = 4, that is, the configuration period of beam clusters 1 and 2 is 1 hour, and the configuration period of beam cluster 3 is 4 hours.

[0134] Specifically, in the prediction model, NGP (multiple kernel Gaussian process) can be used as an example to predict traffic, because Gaussian process prediction can capture the uncertainty and nonlinearity of time series. Figure 8a An example of a multi-kernel Gaussian prediction model according to an exemplary embodiment of the present disclosure is shown. Figure 8aAs shown, NGP supports multiple kernels (also called time period kernels or Gaussian kernels), which can capture multiple periodic characteristics in the data sequence, such as weekly period characteristics and daily period characteristic kernels. In order to predict the traffic of a beam cluster more precisely and better, a finer-grained time period kernel, that is, a Gaussian kernel with a shorter period, can be defined, such as a kernel that captures hourly period characteristics, that is, a period kernel based on hourly period granularity, such as adding 15-minute / 30-minute / 60-minute period kernels. By adding a finer-grained time period kernel, the prediction model's sensitivity to time can be improved, and the traffic changes of the beam cluster can be captured more finely, making the control of the beam more precise and flexible. Respond quickly to traffic changes.

[0135] The prediction step is divided into the following three steps:

[0136] The first step is to calculate the parameters of each kernel (k1, k2, k3, k4, k5) based on the historical data Y(t) of the beam cluster in the database (the beam data within the preset time period of the beam in the beam cluster), where k1 is a weekly kernel (a kernel based on a weekly periodic granularity), which mainly captures the changing pattern of the input sequence in a week; k2 is a daily kernel (a kernel based on a daily periodic granularity), which mainly captures the changing pattern of the input sequence in a day; k3 is an hourly kernel (a kernel based on an hourly periodic granularity), which can be set to 15 minutes, 30 minutes, or 60 minutes, and mainly captures the changing pattern of every quarter hour, half an hour, or 1 hour; k4 is a rational quadratic kernel, which is used to fit dynamic deviations; k5 is a noise kernel, which is used to fit noise.

[0137] In the second step, the final prediction model K is the sum of all the kernels, that is, by solving these 5 kernels, the prediction model K can be obtained.

[0138] The third step is to enter the time t that needs to be predicted i , that is, the beam cluster at time t i The predicted flow value y'(t i ).

[0139] The structure of each Gaussian kernel can be, for example, Figure 8b The structure shown. Figure 8b An example structure of a Gaussian kernel according to an exemplary embodiment of the present disclosure is shown.

[0140] for Figure 8b The five kernels k1, k2, k3, k4, k5 of the given structure shown have the following parameter sets: This set is the parameter set that needs to be solved in the prediction model. This parameter set can be solved by the following formula:

[0141]

[0142] After the parameter set is obtained, they can be substituted into the formula K = k1 + k2 + k3 + k4 + k5, and K is solved. Then, for any given prediction time t i , we can get the predicted value y'(t i ).

[0143] It should be noted that the time series can be the number of access users in the time dimension, the load percentage of the base station, or the throughput of the cell, or the RB utilization rate of the cell. And these indicators are positively correlated, which can reflect the network usage of users in a region and facilitate the base station to make energy-saving decisions.

[0144] In addition, classic time series prediction methods such as the Long Short-Term Memory (LSTM) method and the Auto Regressive model Moving-Average (ARMA) method can also be used to predict the traffic within the beam cluster.

[0145] In an exemplary embodiment of the present disclosure, all beams in the same beam cluster use the same configuration; the configuration of the beam cluster can change over time or remain unchanged, depending on whether the number of users in the beam cluster changes. The exemplary embodiments of the present disclosure can cope with the number of users that changes over time, which can make the network more adaptable in time.

[0146] In step S203, the beams in the at least one beam cluster are configured accordingly according to the determined configuration.

[0147] In an exemplary embodiment of the present disclosure, beams with different beam characteristics are clustered into different beam clusters, and different configuration parameters are configured for different beam clusters, thereby achieving smaller-scale and more accurate regional division of the space within the cell coverage, realizing fine-grained control, and saving more base station energy consumption.

[0148] For the SSB cycle, Figure 4 In the scenario shown, beam cluster 1 mainly serves the access of pedestrian users, and a long-period synchronization signal SSB can be configured; for beam cluster 2 and beam cluster 3, these two beam clusters mainly serve the access of vehicle users, and a short-period synchronization signal SSB can be configured.

[0149] For the switch and direction of the beam cluster, the direction and number of the beam cluster can be adjusted according to the time-varying number of user access in a certain area. If the number of user access is small in a certain period of time, the number of beam clusters in the area can be adjusted to be reduced.

[0150] Regarding the power and bandwidth of the beam cluster, if the number of users accessing the beam cluster is small, the power and / or bandwidth of the beam cluster can be reduced. Conversely, if the number of users accessing the beam cluster is high, the power and / or bandwidth of the beam cluster can be increased.

[0151] In actual scenarios, beam clusters are not static. Due to spatial changes or the occurrence of certain events, the number of user accesses in an area may change dramatically, and this change causes the previously formed beam clusters to no longer be accurate. In order to ensure the accuracy of beam cluster generation and improve the robustness of the present disclosure, the accuracy check step requires periodic comparison of the difference between the traffic prediction value and the actual traffic value of the beam cluster. The accuracy detection period (which can be called T2) is an integer multiple of the minimum configuration period Tk, that is, this step will be triggered every T2. The specific value of period T2 depends on the distribution of the number of user accesses in each beam cluster. If the number of user accesses in a beam cluster changes frequently, the detection period of the accuracy detection is shorter. Otherwise, if the number of user accesses in a beam cluster changes slowly, the detection period of the accuracy detection is longer.

[0152] The embodiment of the present application proposes that the prediction deviation of the beam cluster can be calculated according to the traffic prediction value and the actual traffic value of the beam cluster, and this step can be performed periodically;

[0153] Based on the calculated prediction deviation, determine whether to re-cluster the beams in the beam cluster and the beams associated with the beam cluster, for example, determine whether there is a beam cluster in which the prediction deviation is greater than a preset deviation for a consecutive number of times exceeding a threshold. If the judgment result is yes, it indicates that re-clustering is required, and the beams in the beam cluster and the beams associated with the beam cluster are re-clustered; determine the configuration of the beam cluster obtained by re-clustering; and configure the beams in the beam cluster obtained by re-clustering accordingly according to the determined configuration.

[0154] It should be noted that this judgment of inaccurate beam cluster formation can be a one-time inaccuracy of the prediction value; it can also be multiple consecutive inaccuracies. Because an occasional inaccuracy may be just a small deviation, which is not enough to indicate that the performance of the original beam cluster clustering model (such as the above-mentioned average displacement aggregation method) is low, but multiple consecutive prediction inaccuracies can better indicate that the performance of the original beam cluster clustering model is low, and the current beam cluster is no longer accurate, and the beam cluster needs to be re-formed. When it is detected that a beam cluster has a continuous large prediction deviation (such as a prediction deviation greater than a set threshold for N consecutive times, N is an integer greater than zero), it means that the beam cluster needs to be re-formed. Here, the prediction deviation of a beam cluster refers to the difference between the traffic prediction value and the actual traffic value of the beam cluster. For example, the traffic prediction value at the historical time T and the actual traffic value at that time T can be used to calculate the prediction deviation. Fig.10 Schematic diagram showing the beams used when regenerating the beam cluster. Fig.10 In the case of P=3 consecutive large deviations, the formation of the beam cluster is considered inaccurate and the beam cluster needs to be regenerated. For example, if a large deviation is detected for beam cluster 2, the beam cluster does not need to be reorganized. If three consecutive large deviations are detected for beam cluster 1, the performance of the original beam cluster model is considered to be low and no longer accurate, and the beam cluster needs to be regenerated. The beams used to regenerate the beam cluster include all beams from within the beam cluster as well as beams from other surrounding beam clusters. Fig.10 In FIG. 1 , 18 beams and 3 beam clusters are shown. When it is detected that the prediction value of beam cluster 1 is no longer accurate, it is necessary to re-cluster the beams in beam cluster 1, beams {1, 2, 7, 8, 9, 13, 14} and the beams of beam cluster 2 around them {3, 4, 10, 15, 16}, where the beams of beam cluster 2 can be considered as beams associated with beam cluster 1.

[0155] The beams associated with a certain beam cluster may be beams around the beam cluster, and may include: beams adjacent to beams in the beam cluster, or beams in the beam cluster where beams adjacent to beams in the beam cluster are located.

[0156] In practice, in some scenarios, there is often a surge in the number of user accesses due to emergencies. For example, concerts, football matches, or street performances. Therefore, it is necessary to periodically detect the occurrence of such events and re-form beam clusters based on historical similar events. The period for detecting such events may be T3. The beam data of the beam of the base station collected may also include data for predicting traffic demand change events, which may specifically include at least one of the following: climate data, network application data (such as multimedia social platform data, online news media data, etc.), multimedia data (such as data from public cameras, etc.), traffic data, public utility data, holiday data, etc. Through functions such as keyword search or event detection, these data can be used to predict some unconventional events, such as football matches or concerts that will occur at a certain time and place, which will cause a surge in wireless services.

[0157] The method of detecting such events requires base stations or other servers to regularly collect data from public cameras or network media, and artificial intelligence methods can be used to detect these events. When an emergency event that may cause a sudden change in the number of user access is found, it is necessary to re-form the beam cluster in the area where the event occurred to meet the user access needs when the event occurs.

[0158] Specifically, beam data of a beam of a base station is collected, wherein the beam data at least includes data for predicting traffic demand change events.

[0159] After configuring the beams within the beam cluster, the beam configuration method may further include: predicting a traffic demand change event based on data for predicting a traffic demand change event included in the beam data; when a traffic demand change event is predicted, for at least one beam cluster associated with the predicted traffic demand change event, re-clustering the beams in the at least one beam cluster and the beams associated with the at least one beam cluster; determining the configuration of the re-clustered beam cluster; and correspondingly configuring the beams within the re-clustered beam cluster according to the determined configuration.

[0160] In an exemplary embodiment of the present disclosure, beams with different beam characteristics are clustered into different beam clusters, and different configuration parameters are configured for different beam clusters, thereby achieving a smaller-scale and more accurate regional division of the space within the cell coverage, achieving fine-grained control, and saving more base station energy consumption. In addition, the traffic prediction and configuration of the beam cluster are periodically adaptive to dynamic business changes to achieve accurate energy saving.

[0161] Figure 9a A flowchart of a beam configuration method according to an exemplary embodiment of the present disclosure is shown.

[0162] In step 1, beam data collection is performed.

[0163] The beam data contains data used for communication and data used to predict when traffic demand will change.

[0164] In step 2, beam clusters are formed according to the beam data.

[0165] Beams with similar characteristics are clustered into a beam cluster.

[0166] In step 3, the traffic prediction of the beam cluster and the beam cluster configuration calculation are performed.

[0167] The traffic within the beam cluster is predicted and the corresponding configuration is calculated.

[0168] In step 4, according to the configuration of the beam cluster, corresponding configuration is performed on the beams in the beam cluster.

[0169] Figure 9b Another flowchart of a beam configuration method according to an exemplary embodiment of the present disclosure is shown.

[0170] Step 2 further includes: step 2-1, beam feature extraction (also referred to as beam service feature extraction); step 2-2, beam cluster formation. Among them, beam cluster formation can reduce the complexity of network control.

[0171] Step 3 may be executed periodically, and the execution period may be T1.

[0172] The execution of step 4 refers to performing corresponding configuration on the beams in the beam cluster, so as to adapt to dynamic business changes and accurately achieve base station energy saving.

[0173] In addition, the method further comprises steps 5 and 6, wherein:

[0174] In step 5, the beam cluster prediction accuracy is detected. This step can be performed periodically, and the period can be T2. If there are continuous large prediction deviations, it can go to step 2-1. Otherwise, continue to perform step 5 periodically.

[0175] In step 6, a traffic demand change event prediction is performed. This step can be performed periodically, and the period can be T3. If there is an emergency event of a traffic surge, go to step 2-1, otherwise, continue to perform step 6 periodically.

[0176] The above periods T1, T2, T3 may be the same or different.

[0177] Steps 5 and 6 can improve the robustness of the beam configuration method proposed in this application, so that the beam cluster can adapt to changes in the environment.

[0178] Fig.9c Another flowchart of a beam configuration method according to an exemplary embodiment of the present disclosure is shown.

[0179] In step 1, beam data is collected, wherein the beams of the base station include beam 1, beam 2, ..., beam N.

[0180] In step 2-1, the service characteristics of each beam are extracted according to the historical data of each beam (such as historical data of X days), wherein the service characteristics of the beam include at least one of service flow, service type, and service distribution.

[0181] In step 2-3, beam clusters are formed based on the extracted service characteristics of each beam, for example, beam cluster 1 (the main service users are pedestrians), beam cluster 2 (the main service users are vehicle users, covering vehicle area 1) and beam cluster 3 (the main service users are vehicle users, covering vehicle area 2) are obtained.

[0182] In step 3, traffic prediction is performed for each beam cluster. For example, pedestrian traffic prediction is performed for beam cluster 1, vehicle area 1 traffic prediction is performed for beam cluster 2, and vehicle area 3 traffic prediction is performed for beam cluster 3. Based on the predicted traffic, the beam cluster configuration is calculated to obtain the configuration of each beam cluster in different time periods, such as Fig.9cAs shown, the configuration period of each beam cluster can be different, such as the configuration period of beam cluster 1 is n1*Tk, the configuration period of beam cluster 2 is n2*Tk, and the configuration period of beam cluster 3 is n3*Tk. At the same time, the configuration of each beam cluster can be different, and the configuration of the same beam cluster can be different at different times.

[0183] In step 4, the configuration of each beam cluster is performed. For example, the corresponding configuration for pedestrian services is performed for beam cluster 1, the corresponding configuration for vehicle services in vehicle area 1 is performed for beam cluster 2, and the corresponding configuration for vehicle services in vehicle area 2 is performed for beam cluster 3.

[0184] Figure 9d A flowchart of a beam cluster prediction accuracy checking method according to an exemplary embodiment of the present disclosure is shown.

[0185] The prediction accuracy of the beam cluster is checked according to the period T2. According to the beam data of each beam collected in step 1, the real traffic data of each beam cluster can be obtained. The predicted traffic data of the beam cluster can be obtained through step 3. The real data (real traffic data) and the predicted traffic data are compared to determine whether there are P consecutive large prediction deviations. P can be a preset number threshold. If so, the beam cluster and the beams around it are re-clustered and the process goes to step 2. If not, re-clustering is not performed and the accuracy check is continued periodically. For example, beam cluster 1 has a large deviation in three consecutive periods T2, and P is 3, so it is considered that the beams in beam cluster 1 and the beams around it need to be re-clustered. Beam cluster 2 only has a large deviation in one period, so it is considered that the beams in beam cluster 2 and the beams around it do not need to be re-clustered.

[0186] Fig.9e A flow chart of a method for detecting a flow demand change event according to an exemplary embodiment of the present disclosure is shown.

[0187] Among them, the traffic demand change event detection method can also be called a traffic demand change event prediction method, and can also be called an emergency event detection method or an emergency event prediction method.

[0188] In actual scenarios, there are often some unconventional or non-periodic events that can lead to a surge in the number of users, such as concerts, football matches, large brand exhibitions, or festival performances. Artificial intelligence methods can be used to detect these events. From the data used to predict traffic demand change events in beam data, such as online news, public camera information, or mobile phone application software information, artificial intelligence methods (language detection, keyword detection, image recognition, etc.) are used to identify the location and time of the event.

[0189] In step 6-1, traffic demand change event detection is performed. Based on the data used to predict traffic demand change events in the beam data, language detection, keyword detection, image recognition and other methods can be used to detect whether there is a traffic demand change event, such as a sudden event of a traffic surge. Through 6-1, sudden events can be detected, and the start and end time points of the event can also be detected.

[0190] In step 6-2, when a traffic demand change event is detected, the traffic of the event is predicted based on the data used for communication in the beam cluster, which can also be called event-triggered prediction.

[0191] According to the data used for communication in the beam data of each beam in the beam cluster, the service characteristics of each beam cluster are extracted, including the number of services, service types, service distribution, etc. According to the service characteristics of each beam cluster, the traffic of the event is predicted, which can be called the event trigger prediction value.

[0192] In step 6-3, for each beam cluster, the event trigger prediction value is compared with the actual traffic value (also called actual value) of the beam cluster. When there is a prediction deviation, it is confirmed that an emergency has occurred and the traffic has surged. At this time, beam aggregation can be performed again to obtain a beam cluster.

[0193] Among them, when the difference between the event-triggered predicted value and the actual flow value is greater than the set threshold for more than the set number of times, it is considered that there is a prediction deviation. In other words, the difference between the event-triggered predicted value and the actual flow value is greater than the set threshold for the number of times it occurs can occur once or multiple times in succession.

[0194] For example, the services of beam cluster 1 and beam cluster 2, and compare the difference between the event trigger prediction value and the actual traffic value of beam cluster 1 and beam cluster 2. If they have prediction deviations within their respective beam configuration cycles, it means that an emergency event has begun to occur and the beam cluster needs to be regenerated, that is, repeat steps 1, 2, 3 and 4.

[0195] It should be noted that: usually, the traffic surge starts earlier than the accurate time in the event detection method. Especially for large-scale events and activities, such as a football match starting at 19:00, users may arrive at the scene earlier.

[0196] Figure 9f A flowchart of a beam configuration method according to an exemplary embodiment of the present disclosure is shown.

[0197] Reference Figure 9f In step S901, at least one beam cluster is formed from each beam of the base station.

[0198] In an exemplary embodiment of the present disclosure, before forming at least one beam cluster from each beam of the base station, beam data of the base station may be collected. Here, the beam data may include at least data for communication. Figure 2 The above beam data has been explained in the description and will not be repeated here.

[0199] In an exemplary embodiment of the present disclosure, when beams with similar characteristics are clustered into the same beam cluster according to the beam data, service characteristic information of each beam may be firstly extracted from the data used for communication in the beam data.

[0200] In an exemplary embodiment of the present disclosure, when clustering at least one beam with similar characteristics into a beam cluster according to the extracted service feature information, at least one beam whose feature similarity exceeds a first threshold can be clustered into a beam cluster according to the extracted service feature information.

[0201] In an exemplary embodiment of the present disclosure, when beams with similar characteristics are clustered into the same beam cluster according to the beam data, the beams with similar characteristics may be clustered into the same beam cluster based on a mean shift aggregation method according to the beam data.

[0202] In step S902, the configuration of the at least one beam cluster is determined.

[0203] In an exemplary embodiment of the present disclosure, when determining the configuration of the at least one beam cluster, the traffic prediction value of each beam cluster may be calculated first, and then the configuration of each beam cluster may be determined based on the traffic prediction value of each beam cluster. Figure 2 The configuration of the above beam clusters is explained in the description and will not be repeated here.

[0204] In step S903, the beams in the at least one beam cluster are configured accordingly according to the determined configuration.

[0205] In step S904, the prediction deviation of each beam cluster is calculated according to the traffic prediction value and the actual traffic value of the beam cluster.

[0206] In step S905, it is determined whether there is a beam cluster whose prediction deviation is greater than the preset deviation for a number of consecutive times exceeding a threshold, if yes, step S906 is executed, otherwise, step S905 is continued to be executed.

[0207] In step S906, for a beam cluster whose prediction deviation is greater than a preset deviation for a number of consecutive times exceeding a number threshold, the beams in the beam cluster and the beams associated with the beam cluster are re-clustered. The beams associated with a beam cluster may be beams around the beam cluster, and may include: beams adjacent to the beams in the beam cluster, or beams in the beam cluster where beams adjacent to the beams in the beam cluster are located.

[0208] In step S907, the at least one beam cluster in step S902 is updated to the beam cluster obtained by re-clustering, and the process jumps to step S902 to reconfigure the beams in the beam cluster obtained by re-clustering.

[0209] In step S908, a traffic demand change event is predicted based on the data for predicting a traffic demand change event included in the beam data. Here, the beam data may also include data for predicting a traffic demand change event.

[0210] In an exemplary embodiment of the present disclosure, the data used to predict traffic demand change events may include at least one of climate data, network application data, multimedia data, traffic data, public utility data, and holiday data. The data used to predict traffic demand change events may be collected periodically in real time or may be collected aperiodically.

[0211] In step S909, when a traffic demand change event is predicted, the beam cluster associated with the predicted traffic demand change event and the associated beams are re-clustered.

[0212] In step S910, the at least one beam cluster in step S902 is updated to the beam cluster obtained by re-clustering, and the process jumps to step S902 to reconfigure the beams in the beam cluster obtained by re-clustering.

[0213] In the exemplary embodiments of the present disclosure, in response to the uneven service distribution and time-varying service distribution in the 5G cell, customized variable configuration of the cell is achieved through beam clustering and traffic prediction, thereby achieving a more precise energy control granularity to save more 5G base station energy consumption, which can save about 50% of the average total energy consumption of the base station.

[0214] Fig.11 An example of forming different beam clusters for different service types is shown. Fig.11 As shown, for a certain street model, there are pedestrians on one side and motor vehicles on the other side. The beams covering pedestrian users can be clustered into a beam cluster and configured with a long-period synchronization signal; the beams covering vehicle users can be clustered into a beam family and configured with a short-period synchronization signal.

[0215] The specific steps of forming a beam cluster are:

[0216] The first step is to collect wireless data near the business district (corresponding to the data used for communication mentioned above). These basic wireless data include user location, time, beam service type, beam traffic service, beam synchronization signal period, transmission power, signal quality, beam number, etc.; among them, the measurement indicators of the amount of service can be base station throughput, RB utilization rate, number of access users, etc.; the classification of service types can be based on mobile speed (different levels of speed such as 30 kilometers per hour, 60 kilometers per hour, 120 kilometers per hour, etc.), or QCI, uplink and downlink data ratio, etc.

[0217] In the second step, the average shift method is used to compare the traffic characteristics in each beam, and the beams in the pedestrian area with similar characteristics are clustered into beam cluster 1, and the beams in the vehicle area with similar characteristics are clustered into beam cluster 2.

[0218] The third step is to predict the traffic flow for each beam cluster and calculate the configuration of each beam cluster. Pedestrian beam cluster 1 uses a long-period synchronization signal, and vehicle beam cluster 2 uses a short-period synchronization signal.

[0219] Step 4: Execute the configuration in step 3.

[0220] Fig.12 An example of forming beam clusters for a time-varying number of user accesses in a certain area is shown.

[0221] like Fig.12 As shown, for a time-varying number of user accesses in a certain area (eg, a park area), if the number of user accesses is small within a certain time period, the direction of the beam can be adjusted and the number of beams can be reduced.

[0222] The specific steps of forming a beam cluster are:

[0223] The first step is to collect wireless data near the park area. These basic wireless data include user location, time, beam service type, beam traffic service volume, beam synchronization signal period, transmission power, signal quality, beam number, etc. Among them, the measurement indicators of the volume of services can be base station throughput, RB utilization rate, number of access users, etc.; the classification of service types can be based on mobile speed (30 kilometers per hour, 60 kilometers per hour, 120 kilometers per hour, etc.), or QCI, uplink and downlink data ratio, etc.

[0224] The second step is to use the GPS image recognition method to compare the traffic characteristics in each beam. For example, for beams in different areas, they have the same traffic characteristics within a day, with a flow density of 0.3 at 7 am, 0.8 at 9 am, 0.5 at 5 pm, and 0 at 9 pm, so they are clustered into a beam cluster.

[0225] The third step is to predict the number of users accessing the beam cluster and calculate the corresponding configuration. For example, at 7 a.m., the density of people is 0.3, and the direction, power and switch of the beam cluster are adjusted to open 30% of the beams to cover the area; at 9 a.m., the density of people is 0.8, and the direction, power and switch of the beam cluster are adjusted to open 100% of the beams to cover the area; at 5 p.m., it is 0.5, and the direction, power and switch of the beam are adjusted to open 60% of the beams to cover the area; at 9 p.m., it is 0, and all beams are closed.

[0226] Step 4: Execute the configuration in step 3.

[0227] Fig.13 An example of adaptively changing the beam cluster in a scenario where users are unevenly distributed is shown, and the shape of the beam cluster changes according to the change of RB usage rate.

[0228] like Fig.13 As shown, in a scenario where users are unevenly distributed, for example, if the number of users in beam cluster A and beam cluster B changes, coordination between beam clusters can be achieved by re-clustering the beam clusters.

[0229] The specific steps of forming a beam cluster are:

[0230] The first step is to collect wireless data of a certain square. These basic wireless data include user location, time, beam service type, beam traffic service, beam synchronization signal period, transmission power, signal quality, beam number, etc.

[0231] In the second step, the text recognition method of the adjacent beam switching table is used to compare the traffic characteristics in each beam and cluster them into beam cluster A and beam cluster B respectively.

[0232] The third step is to predict the load of each beam cluster and calculate the configuration of each beam cluster.

[0233] Step 4: Execute the configuration in step 3.

[0234] Step 5: As time changes, user density also changes. After multiple consecutive errors, it is necessary to readjust the formation of the beam clusters, i.e., beam cluster A' and beam cluster B', and recalculate the configuration of the beam clusters.

[0235] Step 6. Execute the new configuration in step 5.

[0236] Fig.14 An example of forming a new beam cluster when a traffic demand change event is detected is shown.

[0237] like Fig.14 As shown, there may be festival activities or other celebrations on the streets or communities from time to time, which will gather many mobile users. In this case, it is necessary to trigger the event detection step, re-cluster the beam clusters, and recalculate the configuration of the beam clusters, so as to meet the surging traffic demand.

[0238] The specific steps of forming a beam cluster are:

[0239] The first step is to collect wireless data near residential areas. These basic wireless data include user location, time, beam service type, beam traffic service volume, beam synchronization signal period, transmission power, signal quality, beam number, etc.

[0240] In the second step, the average shift method is used to compare the RB usage characteristics in each beam and cluster the beam clusters.

[0241] The third step is to predict the traffic flow of each beam cluster and calculate the configuration of each beam cluster.

[0242] Step 4: Execute the configuration in step 3.

[0243] The fifth step is to predict that there will be a lion dance performance in the afternoon by searching keywords on the social platform.

[0244] Step 6: As the lion dance performance approaches, the predicted values ​​and actual values ​​of the existing beam clusters become increasingly inaccurate, so the beam clusters need to be readjusted to form new beam clusters.

[0245] The seventh step is to predict the resource block (RB) utilization rate within the beam cluster and perform configuration calculation.

[0246] Step 8. Execute the configuration in step 7.

[0247] In the ninth step, when the lion dance performance is about to end, the predicted value and actual value of the resource block RB utilization rate of the existing beam cluster are becoming increasingly inaccurate, so the beam cluster needs to be readjusted to form a new beam cluster.

[0248] The tenth step is to predict the traffic within the beam cluster and perform configuration calculations.

[0249] Step 11: Execute the configuration in step 10.

[0250] Fig.15 The time flow curves for four land use types that are typical in cities are shown. Fig.15 There are significant differences in the business volumes of office areas, public facilities areas, residential areas and green areas at different times of the day. Fig.16 An example of energy saving effect based on the present disclosure is shown. Fig.16 As shown, the electricity saved in office area, public facilities area, residential area and green area is 26%, 31%, 71% and 74% respectively. Fig.17 The following table shows the occupancy ratios of office area, public facilities area, residential area and green area (such as park) in the example city land planning. Fig.17 In the figure, the occupancy ratios of office area, public facilities area, residential area and green area are 25%, 9%, 22% and 44% respectively. Therefore, the average total energy saving Es can be calculated according to the following formula.

[0251] Es≈25%*26%+9%*31%+22%*71%+44%*74%=57.38%.

[0252] In summary, the average total energy saving based on the present disclosure is 57.38%.

[0253] The above has been combined Figures 2 to 17 The beam configuration method according to the exemplary embodiment of the present disclosure is described. Fig.18 and Fig.19 A beam configuration device and units thereof according to an exemplary embodiment of the present disclosure are described.

[0254] In the embodiment of the present application, the execution entity of the beam configuration method may be a base station or a server.

[0255] Fig.18 A block diagram of a beam configuration device according to an exemplary embodiment of the present disclosure is shown.

[0256] Reference Fig.18 The beam configuration device includes a beam cluster forming unit 181, a configuration determining unit 182 and a beam configuration unit 183.

[0257] The beam cluster forming unit 181 is configured to form at least one beam cluster from the respective beams of the base station.

[0258] In an exemplary embodiment of the present disclosure, the beam configuration device may further include: a beam data acquisition unit (not shown), configured to collect beam data of the beam of the base station. Here, the beam data may include at least data used for communication. Here, the data used for communication may include at least one of user location information, current time information, beam service type information, beam service flow information, service distribution information, beam synchronization signal period information, transmission power information, signal quality information, and beam number information. Here, the service type information includes at least one of the statistical value of the user's moving speed, the signal's service quality level identifier, and the uplink / downlink data ratio; the service flow includes at least one of the throughput, resource block utilization rate, the number of IP packets, and the number of access users. In an exemplary embodiment of the present disclosure, the beam cluster forming unit 181 may be configured to form at least one beam cluster from each beam of the base station according to the beam data.

[0259] In an exemplary embodiment of the present disclosure, the beam cluster forming unit 181 may be configured to cluster beams having similar characteristics into the same beam cluster according to the beam data.

[0260] In an exemplary embodiment of the present disclosure, the beam cluster forming unit 181 can also be configured to: extract service characteristic information of each beam from the data used for communication in the beam data, the service characteristic information including at least one of service traffic information, service type information, and service distribution information; and cluster at least one beam with similar characteristics into a beam cluster according to the extracted service characteristic information.

[0261] In an exemplary embodiment of the present disclosure, the beam cluster forming unit 181 may also be configured to: cluster at least one beam whose characteristic similarity exceeds a first threshold into a beam cluster according to the extracted service characteristic information.

[0262] In an exemplary embodiment of the present disclosure, the beam cluster forming unit 181 may be further configured to: cluster beams having similar characteristics into the same beam cluster based on the mean shift aggregation method according to the beam data.

[0263] In an exemplary embodiment of the present disclosure, the beam cluster forming unit 181 may also be configured to: first obtain the sample points corresponding to the beam data of each beam of the base station, then randomly select a sample point from all the sample points as the center, find all the sample points within a preset radius of the center, calculate the average displacement vector of the displacement vectors from all the sample points within the preset radius of the center to the center, cluster the beams corresponding to all the sample points within the preset radius according to the average displacement vector, and when there are unclustered beams, take all the sample points of the unclustered beams as all the sample points, and jump to the step of randomly selecting a sample point from all the sample points as the center. The above has been discussed in detail with regard to Figure 2 The above sample points and radii are explained in the description of , so they will not be repeated here.

[0264] In an exemplary embodiment of the present disclosure, the beam cluster forming unit 181 may be configured to: first obtain a service intensity heat map corresponding to the base station, and then cluster at least one beam of the base station into one beam based on the color similarity in the service intensity heat map. Figure 2 The above business intensity heat map and business intensity are explained in the description of , so they are not repeated here.

[0265] In an exemplary embodiment of the present disclosure, the beam cluster forming unit 181 may be configured to: determine the switching correlation between the beams according to the beam switching relationship corresponding to the base station; and cluster at least one beam of the base station into a beam cluster based on the switching correlation between the beams.

[0266] The configuration determining unit 182 is configured to determine the configuration of the at least one beam cluster.

[0267] In an exemplary embodiment of the present disclosure, the configuration determination unit 182 may be configured to: calculate a traffic prediction value of a beam cluster; and determine a configuration of the beam cluster based on the traffic prediction value of the beam cluster, wherein the configuration of the beam cluster includes at least one of a period of a synchronization signal SSB, a power level, a beam cluster bandwidth, a direction of the beam cluster, and a switch of the beam cluster.

[0268] In an exemplary embodiment of the present disclosure, the configuration determination unit 182 can also be configured to: determine the parameters of each time period kernel in the prediction model based on the data used for communication in the beam data of each beam in the beam cluster, and obtain a prediction model based on a time series, wherein the prediction model includes a periodic kernel based on an hour as a periodic granularity; use the prediction model based on a time series to determine the traffic prediction value of each beam cluster at a predetermined time.

[0269] The beam configuration unit 183 is configured to configure the beams within the at least one beam cluster according to the determined configuration.

[0270] Fig.19 A block diagram of a beam configuration device according to an exemplary embodiment of the present disclosure is shown.

[0271] Reference Fig.19 The beam configuration device includes a beam cluster forming unit 191, a configuration determination unit 192, a beam configuration unit 193, a deviation calculation unit 194, a first re-clustering unit 195, an event prediction unit 196, a second re-clustering unit 197, a configuration re-determination unit 198 and a reconfiguration unit 199.

[0272] The beam cluster forming unit 191 is configured to form at least one beam cluster from the respective beams of the base station.

[0273] The configuration determining unit 192 is configured to determine the configuration of the at least one beam cluster.

[0274] The beam configuration unit 193 is configured to configure the beams within the at least one beam cluster according to the determined configuration.

[0275] The deviation calculation unit 194 is configured to calculate the prediction deviation of the beam cluster according to the traffic prediction value and the actual traffic value of the beam cluster.

[0276] The first re-clustering unit 195 is configured to re-cluster the beams in the beam cluster and the beams associated with the beam cluster for a beam cluster whose prediction deviation is greater than a preset deviation or the number of consecutive times that the prediction deviation is greater than the preset deviation exceeds a number threshold.

[0277] The event prediction unit 196 is configured to predict a traffic demand change event according to data for predicting a traffic demand change event included in the beam data. Here, the beam data may also include data for predicting a traffic demand change event.

[0278] In an exemplary embodiment of the present disclosure, the data used to predict traffic demand change events may include at least one of climate data, network application data, multimedia data, traffic data, public utility data, and holiday data. The data used to predict traffic demand change events may be collected periodically in real time or may be collected aperiodically.

[0279] The second re-clustering unit 197 is configured to, when a traffic demand change event is predicted, re-cluster the beams in the at least one beam cluster and the beams associated with the at least one beam cluster associated with the preset traffic demand change event.

[0280] The configuration re-determining unit 198 is configured to determine the configuration of the re-clustered beam clusters.

[0281] The reconfiguration unit 199 is configured to configure the beams in the re-clustered beam clusters accordingly according to the determined configuration.

[0282] In addition, according to an exemplary embodiment of the present disclosure, a computer-readable storage medium is also provided, on which a computer program is stored. When the computer program is executed, the beam configuration method according to the exemplary embodiment of the present disclosure is implemented.

[0283] In an exemplary embodiment of the present disclosure, the computer-readable storage medium may carry one or more programs, and when the computer program is executed, the following steps can be implemented: determining the configuration of the beam cluster in the at least one beam cluster; and configuring the beams within the beam cluster according to the determined configuration of the beam cluster.

[0284] The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In an embodiment of the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a computer program that may be used by or in conjunction with an instruction execution system, device or device. The computer program contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination thereof. The computer-readable storage medium may be contained in any device; it may also exist alone without being assembled into the device.

[0285] The above has been combined Figure 18 to Figure 19 The beam configuration device according to the exemplary embodiment of the present disclosure is described. Fig. 20 An electronic device according to an exemplary embodiment of the present disclosure is described.

[0286] Fig. 20 A schematic diagram showing an electronic device according to an exemplary embodiment of the present disclosure.

[0287] Reference Fig. 20 According to an exemplary embodiment of the present disclosure, the electronic device 20 includes a memory 191 and a processor 202, wherein the memory 201 stores a computer program. When the computer program is executed by the processor 202, the beam configuration method according to the exemplary embodiment of the present disclosure is implemented.

[0288] In an exemplary embodiment of the present disclosure, when the computer program is executed by the processor 202, the following steps can be implemented: forming at least one beam cluster from the various beams of the base station; determining the configuration of the at least one beam cluster; and configuring the beams within the at least one beam cluster accordingly according to the determined configuration.

[0289] Fig. 20The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0290] The above has been referred to Figures 2 to 20 The beam configuration method and apparatus, electronic device and computer storage medium according to exemplary embodiments of the present disclosure are described. However, it should be understood that: Fig.18 and Fig.19 The beam configuration device and its units shown in the figure may be respectively configured as software, hardware, firmware or any combination of the above items to perform specific functions. Fig. 20 The electronic device shown in is not limited to including the components shown above, but some components may be added or deleted as needed, and the above components may also be combined.

[0291] According to the beam configuration method and apparatus, electronic device and computer storage medium of the exemplary embodiments of the present disclosure, at least one beam cluster is first formed from the various beams of the base station, and then the configuration of the at least one beam cluster is determined. Finally, the beams in the at least one beam cluster are configured accordingly according to the determined configuration, so that beams with different beam characteristics are clustered into different beam clusters, and different configuration parameters are configured for different beam clusters, thereby achieving a smaller-scale and more accurate regional division of the space within the cell coverage, realizing fine-grained control, and saving the energy consumption of the base station.

[0292] While the present disclosure has been particularly shown and described with reference to exemplary embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the present disclosure as defined by the following claims.

Claims

1. A beam configuration method in a wireless communication system, comprising: Obtaining beam data of multiple beams of a base station; extracting service characteristic information about each of the plurality of beams from the beam data; Based on the service characteristic information of the beam data, clustering at least two beams having similar characteristics from the multiple beams into beam clusters to form at least one beam cluster; determining a configuration of the at least one beam cluster; The plurality of beams in the at least one beam cluster are configured according to the determined configuration.

2. The beam configuration method according to claim 1, wherein: The beam data includes at least one of the following: User location information, current time information, service type information about at least one beam, service traffic information about the at least one beam, service distribution information about the at least one beam, synchronization signal period information about the at least one beam, transmission power information about the at least one beam, signal quality information, and beam number information.

3. The beam configuration method according to claim 2, wherein: The service type information includes at least one of the following: a statistical value of the user's moving speed, a signal's service quality level identifier, and an uplink / downlink data ratio. The service traffic includes at least one of the following: throughput, resource block utilization rate, amount of Internet Protocol IP data packets, and number of access users.

4. The beam configuration method according to claim 1, wherein: The service characteristic information includes at least one of service flow information, service type information, and service distribution information.

5. The beam configuration method according to claim 1, wherein: The step of clustering at least two beams having similar characteristics from the multiple beams into a beam cluster based on the service characteristic information of the beam data comprises: According to the service characteristic information, based on determining that a characteristic similarity between at least two beams having similar characteristics exceeds a threshold, the at least two beams are clustered into a beam cluster.

6. The beam configuration method according to claim 1, further comprising: Based on the beam data, at least one beam having similar characteristics among the plurality of beams is clustered into a beam cluster based on mean shift aggregation.

7. The beam configuration method according to claim 6, wherein: The step of clustering at least one beam having similar characteristics among the plurality of beams into a beam cluster by average shift aggregation according to the beam data comprises: Acquire sample points corresponding to the beam data; Randomly select a sample point from the sample points as the center, and determine other sample points within a preset radius of the center; Calculate the average displacement vector of the displacement vectors from other sample points within the preset radius to the center; Clustering other beams corresponding to the other sample points according to the average displacement vector; When there are unclustered beams, selection, calculation, and clustering are repeated by using the unclustered beams as sample points.

8. The beam configuration method according to claim 1, wherein: The step of forming at least one beam cluster from the plurality of beams comprises: Obtaining a service intensity heat map corresponding to the base station; Based on the color similarity in the service intensity heat map, at least two beams of the base station are clustered into the beam cluster.

9. The beam configuration method according to claim 1, wherein: The step of forming at least one beam cluster from the plurality of beams comprises: Determining a switching correlation between the multiple beams according to a beam switching relationship corresponding to the base station; At least one beam of the base station is clustered into one beam cluster among the at least one beam cluster based on the switching correlation between the multiple beams.

10. The beam configuration method according to claim 1, wherein: The step of determining the configuration of the at least one beam cluster comprises: Calculating a traffic prediction value of the at least one beam cluster; and determining a configuration of the at least one beam cluster based on a traffic prediction value of the at least one beam cluster, The configuration of the at least one beam cluster includes at least one of the period of the synchronization signal SSB, the power size, the beam cluster bandwidth, the direction of the at least one beam cluster, and the switching state of the at least one beam cluster.

11. The beam configuration method according to claim 10, wherein: The steps of calculating the traffic prediction value of the beam cluster include: Determine the parameters of each time period kernel in the prediction model according to the communication data in the beam data of each beam in the at least one beam cluster, and obtain the prediction model based on the time series, wherein the prediction model includes a period kernel based on the hour as the period granularity; The prediction model based on the time series is used to determine a traffic prediction value of the at least one beam cluster at a predetermined time.

12. The beam configuration method according to claim 1, after configuring the multiple beams in the at least one beam cluster, further comprising: Calculating a prediction deviation of the at least one beam cluster according to a flow prediction value and an actual flow value of the at least one beam cluster; Based on the calculated prediction deviation, determining whether to re-cluster the plurality of beams and other beams associated with the at least one beam cluster; Determining a re-clustering configuration of the re-clustered beam clusters obtained by the re-clustering; The multiple beams in the re-clustered beam cluster and the other beams are configured according to the determined re-clustering configuration.

13. The beam configuration method according to claim 1, wherein: The beam data at least includes data for predicting traffic demand change events, Wherein, after configuring the multiple beams in the at least one beam cluster, the beam configuration method further includes: predicting a traffic demand change event according to data for predicting a traffic demand change event included in the beam data; When a traffic demand change event is predicted, based on determining that the at least one beam cluster is associated with the predicted traffic demand change event, re-clustering the plurality of beams in the at least one beam cluster and other beams associated with the at least one beam cluster; Determining a re-clustering configuration of the re-clustered beam clusters obtained by the re-clustering; The multiple beams in the re-clustered beam cluster and the other beams are configured according to the determined re-clustering configuration.

14. The beam configuration method according to claim 13, wherein: The data used to predict traffic demand change events include at least one of the following: climate data, network application data, multimedia data, traffic data, public utility data, and holiday data.

15. A beam configuration device in a wireless communication system, comprising: A beam cluster forming unit, configured to obtain beam data of multiple beams of a base station; Extracting service characteristic information about each of the multiple beams from the beam data; clustering at least two beams having similar characteristics from the multiple beams into beam clusters based on the service characteristic information of the beam data to form at least one beam cluster; a configuration determining unit, configured to determine a configuration of the at least one beam cluster; and A beam configuration unit is configured to configure the multiple beams in the at least one beam cluster according to the determined configuration.

16. A non-transitory computer-readable storage medium, wherein: The non-transitory computer-readable storage medium is configured to store instructions, which, when executed by a processor, implement the beam configuration method according to any one of claims 1 to 14.

17. An electronic device in a wireless communication system, comprising: processor; A memory storing a computer program, wherein when the computer program is executed by a processor, the beam configuration method according to any one of claims 1 to 14 is implemented.

Images