AI-based base station power consumption intelligent regulation and control method, apparatus and device, and medium

Abstract

The invention relates to an AI-based base station power consumption intelligent regulation and control method and device, equipment and a medium. The method comprises the following steps of: constructing a base station load adjustment graph convolutional network by taking each base station as a node and taking user load capacity in an overlapping range of a maximum adjustment area of an antenna unit of each base station in a connecting line direction among the base stations as an edge; acquiring real-time load state information of each base station; inputting the real-time load state information into a base station load adjustment graph convolutional network, and generating load distribution information of each base station; and generating antenna unit power consumption adjustment information based on the load distribution information of each base station, the antenna unit power consumption adjustment information being used for adjusting the power consumption of each antenna unit of each base station. By adopting the method, the interpretability of the base station power consumption intelligent regulation and control model can be improved, the intelligent management of the base station power consumption is realized, the network service quality is ensured, and the dynamic adaptive capacity of the base station power consumption intelligent regulation and control is enhanced.

Description

Technical Field

[0001] The present invention belongs to the field of base station power management, and in particular relates to an AI-based intelligent control method, device, equipment and medium for base station power consumption. Background Art

[0002] With the rapid development of communications technology, mobile data traffic has exploded. The widespread adoption of next-generation communications technologies like 5G has made base station energy consumption an increasingly prominent issue. Traditional base station energy conservation efforts rely primarily on methods such as switching power supplies, air conditioning, carrier frequency, and time-based device shutdown. While these methods can reduce energy consumption to a certain extent, they also have significant limitations, such as weakening wireless signal strength in the base station's coverage area, impacting user experience, and making it difficult to strike a balance between energy conservation and ensuring communication quality.

[0003] In recent years, artificial intelligence (AI) technology has been gradually introduced into the field of base station energy conservation, demonstrating enormous potential. By analyzing and learning from base station operating data through AI technologies such as machine learning and deep learning, intelligent regulation of base station power consumption can be achieved. However, existing AI technologies for base station energy conservation still suffer from limitations such as high data requirements, poor model interpretability, and limited adaptability to dynamic environments. Summary of the Invention

[0004] Based on this, it is necessary to provide an AI-based base station power consumption intelligent control method, device, equipment and medium to address the above technical problems, which can improve the model interpretability, realize intelligent management of base station power consumption, ensure network service quality and enhance the model's ability to adapt to dynamic environments.

[0005] In a first aspect, the present application provides an AI-based intelligent control method for base station power consumption, comprising:

[0006] With each base station as a node and the user load within the overlapping range of the maximum adjustment area of ​​the antenna units of each base station in the direction of the connection between the base stations as an edge, a base station load adjustment graph convolutional network is constructed. The base station load adjustment graph convolutional network is used to adjust the load distribution of each base station according to the real-time load status of each base station;

[0007] Obtain real-time load status information of each base station;

[0008] Input the real-time load status information into the base station load adjustment graph convolutional network to generate the load distribution information of each base station;

[0009] Based on the load distribution information of each base station, antenna unit power consumption adjustment information is generated, and the antenna unit power consumption adjustment information is used to adjust the power consumption of each antenna unit of each base station.

[0010] In one of the embodiments, the antenna unit power consumption adjustment information includes antenna unit channel enabled number information and antenna unit channel operating frequency information, the load distribution information includes antenna unit load distribution information of each antenna unit, the antenna unit power consumption adjustment information is generated based on the load distribution information of each base station, including:

[0011] The channel allocation base and the channel rated load parameter of each antenna unit are obtained, and the channel allocation base is used to represent the minimum allocated channel number of each antenna unit;

[0012] The antenna unit channel enabled number information is generated based on the channel allocation base, the channel rated load parameter and the antenna unit load distribution information;

[0013] The channel load rate is calculated in combination with the antenna unit channel enabled number information, the channel rated load parameter and the antenna unit load distribution information;

[0014] The antenna unit channel operating frequency information is distributed according to the channel load rate;

[0015] The expression of the antenna unit channel operating frequency information is:

[0016]

[0017] {i|i∈(N0,N max )∩Z}

[0018] In the formula, f CW is the antenna unit channel operating frequency information, is the antenna unit channel minimum operating frequency, is the channel load rate threshold corresponding to the (N0+1)th antenna unit channel operating frequency, N max is the highest number of antenna unit channel operating frequencies, f i is the i-th antenna unit channel operating frequency, μ i is the channel load rate threshold corresponding to the i-th antenna unit channel operating frequency, is the antenna unit channel maximum operating frequency, is the channel load rate threshold corresponding to the antenna unit channel maximum operating frequency, R CL is the channel load rate, I AULD is the antenna unit load distribution information, N AUCE is the antenna unit channel enabled number information, P CRL is the channel rated load parameter, and Z is an integer set symbol.

[0019] In one of the embodiments, the AI-based base station power consumption intelligent regulation method further includes:

[0020] The loss value of each channel of each antenna unit is obtained.

[0021] Generate channel activation selection information in descending order of the loss values ​​of the channels and the number of channels activated in the antenna unit. The channel activation selection information is used to represent the channels activated in the antenna unit.

[0022] The loss value of the enabled channel is updated based on the antenna unit channel operating frequency information and the enabled duration of the enabled channel.

[0023] In one embodiment, the AI-based intelligent control method for base station power consumption further includes:

[0024] Identify the antenna unit whose channel activation quantity information is equal to the channel allocation base as a candidate dormant antenna unit;

[0025] Identify the candidate dormant antenna unit whose antenna unit channel operating frequency information is equal to the lowest operating frequency of the antenna unit channel as the antenna unit to be dormant;

[0026] Control the antenna unit to be dormant to enter the dormant state.

[0027] In one embodiment, the node characteristics of the node include connected user load characteristics, antenna unit direction characteristics, antenna unit rated load characteristics, and base station location characteristics. The node characteristic matrix of the node is expressed as follows:

[0028]

[0029] {κ|κ∈(1,M)∩Z}

[0030] Where, is the node feature matrix, L κ is the load characteristic of connected users of the κth base station, is the directional characteristic of the antenna unit of the κ-th base station, is the rated load characteristic of the antenna unit of the κth base station, is the base station location feature of the κth base station, M is the total number of base stations, is the antenna unit directional characteristic of the λth antenna unit of the κth base station, C λ is the antenna unit rated load characteristic of the λth antenna unit of the κth base station, N κ is the total number of antenna units of the κth base station.

[0031] In one embodiment, the AI-based intelligent control method for base station power consumption further includes:

[0032] Obtaining a predicted overlapping user load corresponding to the overlapping user load of each base station and a predicted connected user load corresponding to the connected user load characteristics;

[0033] inputting the predicted overlapping user load amount and the predicted connected user load amount into a base station load adjustment graph convolution network to generate load allocation prediction information of each base station;

[0034] based on the load allocation prediction information of each base station, antenna unit load prediction information is generated;

[0035] According to the antenna unit load prediction information, antenna unit power consumption prediction adjustment information of the antenna unit of each base station is generated, which is used to represent the antenna unit channel enabling quantity information and the antenna unit channel working frequency information of each antenna unit of each base station in the prediction period.

[0036] In one of the embodiments, the predicted overlapping user load amount corresponding to the overlapping user load amount of each base station and the predicted connected user load amount corresponding to the connected user load amount characteristic are obtained, including:

[0037] The historical load state information data of each base station is obtained, including the historical overlapping user load amount corresponding to the overlapping user load amount and the historical connected user load amount corresponding to the connected user load amount characteristic;

[0038] Based on the historical load state information data, an inter-day long short-term memory network model, an inter-week long short-term memory network model and a holiday long short-term memory network model are constructed respectively;

[0039] According to the inter-day long short-term memory network model, the inter-week long short-term memory network model and the holiday long short-term memory network model, the inter-day predicted overlapping user load amount, the inter-week predicted overlapping user load amount and the holiday predicted overlapping user load amount corresponding to the overlapping user load amount of each base station, and the inter-day predicted connected user load amount, the inter-week predicted overlapping user load amount and the holiday predicted overlapping user load amount corresponding to the connected user load amount characteristic are generated;

[0040] The predicted overlapping user load amount is obtained by combining the inter-day predicted overlapping user load amount, the inter-week predicted overlapping user load amount and the holiday predicted overlapping user load amount, and the predicted overlapping user load amount is obtained by combining the inter-day predicted connected user load amount, the inter-week predicted overlapping user load amount and the holiday predicted overlapping user load amount.

[0041] In a second aspect, the present application also provides an AI-based base station power consumption intelligent regulation device, comprising:

[0042] The load adjustment model construction module is used to construct a base station load adjustment graph convolution network with each base station as a node and the user load amount in the overlapping range of the maximum adjustment region of the antenna unit of each base station in the connection direction between the base stations as an edge, and the base station load adjustment graph convolution network is used to adjust the load allocation of each base station according to the real-time load state of each base station;

[0043] a real-time load state acquisition module, configured to acquire real-time load state information of each base station;

[0044] a load distribution information generation module, configured to input the real-time load state information into a base station load regulation graph convolution network, and generate load distribution information of each base station;

[0045] an antenna unit power consumption adjustment module, configured to generate antenna unit power consumption adjustment information based on the load distribution information of each base station, and the antenna unit power consumption adjustment information is used to adjust power consumption of each antenna unit of each base station.

[0046] In a third aspect, the present application also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method according to any one of the first aspect of the present application when executing the computer program.

[0047] In a fourth aspect, the present application also provides a computer readable storage medium, which stores a computer program, and the computer program implements the steps of the method according to any one of the first aspect of the present application when executed by a processor.

[0048] The AI-based base station power consumption intelligent regulation method, device, equipment and medium described above can effectively model the spatial correlation and load dependent relationship between base stations by abstracting the base stations as graph nodes and the user load of the antenna unit overlapping area as edges, improve the interpretability of the base station model, and solve the problem of insufficient synergy caused by isolated regulation of base station power consumption in traditional methods, thereby improving the accuracy of multi-base station collaborative regulation. By generating load distribution information of each base station, the network traffic fluctuations can be responded in real time, and intelligent balancing of load between base stations can be achieved, thereby optimizing the network resource utilization efficiency. By generating antenna unit power consumption adjustment strategies based on load distribution information, the network layer load state can be directly associated with the physical layer antenna power consumption, and the redundant power consumption can be dynamically reduced under the premise of ensuring the coverage quality, thereby achieving balanced optimization of energy consumption and performance. By using the joint feature extraction capability of the graph convolution network on node attributes and edge weights, the limitations of traditional regulation based on single base station data can be broken through, and multi-dimensional coupling relationships in complex network environments can be captured, thereby ensuring that the regulation strategy is globally optimal. BRIEF DESCRIPTION OF DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related art, the drawings needed in the embodiment or related art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0050] Figure 1 A flowchart of an AI-based intelligent control method for base station power consumption provided in one embodiment of the present application;

[0051] Figure 2 A schematic diagram of a process for generating antenna unit power consumption adjustment information provided in one embodiment of the present application;

[0052] Figure 3 A flowchart of another AI-based method for intelligently controlling base station power consumption provided in one embodiment of the present application;

[0053] Figure 4 A schematic structural diagram of a base station power consumption intelligent control device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0055] In one embodiment, Figure 1 As shown, a method for intelligent control of base station power consumption based on AI is provided. This embodiment takes the application of this method to a server as an example. It can be understood that this method can also be applied to a base station terminal, and can also be applied to a system including a base station terminal and a server, and is implemented through the interaction between the base station terminal and the server. The constituent devices of the base station terminal may include, but are not limited to, an active antenna unit (AAU), a baseband processing unit (BBU), a remote radio unit (RRU), a power supply device, and a heat dissipation device. The antenna unit in the base station involved in this application may be an active antenna unit, or it may be other antenna units that can transmit and receive wireless signals to realize communication between the base station and the user equipment, which is not limited here. In this embodiment, the method includes the following steps:

[0056] Step S101 , constructing a base station load adjustment graph convolutional network with each base station as a node and the user load within the overlapping range of the maximum adjustment area of ​​the antenna unit of each base station in the connection direction between the base stations as an edge.

[0057] Optionally, the server can map each base station terminal to a node in the base station load regulation graph convolutional network with the load characteristics and spatial distribution characteristics of each base station terminal as the feature attributes of the node. The server can construct the base station load regulation graph convolutional network with each base station terminal as a node, and the user load amount in the overlapping range of the maximum regulation area of the antenna unit of each base station terminal in the connection direction between each base station terminal as an edge. M The difference set between the maximum working range G Max and the minimum working range G Min , that is: The base station load regulation graph convolutional network can be loaded on the server.

[0058] Illustratively, the base station load regulation graph convolutional network loaded on the server is used to adjust the load distribution of each base station terminal according to the real-time load state of each base station terminal.

[0059] Illustratively, the server can calculate the coverage overlap area between adjacent base station terminals based on the geometric overlap detection algorithm based on the circular area formula. The server can calculate the user load amount in the overlapping range of the maximum regulation area of the antenna unit of each base station terminal in the connection direction between each base station terminal based on the coverage overlap area between adjacent base station terminals and the user load information in the coverage overlap area between adjacent base station terminals, and use the user load amount in the overlapping range of the maximum regulation area of the antenna unit of each base station terminal in the connection direction between each base station terminal as the edge weight of the base station load regulation graph convolutional network.

[0060] Step S102, obtaining the real-time load state information of each base station.

[0061] Optionally, the real-time load state information of each base station terminal can include but is not limited to user load amount information in the overlapping range and base station user load total amount information.

[0062] Step S103, inputting the real-time load state information into the base station load regulation graph convolutional network to generate the load distribution information of each base station.

[0063] Specifically, the server can input the obtained real-time load state information of each base station terminal into the base station load regulation graph convolutional network loaded on the server to generate the load distribution information of each base station terminal.

[0064] Step S104, generating antenna unit power consumption regulation information based on the load distribution information of each base station.

[0065] Specifically, the antenna unit power consumption regulation information can be used to adjust the power consumption of each antenna unit of each base station.

[0066] In the above-mentioned AI-based intelligent control method for base station power consumption, the load status of each base station can be tracked in real time through a refined load monitoring and distribution mechanism; by formulating a load optimization plan for each base station based on a graph convolutional network, it can effectively solve the problems of slow response and insufficient accuracy of traditional load balancing technology, thereby further realizing efficient utilization of network resources and effective control of energy consumption; by integrating graph convolutional networks with base station power control technology, it can realize power consumption management at the base station antenna unit level; by generating differentiated power consumption adjustment instructions according to the actual load conditions of the antenna unit, the report effectively solves the problem of extensive power consumption adjustment of the base station, improves the adaptability of the base station to complex scenarios, and further enhances the ability of the base station to cope with diversified communication needs, thereby realizing multi-objective optimization of communication quality and energy consumption control.

[0067] In an optional embodiment of the present application, the antenna unit power consumption adjustment information may include the number of antenna unit channel activation information and the antenna unit channel operating frequency information, and the load distribution information may include the antenna unit load distribution information of each antenna unit. This embodiment is described by applying the AI-based base station power consumption intelligent control method to a system including a base station terminal and a server as an example. Figure 2 As shown, generating antenna unit power consumption adjustment information based on the load distribution information of each base station may include:

[0068] Step S201: Acquire the channel allocation cardinality and channel rated load parameters of each antenna unit. The channel allocation cardinality is used to represent the minimum number of allocated channels of each antenna unit.

[0069] Optionally, taking the total number of channels of the antenna terminal as 64 as an example, the channel allocation base of the antenna terminal may be 16, and the number of enabled antenna unit channels may be, but is not limited to, 16, 32, 48, and 64. The channel rated load parameter may be used to characterize the rated load of a single channel of the antenna unit.

[0070] Step S202: Generate antenna unit channel activation quantity information based on the channel allocation cardinality, the channel rated load parameter, and the antenna unit load allocation information.

[0071] Optionally, the base station terminal may obtain antenna unit load distribution information generated by the server, and may generate antenna unit channel enabled quantity information based on the channel distribution base, the channel rated load parameter, and the antenna unit load distribution information.

[0072] Exemplarily, the base station terminal can generate antenna unit channel activation quantity information based on the channel allocation cardinality, the channel rated load parameter, and the antenna unit load allocation information based on the baseband processing unit in the base station terminal. The antenna unit channel activation quantity information can be a result obtained by multiplying the channel allocation cardinality by rounding up the antenna unit load allocation information divided by the channel rated load parameter and then divided by the channel allocation cardinality, that is: Where N AUCE Enables the number of antenna unit channels, P CRL is the channel rated load parameter, I AULD For antenna unit load distribution information, B CA Assign cardinality to the channel, is the ceiling function.

[0073] Step S203 : Calculate the channel load rate by combining the antenna unit channel activation quantity information, the channel rated load parameter, and the antenna unit load distribution information.

[0074] Step S204: Allocate antenna unit channel operating frequency information according to the channel load rate.

[0075] The expression of the antenna unit channel operating frequency information can be:

[0076]

[0077] {i|i∈(N0,N max )∩Z}

[0078] Where, f CW is the antenna unit channel operating frequency information, is the lowest operating frequency of the antenna unit channel, is the channel load rate threshold corresponding to the (N0+1)th antenna unit channel operating frequency, N max is the highest frequency of the antenna unit channel operating frequency, f i is the operating frequency of the antenna unit channel in the i-th gear, μ i is the operating frequency f of the antenna unit channel in the i-th gear i The corresponding channel load rate threshold, is the highest operating frequency of the antenna unit channel, is the channel load rate threshold corresponding to the highest operating frequency of the antenna unit channel, R CL is the channel load rate, I AULD The load distribution information for antenna units, N AUCE Enables the number of antenna element channels, P CRL is the channel rated load parameter, and Z is the integer set symbol.

[0079] For example, the gear number N0 corresponding to the lowest working frequency of the antenna unit channel is 0, and the highest gear number N4 of the working frequency of the antenna unit channel is 4. max For example, the gear number N0 corresponding to the lowest working frequency of the antenna unit channel is 0, and the highest gear number N4 of the working frequency of the antenna unit channel is 4.

[0080]

[0081] In the formula, f0 is the lowest working frequency of the antenna unit channel, and f4 is the highest working frequency of the antenna unit channel.

[0082] For example, the gear number N0 corresponding to the lowest working frequency of the antenna unit channel is 0, and the highest gear number N4 of the working frequency of the antenna unit channel is 4. And It can be seen that the channel load rate R CL The value range can belong to (0, 1). Therefore, the value range of the channel load rate threshold corresponding to the working frequency of the antenna unit channel of each gear can also belong to (0, 1). And the channel load rate threshold corresponding to the working frequency of the antenna unit channel of the high gear can be greater than the channel load rate threshold corresponding to the working frequency of the antenna unit channel of the low gear.

[0083] In the above AI-based base station power consumption intelligent control method, by controlling the number of enabled channels and the working frequency, the load distribution of each antenna unit can be accurately allocated, and the channel resources and working state of the antenna unit can be reasonably controlled, thereby effectively improving the utilization rate of the antenna unit resources, further optimizing the energy consumption management of the antenna unit, and improving the adaptability of the base station to different load conditions; by using the channel allocation base and the rated load parameter, in combination with the antenna unit load distribution information, the channel enabled number information can be generated, which can flexibly adjust the number of channels, improve the utilization efficiency of the antenna unit resources, and avoid waste of channel resources; by allocating the working frequency of the antenna unit channel based on the channel load rate, the working frequency of the antenna unit can be reasonably controlled, the frequency resources can be finely managed, and the stability and reliability of the base station communication can be ensured.

[0084] In an optional embodiment of the present application, please refer to Figure 3 The AI-based base station power consumption intelligent control method can further include:

[0085] Step S305, obtaining the loss value of each channel of each antenna unit.

[0086] Step S306, in the order from low to high of the loss value of each channel, in combination with the antenna unit channel enabled number information, generating channel enabled selection information.

[0087] For example, the channel enabled selection information is used to represent the enabled channels in the antenna unit.

[0088] In step S307, the loss value of the enabled channel is updated based on the antenna unit channel working frequency information and the enabled duration of the enabled channel.

[0089] In the AI-based base station power consumption intelligent regulation method, the channel enabling selection information is generated, the channel with lower loss is accurately selected for enabling, the efficiency and stability of signal transmission are ensured, the service life of the base station equipment is improved, and the long-term stable operation of the communication network is ensured.

[0090] In an optional embodiment of the present application, the AI-based base station power consumption intelligent regulation method can further include:

[0091] The antenna unit channel enabling number information is equal to the antenna unit identified as the candidate dormant antenna unit.

[0092] The candidate dormant antenna unit with the antenna unit channel working frequency information equal to the lowest working frequency of the antenna unit channel is identified as the antenna unit to be put into dormancy.

[0093] The antenna unit to be put into dormancy is controlled to enter the dormant state.

[0094] In an optional embodiment of the present application, the node features of the node include connected user load features, antenna unit direction features, antenna unit rated load features, and base station location features, and the expression of the node feature matrix of the node can be:

[0095]

[0096] {κ|κ∈(1,M)∩Z}

[0097] In the formula, is the node feature matrix, L κ is the connected user load feature of the v-th base station, is the antenna unit direction feature of the κ-th base station, is the antenna unit rated load feature of the κ-th base station, is the base station location feature of the κ-th base station, and M is the total number of base stations, is the antenna unit direction feature of the λ-th antenna unit of the κ-th base station, C λ is the antenna unit rated load feature of the λ-th antenna unit of the κ-th base station, N κ is the total number of antenna units of the κ-th base station.

[0098] In an optional embodiment of the present application, please refer to Figure 3 The AI-based base station power consumption intelligent regulation method can further include:

[0099] Step S308, obtain the predicted overlapping user load corresponding to the overlapping user load of each base station and the predicted connected user load corresponding to the connected user load characteristics.

[0100] Step S309, input the predicted overlapping user load and the predicted connected user load into the base station load adjustment graph convolution network to generate load allocation prediction information of each base station.

[0101] Step S310, based on the load allocation prediction information of each base station, generate antenna unit load prediction information.

[0102] Step S311, generate antenna unit power consumption prediction adjustment information of the antenna unit of each base station according to the antenna unit load prediction information.

[0103] Illustratively, the antenna unit power consumption prediction adjustment information is used to represent the antenna unit channel enabling quantity information and the antenna unit channel working frequency information of each antenna unit of each base station in the prediction period.

[0104] In the above AI-based base station power consumption intelligent control method, by generating load allocation prediction information, the future load and power consumption of the base station can be accurately predicted, and reasonable power consumption planning and resource allocation can be performed in advance, so that the network traffic fluctuations can be effectively responded to, and the overall stability and reliability of the network can be improved; through the predicted antenna unit power consumption adjustment information, the working mode of the antenna unit can be adjusted in advance, which can reduce the influence of the delay caused by the working mode adjustment on the user, so that the utilization efficiency of network resources can be optimized, and user satisfaction can be improved; by understanding the load trend in advance, the maintenance and upgrade plan of the base station can be reasonably arranged, the risk of service quality decline caused by network congestion or resource shortage can be reduced, and user experience can be improved.

[0105] In an optional embodiment of the present application, step S308, obtaining the predicted overlapping user load corresponding to the overlapping user load of each base station and the predicted connected user load corresponding to the connected user load characteristics, can include:

[0106] Obtain historical load state information data of each base station, and the historical load state information data includes historical overlapping user load corresponding to the overlapping user load and historical connected user load corresponding to the connected user load characteristics.

[0107] Based on the historical load state information data, respectively construct an inter-day long short-term memory network model, a week-long long short-term memory network model, and a holiday long short-term memory network model.

[0108] According to the daytime long short-term memory network model, the weekday long short-term memory network model and the holiday long short-term memory network model, the daytime predicted overlapping user load, the weekday predicted overlapping user load and the holiday predicted overlapping user load corresponding to the overlapping user load of each base station are generated, as well as the daytime predicted connected user load, the weekday predicted overlapping user load and the holiday predicted overlapping user load corresponding to the connected user load characteristics.

[0109] The predicted overlapping user load is obtained by combining the daytime predicted overlapping user load, the weekday predicted overlapping user load and the holiday predicted overlapping user load, and the predicted overlapping user load is obtained by combining the daytime predicted connected user load, the weekday predicted overlapping user load and the holiday predicted overlapping user load.

[0110] In an exemplary embodiment of the present application, Figure 3 As shown, a method for intelligently controlling base station power consumption based on AI is provided, which may include:

[0111] Step S301 , constructing a base station load adjustment graph convolutional network with each base station as a node and the user load within the overlapping range of the maximum adjustment area of ​​the antenna unit of each base station in the connection direction between the base stations as an edge.

[0112] Step S302: Acquire real-time load status information of each base station.

[0113] Step S303: input the real-time load status information into the base station load adjustment graph convolutional network to generate load distribution information for each base station.

[0114] Step S304: generating antenna unit power consumption adjustment information based on the load distribution information of each base station.

[0115] Step S305: Obtain the loss value of each channel of each antenna unit.

[0116] Step S306 , generating channel activation selection information in the order of the loss values ​​of the channels from low to high and in combination with the information on the number of activated channels of the antenna unit.

[0117] Step S307: updating the loss value of the enabled channel based on the antenna unit channel operating frequency information and the enabled duration of the enabled channel.

[0118] Step S308: Obtain the predicted overlapping user load corresponding to the overlapping user load of each base station and the predicted connected user load corresponding to the connected user load feature.

[0119] Step S309: Input the predicted overlapping user load and the predicted connected user load into the base station load adjustment graph convolutional network to generate load distribution prediction information for each base station.

[0120] At step S310, antenna unit load prediction information is generated based on the load distribution prediction information of each base station.

[0121] At step S311, antenna unit power consumption prediction adjustment information of the antenna units of each base station is generated according to the antenna unit load prediction information.

[0122] In the AI-based base station power consumption intelligent regulation method described above, precise load distribution and power consumption regulation of the base station can be achieved, the channel life management of the antenna units of the base station can be optimized, the forward-looking regulation capability of the base station power consumption can be improved, and the stability and reliability of the base station network can be enhanced.

[0123] It should be understood that, although each step in the flowchart involved in each embodiment described above is displayed in sequence according to the arrow, these steps are not necessarily executed in the order indicated by the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other orders. Moreover, at least part of the steps in the flowchart involved in each embodiment described above can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of these steps or stages is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or steps or stages in other steps.

[0124] Based on the same inventive concept, the present application also provides an AI-based base station power consumption intelligent regulation device for implementing the AI-based base station power consumption intelligent regulation method described above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, and therefore the specific limitations in one or more AI-based base station power consumption intelligent regulation device embodiments provided below can refer to the limitations of the AI-based base station power consumption intelligent regulation method described above, which will not be described here again.

[0125] In one exemplary embodiment, as shown in Figure 4 An AI-based base station power consumption intelligent regulation device 400 is provided, which can include:

[0126] The load adjustment model construction module 401 can be used to take each base station as a node, and the user load amount in the overlapping range of the maximum adjustment area of the antenna units of each base station in the direction of the connection between the base stations as an edge, to construct a base station load adjustment graph convolution network, which is used to adjust the load distribution of each base station according to the real-time load state of each base station.

[0127] The real-time load state acquisition module 402 can be used to acquire the real-time load state information of each base station.

[0128] The load distribution information generation module 403 can be configured to input the real-time load state information into the base station load adjustment map convolutional network to generate load distribution information of each base station.

[0129] The antenna unit power consumption adjustment module 404 can be configured to generate antenna unit power consumption adjustment information based on the load distribution information of each base station, and the antenna unit power consumption adjustment information is used to adjust the power consumption of each antenna unit of each base station.

[0130] In an optional embodiment of the present application, the antenna unit power consumption adjustment module 404 can be further configured to:

[0131] Obtain a channel allocation base and a channel rated load parameter of each antenna unit, and the channel allocation base is used to represent the minimum number of allocated channels of each antenna unit.

[0132] Generate antenna unit channel enabled number information based on the channel allocation base, the channel rated load parameter, and the antenna unit load distribution information.

[0133] Combine the antenna unit channel enabled number information, the channel rated load parameter, and the antenna unit load distribution information to calculate a channel load rate.

[0134] Distribute antenna unit channel operating frequency information according to the channel load rate.

[0135] In an optional embodiment of the present application, the AI-based base station power consumption intelligent regulation device 400 can be further configured to:

[0136] Obtain a loss value of each channel of each antenna unit.

[0137] In the order from low to high of the loss value of each channel, combine the antenna unit channel enabled number information to generate channel enabled selection information, and the channel enabled selection information is used to represent the enabled channels in the antenna unit.

[0138] Update the loss value of the enabled channel based on the antenna unit channel operating frequency information and the enabled duration of the enabled channel.

[0139] In an optional embodiment of the present application, the AI-based base station power consumption intelligent regulation device 400 can be further configured to:

[0140] Identify an antenna unit with the antenna unit channel enabled number information equal to the channel allocation base as a candidate dormant antenna unit.

[0141] Identify a candidate dormant antenna unit with the antenna unit channel operating frequency information equal to the antenna unit channel minimum operating frequency as a to-be-dormant antenna unit.

[0142] Control the antenna unit to be hibernated to enter a hibernation state.

[0143] In an optional embodiment of the present application, the AI-based base station power consumption intelligent regulation device 400 can also be used for:

[0144] Obtain the predicted overlapping user load corresponding to the overlapping user load of each base station and the predicted connected user load corresponding to the connected user load characteristics.

[0145] Input the predicted overlapping user load and the predicted connected user load into the base station load regulation graph convolution network to generate load distribution prediction information of each base station.

[0146] Based on the load distribution prediction information of each base station, antenna unit load prediction information is generated.

[0147] According to the antenna unit load prediction information, antenna unit power consumption prediction regulation information of each antenna unit of each base station is generated, which is used to represent antenna unit channel enablement quantity information and antenna unit channel operating frequency information of each antenna unit of each base station in the prediction period.

[0148] In an optional embodiment of the present application, the AI-based base station power consumption intelligent regulation device 400 can also be used for:

[0149] Obtain historical load state information data of each base station, including historical overlapping user load corresponding to overlapping user load and historical connected user load corresponding to connected user load characteristics.

[0150] Based on the historical load state information data, an inter-day long short-term memory network model, an inter-week long short-term memory network model and a holiday long short-term memory network model are constructed.

[0151] According to the inter-day long short-term memory network model, the inter-week long short-term memory network model and the holiday long short-term memory network model, the inter-day predicted overlapping user load, the inter-week predicted overlapping user load and the holiday predicted overlapping user load corresponding to the overlapping user load of each base station, and the inter-day predicted connected user load, the inter-week predicted overlapping user load and the holiday predicted overlapping user load corresponding to the connected user load characteristics are generated.

[0152] The predicted overlapping user load is obtained by combining the inter-day predicted overlapping user load, the inter-week predicted overlapping user load and the holiday predicted overlapping user load, and the predicted overlapping user load is obtained by combining the inter-day predicted connected user load, the inter-week predicted overlapping user load and the holiday predicted overlapping user load.

[0153] In one embodiment, a computer device is provided, comprising a memory and a processor, the memory stores a computer program, and the processor implements the steps of the AI-based base station power consumption intelligent regulation method as described above when executing the computer program.

[0154] In one embodiment, a computer readable storage medium is provided, which stores a computer program, and the computer program is executed by a processor to implement the steps of the above method embodiments.

[0155] For the device embodiment, since it basically corresponds to the method embodiment, the relevant part is described in the part of the method embodiment. The above described device embodiment is only schematic, wherein the components described as separate components can or can not be physically separate, and the components displayed as a unit can or can not be a physical unit, i.e. can be located in one place, or also distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the present disclosure according to the actual needs. Those skilled in the art can understand and implement it without creative labor.

[0156] The above described embodiments only express several implementation manners of the present application, which are described in detail, but it should not be understood as a limitation to the patent scope of the application. It should be pointed out that, for those skilled in the art, without departing from the concept of the present application, several modifications and improvements can be made, which are all within the protection scope of the present application.

Claims

1. An AI-based intelligent control method for base station power consumption, characterized in that: The method comprises: A base station load adjustment graph convolutional network is constructed with each base station as a node and the user load within the overlapping range of the maximum adjustment area of ​​the antenna unit of each base station in the connection direction between the base stations as an edge. The base station load adjustment graph convolutional network is used to adjust the load distribution of each base station according to the real-time load status of each base station; Obtaining real-time load status information of each base station; Inputting the real-time load status information into the base station load adjustment graph convolutional network to generate load distribution information for each base station; Antenna unit power consumption adjustment information is generated based on the load distribution information of each base station, and the antenna unit power consumption adjustment information is used to adjust the power consumption of each antenna unit of each base station.

2. The method according to claim 1, characterized in that The antenna unit power consumption adjustment information includes information about the number of antenna unit channel activations and information about the operating frequency of the antenna unit channels, the load distribution information includes antenna unit load distribution information of each antenna unit, and generating the antenna unit power consumption adjustment information based on the load distribution information of each base station includes: Acquire a channel allocation cardinality and a channel rated load parameter of each antenna unit, wherein the channel allocation cardinality is used to represent a minimum number of allocated channels of each antenna unit; Generate the antenna unit channel activation quantity information based on the channel allocation cardinality, the channel rated load parameter and the antenna unit load allocation information; Calculate the channel load rate by combining the antenna unit channel enabled quantity information, the channel rated load parameter and the antenna unit load distribution information; Allocate the antenna unit channel operating frequency information according to the channel load rate; The expression of the antenna unit channel operating frequency information is: {i|i∈(N0,N max )∩Z} Where, f CW is the operating frequency information of the antenna unit channel, is the lowest operating frequency of the antenna unit channel, is the channel load rate threshold corresponding to the (N0+1)th antenna unit channel operating frequency, N max is the highest frequency of the antenna unit channel operating frequency, f i is the operating frequency of the antenna unit channel in the i-th gear, μ i is the channel load rate threshold corresponding to the operating frequency of the antenna unit channel in the i-th gear, is the highest operating frequency of the antenna unit channel, is the channel load rate threshold corresponding to the highest operating frequency of the antenna unit channel, R CL is the channel load rate, I AULD The load distribution information of the antenna unit, N AUCE Information about the number of antenna unit channels enabled, P CRL is the rated load parameter of the channel, and Z is an integer set symbol.

3. The method according to claim 2, characterized in that The method further comprises: Obtaining a loss value for each channel of each of the antenna units; Generate channel activation selection information in an ascending order of the loss values ​​of the channels and in combination with the information on the number of activated channels of the antenna unit, wherein the channel activation selection information is used to represent the activated channels in the antenna unit; The loss value of the enabled channel is updated based on the antenna unit channel operating frequency information and the activation time of the enabled channel.

4. The method according to claim 2, characterized in that The method further comprises: Identify the antenna unit whose channel activation quantity information of the antenna unit is equal to the channel allocation base as a candidate dormant antenna unit; Identify the candidate dormant antenna unit whose antenna unit channel operating frequency information is equal to the lowest operating frequency of the antenna unit channel as the antenna unit to be dormant; Control the antenna unit to be dormant to enter a dormant state.

5. The method according to any one of claims 1 to 4, characterized in that The node characteristics of the node include connected user load characteristics, antenna unit direction characteristics, antenna unit rated load characteristics and base station location characteristics. The expression of the node characteristic matrix of the node is: {λ|λ∈(1,N κ )∩Z} {κ|κ∈(1,M)∩Z} Where, is the node feature matrix, L κ is the connected user load characteristic of the κ-th base station, is the antenna unit directional characteristic of the κ-th base station, is the rated load characteristic of the antenna unit of the κ-th base station, is the base station location feature of the κ-th base station, M is the total number of base stations, is the antenna unit directional characteristic of the λth antenna unit of the κth base station, C λ is the rated load characteristic of the antenna unit of the λth antenna unit of the κth base station, N κ is the total number of antenna units of the κth base station.

6. The method according to claim 5, characterized in that The method further comprises: Obtaining a predicted overlapping user load corresponding to the overlapping user load and a predicted connected user load corresponding to the connected user load feature of each base station; Inputting the predicted overlapping user load and the predicted connected user load into the base station load adjustment graph convolutional network to generate load distribution prediction information for each base station; generating antenna unit load prediction information based on the load distribution prediction information of each base station; Antenna unit power consumption prediction and adjustment information of the antenna units of each base station is generated based on the antenna unit load prediction information, and the antenna unit power consumption prediction and adjustment information is used to characterize the antenna unit channel activation quantity information and the antenna unit channel operating frequency information of each antenna unit of each base station in the prediction period.

7. The method according to claim 6, characterized in that The obtaining of the predicted overlapping user load corresponding to the overlapping user load of each base station and the predicted connected user load corresponding to the connected user load characteristic includes: Acquire historical load status information data of each base station, wherein the historical load status information data includes a historical overlapping user load corresponding to the overlapping user load and a historical connected user load corresponding to the connected user load characteristic; Based on the historical load status information data, a daytime long short-term memory network model, a weekday long short-term memory network model, and a holiday long short-term memory network model are respectively constructed; Generate the daytime predicted overlapping user load, the weekday predicted overlapping user load, and the holiday predicted overlapping user load corresponding to the overlapping user load of each base station according to the daytime long short-term memory network model, the weekday long short-term memory network model, and the holiday long short-term memory network model, and the daytime predicted connected user load, the weekday predicted overlapping user load, and the holiday predicted overlapping user load corresponding to the connected user load feature; The predicted overlapping user load is obtained by combining the daytime predicted overlapping user load, the weekday predicted overlapping user load and the holiday predicted overlapping user load, and the predicted overlapping user load is obtained by combining the daytime predicted connected user load, the weekday predicted overlapping user load and the holiday predicted overlapping user load.

8. An AI-based intelligent control device for base station power consumption, characterized in that: The device comprises: A load adjustment model construction module is used to construct a base station load adjustment graph convolutional network with each base station as a node and the user load within the overlapping range of the maximum adjustment area of ​​the antenna unit of each base station in the connection direction between the base stations as an edge. The base station load adjustment graph convolutional network is used to adjust the load distribution of each base station according to the real-time load status of each base station; A real-time load status acquisition module, configured to acquire real-time load status information of each base station; A load distribution information generation module, configured to input the real-time load status information into the base station load adjustment graph convolutional network to generate load distribution information for each base station; The antenna unit power consumption adjustment module is used to generate antenna unit power consumption adjustment information based on the load distribution information of each base station, and the antenna unit power consumption adjustment information is used to adjust the power consumption of each antenna unit of each base station.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images