5G base station energy saving control method and system

Abstract

The application is a 5G base station energy-saving control method and system, relates to the technical field of communication, and aims at the problem of excessive use or idling of resources in the prior art.The application provides a 5G base station energy-saving control method, which comprises the following steps: real-time load monitoring, service demand analysis, intelligent decision-making, if it is determined that a component meets a sleep condition, sending a sleep instruction to the component to make it enter a sleep state, and if it is determined that the component meets a wake-up condition, sending a wake-up instruction to the component to make it return to a normal working state.The application accurately determines whether various hardware components in the base station meet sleep or wake-up conditions through an intelligent decision-making algorithm, realizes dynamic optimization configuration of hardware resources, and effectively avoids excessive use or idling of hardware resources.According to different types of hardware components and service demands, the weight coefficients and fuzzy membership functions of various factors can be flexibly adjusted, the accuracy and reliability of prediction are improved, and the application has high scalability and good flexibility.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of communication, in particular to a 5G base station energy-saving control method and system. BACKGROUND

[0002] With the rapid development of 5G communication technology, as the core facility supporting the operation of 5G network, the energy consumption problem of 5G base station is increasingly prominent. 5G base station not only needs to provide higher data transmission rate, lower latency and greater connection density, but also needs to achieve energy saving and emission reduction and reduce operating costs while meeting these high performance requirements.

[0003] Chinese patent CN114362279A discloses a 5G base station energy-saving control method and system, wherein the 5G base station energy-saving control method comprises: detecting the charging state of the battery, and determining whether the battery is in energy-saving charging according to the charging state of the battery; if yes, obtaining the power of the battery; determining whether the power of the battery is greater than a first threshold; if yes, triggering the battery power supply mode. This application can realize automatic switching of dual power supply, supply power to the equipment of 5G base station by battery power, charge and supply power at the same time during the period, switch to mains power supply mode when the battery is lower than the first threshold, which can reduce power consumption and achieve energy saving and emission reduction and reduce costs. However, this method still has obvious shortcomings, that is, it cannot accurately determine the sleep or wake-up condition of hardware components, resulting in overuse or idling of resources. SUMMARY

[0004] The present application provides a 5G base station energy-saving control method and system, which can effectively reduce the energy consumption of the base station and efficiently utilize the resources by real-time monitoring of the load condition of the base station, analyzing and predicting the business demand, and accurately determining the sleep or wake-up condition of the hardware components.

[0005] The present application provides the following technical solution: a 5G base station energy-saving control method, comprising:

[0006] S1, real-time load monitoring: real-time monitoring of communication data between the 5G base station and the user equipment to obtain the load condition of the current base station, the communication data including but not limited to data flow, number of connected users, service type information, setting high-precision data collectors at key nodes of the communication link between the base station and the user equipment to collect data samples at fixed time intervals, and pre-processing the collected data;

[0007] S2, business demand analysis: combining the priority of network business, real-time business request condition and historical business law factors to analyze and predict the business demand that may occur in the current and short term;

[0008] S3, intelligent decision: according to the result of step S1 and the result of step S2, an intelligent decision algorithm is constructed based on fuzzy logic and multi-objective optimization theory, and the intelligent decision algorithm is used to judge whether each type of hardware component in the 5G base station meets the sleep condition or the wake-up condition;

[0009] S4, if it is determined that the component meets the sleep condition, a sleep instruction is sent to the component to make it enter the sleep state; if it is determined that the component meets the wake-up condition, a wake-up instruction is sent to the component to make it return to the normal working state. The sending of the instruction is performed through the internal high-speed communication bus of the base station, and the instruction is encoded and transmitted according to the communication protocol format of the hardware component.

[0010] Further, in step S1, a denoising algorithm based on wavelet transform is used to denoise the collected data, and then the denoised data is subjected to multi-scale decomposition by setting a wavelet base function and a decomposition layer number. The high-frequency noise component is removed, and the effective load information is retained, so as to further improve the accuracy of the load data.

[0011] Further, the specific operation of step S2 includes:

[0012] S2.1, according to the dependence of the business on the network resource, the real-time requirement and the factors affecting the user experience, the priority weight of different types of business is set from high to low: emergency call business with extremely high real-time requirement and great influence on user experience, high-definition video streaming media business, and ordinary web browsing business;

[0013] S2.2, the request initiation time, the request quantity and the request business type of each type of business are captured in real time, and the request times of each type of business in a unit of time are recorded;

[0014] S2.3, the base station business historical data are collected and stored, and these historical data include the business type distribution, the business traffic and the user access quantity information in different time periods, a business demand prediction model is constructed based on the historical data, and a historical business rule prediction value is obtained;

[0015] S2.4, the weighted coefficients of the network business priority weight, the real-time business request times and the historical business rule prediction value are respectively assigned by means of weighted summation, and a comprehensive prediction value of the current and short-term business demand is obtained.

[0016] Further, in step S2.1, if there is a virtual reality / augmented reality business, the priority weight value of the virtual reality / augmented reality business is set between the emergency call business with extremely high real-time requirement and great influence on user experience and the high-definition video streaming media business. P4=0.85, the determination of the weight considers not only the characteristics of the service itself, but also the user's attention to the VR / AR service experience in market research and the importance evaluation results of such services in future development trends, and for different levels of VR / AR services, the priority weight can be further subdivided on the basis of 4 to more accurately reflect the service demand according to the specific resource requirement differences. P

[0017] Further, in step S2.2, the source IP address distribution of the service request is analyzed, the users are divided into different regional groups through clustering analysis of the source IP address of the service request, and the service request frequency and service type preference of each regional group are counted, and according to the statistical results, the service demand prediction value is segmented and defined.

[0018] Further, in step S2.3, the service demand prediction model is constructed, and the time series decomposition and adaptive weighted prediction algorithm or gray prediction model are used to realize.

[0019] Further, in step S3, the expression of the intelligent decision algorithm is:

[0020]

[0021] wherein, represents the decision value of the i-th hardware component, which is used to judge whether the component meets the sleep or wake-up condition, k is a preset sleep threshold value, which is used to determine that the component meets the sleep condition, T s T w is a preset wake-up threshold value, which is used to determine that the component meets the wake-up condition, represents the number of factors participating in the decision, including load-related factors and service demand-related factors, is the weight coefficient of each factor, represents the i-th factor about the j-th hardware component, l k is the fuzzy membership function of the i-th factor about the j-th hardware component, which is used to map the actual value of each factor to the interval [0, 1].

[0022] ​​​​​Further, in the S2 service demand analysis, in addition to recording the number of service requests, the source IP address distribution of the service requests is further analyzed, the users are divided into different regional groups through clustering analysis of the source IP addresses of the service requests, and the service request frequency and service type preference of each regional group are counted. Thus, in predicting the service demand, the regional characteristics can be combined to more accurately estimate the service demand situation that may occur in different regions, and provide more targeted service demand input for the intelligent decision algorithm.

[0023] Further, in the S3 intelligent decision algorithm, the weight coefficients of the factors of different types of hardware components can be adjusted according to the characteristics and functional requirements of the components. For the power management unit, since its main responsibility is to allocate and manage the power supply of the base station, the load value has relatively small influence on its decision, and the service demand change rate has relatively large influence on its decision. Therefore, the weight coefficient of the load value can be adjusted to =0.15, the weight coefficient of the service demand change rate can be adjusted to =0.3, and the weight coefficients of other factors are also adjusted accordingly to ensure that the intelligent decision algorithm can more accurately determine whether the component meets the sleep or wake-up condition.

[0024] Further, in the S4 hardware component control step, when sending a sleep instruction to a hardware component, in addition to encoding and transmitting according to the normal communication protocol format, a sleep duration estimate value is also attached to the instruction. The estimate value is calculated by a linear regression model according to the current service demand prediction and the characteristics of the hardware component. For a certain radio frequency unit, assuming that the service demand prediction value is Q t , the rated load is C max , and the sleep duration estimate value is T h , which can be calculated by the following formula:

[0025]

[0026] wherein k 1 and k 2 are parameters fitted by a large amount of experimental data. Through experiments on different combinations of service demand and hardware component load, the actual sleep duration value of each experiment is recorded, and linear regression analysis is performed to obtain the values of k 1 and k 2. In this way, after receiving the sleep instruction, the hardware component can not only enter the sleep state, but also reasonably arrange its low-power mode running time according to the sleep duration estimate value, thereby improving the energy saving effect.

[0027] Further, in the S4 hardware component control step, when sending the wake-up instruction to the hardware component, a wake-up priority identifier is attached in the instruction, which is determined according to the emergency degree of the service demand and the importance of the hardware component in meeting the service demand. For the case that high-bandwidth service transmission is being carried out and is about to be interrupted due to the dormancy of the hardware component, the wake-up priority identifier of the relevant hardware component is set to high. For the general service case, the wake-up priority identifier is set to medium. For the hardware component that can resume normal work at a later time and has less impact on the service, the wake-up priority identifier is set to low. By setting the wake-up priority identifier, when there are multiple hardware components that need to be woken up at the same time, the hardware components can be woken up in order according to the priority, so as to ensure the continuity and efficiency of the service.

[0028] A 5G base station energy-saving control system, the system comprises a real-time load monitoring module, a service demand analysis module, an intelligent decision algorithm module and a hardware component control module;

[0029] The real-time load monitoring module is used for monitoring the communication data between the 5G base station and the user equipment in real time to obtain the load condition of the current base station, and pre-processing the collected data; the communication data covers but is not limited to data flow, number of connected users, service type information, high-precision data collectors are set at the key nodes of the communication link between the base station and the user equipment to collect data samples at fixed time intervals, and real-time data processing algorithms are used to pre-process the collected data to remove outliers and noise interference;

[0030] The service demand analysis module is used for predicting the service demand that may occur in the current and short term based on the priority of network service, real-time service request and historical service law;

[0031] The intelligent decision algorithm module is used for constructing an intelligent decision algorithm based on fuzzy logic and multi-objective optimization theory according to the results of the real-time load monitoring module and the results of the service demand analysis module, and using the intelligent decision algorithm to judge whether each type of hardware component in the 5G base station meets the dormancy condition or the wake-up condition;

[0032] The hardware component control module sends a dormancy instruction to the component to make it enter the dormancy state if it is determined that the component meets the dormancy condition, and sends a wake-up instruction to the component to make it resume the normal working state if it is determined that the component meets the wake-up condition.

[0033] Compared with the prior art, the 5G base station energy-saving control method and system have the following beneficial effects:

[0034] 1.The application accurately judges whether various types of hardware components in the base station meet the sleep or wake-up conditions through an intelligent decision algorithm, realizes dynamic optimization configuration of hardware resources, and effectively avoids overuse or idling of hardware resources, thereby reducing the energy consumption of the base station and achieving significant energy-saving effect, which has important practical application value.

[0035] 2.The application constructs fuzzy logic and multi-objective optimization theory, flexibly adjusts the weight coefficients and fuzzy membership functions of various factors according to different types of hardware components and business demands, to adapt to different application scenarios, and can also combine historical business data and real-time business request conditions of the base station, use time series decomposition, adaptive weighting and grey prediction algorithms to predict business demand, improve the accuracy and reliability of prediction, and has a wide market application prospect due to the high scalability and flexibility of the method to adapt to the development and changes of future 5G base stations. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 a flowchart of the 5G base station energy-saving control method;

[0037] Figure 2 an operation flowchart of the intelligent decision algorithm. DETAILED DESCRIPTION

[0038] The technical solutions in the embodiments of the application will be described clearly and completely below. Obviously, the described embodiments are only some of the embodiments of the application, not all. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the application.

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the drawings needed in the embodiment or prior art description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0040] Embodiment 1

[0041] A 5G base station energy-saving control method, as shown in Figure 1 and Figure 2 , includes:

[0042] S1, real-time monitoring of communication data between 5G base station and user equipment to obtain the load condition of the current base station, and pre-processing the collected data; the collected data is denoised by using a wavelet transform-based denoising algorithm, and then the denoised data is subjected to multi-scale decomposition by setting a wavelet base function and a decomposition layer number.

[0043] S2, predicting the service demand that may occur in the current and short term based on the priority of network service, real-time service request and historical service law; the specific operation includes:

[0044] S2.1, setting the priority weight of different types of services from high to low according to the dependence of the service on network resources, real-time requirement and factors affecting user experience: emergency call service with extremely high real-time requirement and great influence on user experience, virtual reality / augmented reality service, high-definition video streaming media service and ordinary web browsing service; for the emergency call service with extremely high real-time requirement and great influence on user experience, the priority weight is set as P1=0.9, for the high-definition video streaming media service, the priority weight is set as P2=0.7 according to its resolution and frame rate requirement, and for the ordinary web browsing service, the priority weight is set as P3=0.5. These weights are determined based on a large amount of market research and comprehensive evaluation of the importance of different services in actual application scenarios. Since the virtual reality / augmented reality service has extremely high requirements for bandwidth, delay and stability, the priority weight is set as P4=0.85. The determination of this weight considers not only the characteristics of the service itself, but also the market research of user attention to VR / AR service experience and the importance evaluation results of such services in future development trend, and for different levels of VR / AR service, the priority weight can be further subdivided based on P4 according to the difference of specific resource requirements, so as to more accurately reflect the service demand.

[0045] S2.2, real-time capturing of the request initiation time, request quantity and request service type of various services, and recording the request times of each type of service in unit time; a real-time service request monitor is deployed at the network access port of the base station, which can real-time capture the request initiation time, request quantity and request service type information of various services, and record the request times of each type of service in unit time, denoted as R i , wherein i represents different service types;

[0046] S2.3, collect and store base station service history data, build service demand prediction model based on history data, get history service rule prediction value; these history data include service type distribution, service traffic and user access quantity information in different time period, through deep mining and analysis of these history data; build service demand prediction model uses time series decomposition and adaptive weighted prediction algorithm formula as follows:

[0047]

[0048] wherein, Q t represents the service demand prediction value at future time t n and m are the order of trend item and seasonal item in time series decomposition, determined according to periodic characteristics of history data, are adaptive weight coefficients, represents the seasonal component of service demand at time t-i in history data, T t-i represents the trend component of service demand at time t-i in history data, I t-j represents the irregular component of service demand at time t-j

[0049] The basic formula of the grey prediction model used to build the service demand prediction model is:

[0050]

[0051] wherein, represents original service demand time series data, represents new sequence generated by once accumulation of original data, a b are parameters fitted by least square method, the determination process is: put original service demand time series data into the basic formula of grey prediction model, solve parameters a b by least square method, so that the error square sum of prediction value and actual value is minimum, through weighted fusion of results of two prediction methods, the accuracy of service demand prediction can be further improved.

[0052] ​​​​​S2.4, assign the weighted coefficients of network service priority weight, number of real-time service requests and historical service regularity prediction value respectively through weighted summation to obtain the comprehensive prediction value of current and short-term service demand: the weighted coefficients of network service priority weight, number of real-time service requests and historical service regularity prediction value are set as 、 、 ,and The weighting coefficient is determined based on a large number of simulation experiments and quantitative evaluation of the importance of different factors in business demand forecasting. =0.4, =0.3, =0.3.

[0053] S3, based on the results of step S1 and step S2, constructing an intelligent decision-making algorithm based on fuzzy logic and multi-objective optimization theory, and using the intelligent decision-making algorithm to determine whether various hardware components in the 5G base station meet the sleep condition or wake-up condition;

[0054]

[0055] in, Indicates the k The decision value of a hardware component is used to determine whether the component meets the sleep or wake-up conditions. hour, T s is the preset sleep threshold, which determines that the component meets the sleep condition. hour, T w is the preset wake-up threshold, which determines whether the component meets the wake-up condition. Indicates the number of factors involved in the decision, including load-related factors and business demand-related factors, is the weight coefficient of each factor, Indicates the l Factors about k The fuzzy membership function of each hardware component is used to map the actual value of each factor to the interval [0, 1]. Indicates the l Factors about k The fuzzy membership function of each hardware component.

[0056] Fuzzy membership function as decision value D k The core input factors of C k The value of may affect the factor weight Dynamic allocation of D kThe comparison with the threshold value finally determines the sleep or wake-up state of the hardware component.

[0057] S4, if the component meets the sleep condition, a sleep instruction is sent to the component to make it enter the sleep state; if the component meets the wake-up condition, a wake-up instruction is sent to the component to make it return to the normal working state.

[0058] When the sleep instruction is sent to the hardware component, in addition to being encoded and transmitted according to the normal communication protocol format, a sleep duration estimation value is also attached to the instruction, which is calculated by a linear regression model according to the current business demand prediction and the characteristics of the hardware component. For a certain radio frequency unit, assuming that the business demand prediction value is Q t and the rated load is C max , the sleep duration estimation value is T h which can be calculated by the following formula:

[0059]

[0060] wherein, k 1 and k 2 are parameters fitted by a large amount of experimental data. Through experiments on different combinations of business demand and hardware component load, the actual sleep duration value of each experiment is recorded, and linear regression analysis is performed to obtain the values of k 1 and k 2. In this way, after receiving the sleep instruction, the hardware component can not only enter the sleep state, but also reasonably arrange its low-power mode running time according to the sleep duration estimation value, thereby improving the energy saving effect.

[0061] When the wake-up instruction is sent to the hardware component, a wake-up priority identifier is attached to the instruction, which is determined according to the urgency of the business demand and the importance of the hardware component in meeting the business demand. For the case where high-bandwidth business transmission is being performed and will be interrupted due to the sleep of the hardware component, the wake-up priority identifier of the related hardware component is set to high. For general business conditions, the wake-up priority identifier is set to medium. For hardware components that can resume normal work at a later time and have less impact on the business, the wake-up priority identifier is set to low. By setting the wake-up priority identifier, when there are multiple hardware components that need to be woken up at the same time, they can be woken up in order according to the priority, thereby ensuring the continuity and efficiency of the business.

[0062] Embodiment 2

[0063] A 5G base station energy saving control system, comprising:

[0064] a real-time load monitoring module, configured to monitor communication data between the 5G base station and user equipment in real time to obtain a load condition of the current base station, and to pre-process collected data;

[0065] a service demand analysis module, configured to predict service demands that are likely to occur in the current and short term based on priorities of network services, real-time service requests, and historical service rules;

[0066] an intelligent decision algorithm module, configured to construct an intelligent decision algorithm based on fuzzy logic and multi-objective optimization theory according to results of the real-time load monitoring module and the service demand analysis module, and to determine whether each type of hardware component in the 5G base station meets a sleep condition or a wake-up condition by using the intelligent decision algorithm;

[0067] a hardware component control module, configured to send a sleep instruction to a component to make the component enter a sleep state if it is determined that the component meets the sleep condition, and to send a wake-up instruction to the component to make the component return to a normal working state if it is determined that the component meets the wake-up condition.

[0068] Application Example 1

[0069] This application example describes that, in a central urban area, during morning and evening peak hours (for example, 7:00-9:00 am and 5:00-7:00 pm on weekdays), a large number of users simultaneously use the 5G network for high-data-traffic activities, resulting in a sharp increase in the load of the 5G base station. In order to ensure user experience while achieving energy saving, the 5G base station energy-saving control method is adopted.

[0070] High-precision data collectors are arranged at key nodes of the communication link between the base station and the user equipment to collect data samples in units of seconds. It is assumed that at 8:00 am during the morning peak on a certain weekday, the collected data is as follows: the data traffic reaches 500 MB within 10 seconds, the number of connected users is 2000, and the service types are mainly concentrated in high-definition video streaming services (accounting for 60%), online conference services (accounting for 30%), and other services (accounting for 10%). A denoising algorithm based on wavelet transform is used to pre-process the data to remove outliers and noise interference, ensuring the accuracy and reliability of the data.

[0071] According to the market research results, the priority weight of the high-definition video streaming service is set to 0.7, and the priority weight of the online conference service is set to 0.8, because these services have a significant impact on user experience and have high real-time requirements. Through the real-time service request monitor, the request initiation time, the number of requests, and the types of various services are captured. Within 10 minutes from 8:00 to 8:10, the number of high-definition video streaming service requests is 1500, the number of online conference service requests is 800, and the number of other service requests is 200.

[0072] Combined with historical business data, the forecasting algorithm of time series decomposition and adaptive weighting is used to build a business demand forecasting model. It is assumed that the order of trend items in time series decomposition is determined by analyzing the historical data of the same period in the past week. n =3, seasonal term order m =2, after the adaptive weight coefficient is calculated and adjusted, =0.4, =0.3, =0.2, =0.1, =0.05, predicted by the model, the business demand forecast value from 8:10 to 8:20 Q t The predicted business demand value is 1200 (the business demand forecast here is a quantitative value after comprehensively considering various types of business, and can be converted based on factors such as actual business volume). We conducted an in-depth analysis of the source IP address distribution of business requests and divided users into five regional groups through cluster analysis. Among them, regional group A had the highest frequency of business requests, accounting for 40% of the total number of requests. The business type preference of this regional group was mainly high-definition video streaming services (accounting for 70%).

[0073] Based on fuzzy logic and multi-objective optimization theory, an intelligent decision-making algorithm is constructed. Assuming that the number of factors involved in decision-making is L=5, l =1 is the current load value, l =2 is the business demand forecast value, l =3 is the business type priority, l =4 is the service request frequency, l =5 is the business demand change rate, and the weight coefficients of each factor are: =0.2 (load value weight), =0.3 (business demand forecast value weight), =0.15 (business type priority weight), =0.15 (service request frequency weight), =0.2 (weight of business demand change rate). For a hardware component (such as a radio frequency unit), its rated load is 1000 (the unit can be set according to actual conditions, such as Mbps), the current load value is 700, and the minimum load value for normal operation is , calculate the decision value of the hardware component through the intelligent decision algorithm D k : , where the fuzzy membership function of the load value is for:

[0074] Will C 1k =700,C 1max =1000, C 1min =200 Substitute to get .

[0075] Assume that the fuzzy membership values ​​of other factors after calculation and mapping are: F2(C 2k ,B k )=0.6(related to business demand forecast value), F3(C 3k ,B k )=0.7 (business type priority related, corresponding to the priority weight mapping of high-definition video streaming business), F4(C 4k ,B k )=0.5 (related to the frequency of business requests), F5(C 5k ,B k ) = 0.4 (related to the rate of change of business demand, assuming it is calculated based on recent changes in business demand). This segmented definition approach can more meticulously reflect the impact of business demand forecasts on hardware component decisions in different intervals, improving the accuracy of the intelligent decision-making algorithm.

[0076] but .

[0077] Preset sleep threshold T s =0.4, wake-up threshold T w =0.6, due to D k =0.515> T s and D k < T w , the component currently does not meet the sleep or wake-up conditions, which means that the running status of the component does not need to be adjusted and no processing is performed temporarily.

[0078] If the intelligent decision algorithm determines that a component meets the sleep condition, a sleep instruction is sent to the component to put it into sleep mode. The sleep instruction is accompanied by an estimated sleep duration. Assuming that for the radio frequency unit, the service demand forecast value Q t =1200, its rated load is 1000, and the parameters obtained by linear regression model fitting are k 1=0.01, k 2=5, then the estimated sleep duration minute.

[0079] If it is determined that a component meets the wake-up condition, a wake-up instruction is sent to the component to restore it to a normal working state. The wake-up instruction is accompanied by a wake-up priority identifier, which is determined based on the urgency of the business requirement and the importance of the hardware component in meeting the business requirement. For example, for a high-definition video streaming service that is being transmitted and is about to be interrupted due to the sleep of a hardware component, the wake-up priority identifier of the relevant hardware component is set to high. The instruction is sent through the internal high-speed communication bus of the base station and encoded according to the communication protocol format of the hardware component to ensure the accuracy and timeliness of the instruction.

[0080] Application Example 2

[0081] This application example describes a remote area where the user distribution is sparse and the daily 5G base station load is usually low. However, in order to meet the occasional high-load business requirements, the base station needs to maintain a high operating power to ensure business continuity. In order to achieve energy saving while ensuring business continuity, the following 5G base station energy saving control method is adopted.

[0082] Data collectors are set at the key nodes of the communication link between the base station and the user equipment to collect data samples in minutes. Suppose that at 14:00 on a certain day, the collected data is as follows: the data traffic is 10MB in 1 minute, the number of connected users is 50, and the business types are mainly ordinary web browsing business (70%), voice call business (20%), and other businesses (10%). Considering the special nature of the communication environment in remote areas, the data collector should have high stability and anti-interference capability. A denoising algorithm based on wavelet transform is used to preprocess the data to remove abnormal values and noise interference caused by environmental factors or equipment failure, ensuring the accuracy and reliability of the data.

[0083] According to historical business data, the business rules in remote areas are analyzed in depth. Due to the sparse distribution of users, business requirements usually have great volatility and uncertainty. A method combining time series decomposition, adaptive weighting, and grey prediction model is used to build a business demand prediction model. Suppose that through analysis of the historical data of the past month, the order of the trend item in time series decomposition is determined as , the order of the seasonal item is , the adaptive weight coefficient is =0.5, =0.3, =0.2, and at the same time, the business demand is predicted by a grey prediction model. Suppose that the original business demand time series data is , , (the data here is the quantitative value converted according to the business volume), the parameters are obtained by least squares fitting a =-0.1,b =55, the result of the weighted fusion of the results of the two prediction methods (assuming that the time series decomposition and adaptive weighted prediction result weight is 0.6, and the gray prediction model result weight is 0.4), the predicted value of the service demand from 14:10 to 14:20 Q t 55 (also a comprehensive quantitative value).

[0084] Based on fuzzy logic and multi-objective optimization theory, an intelligent decision algorithm is constructed, assuming that the number of factors participating in decision-making L=4, l =1 is the current load value, l =2 is the service demand prediction value, l =3 is the service type priority, l =4 is the service request frequency, and the weight coefficients of each factor are: =0.15 (load value weight), =0.35 (service demand prediction value weight), =0.2 (service type priority weight), =0.3 (service request frequency weight), for a certain RF unit key hardware component, the rated load is 500 (units can be set according to actual conditions, such as Mbps), the current load value is 100, and the minimum load value that can be normally operated is 50. The decision value of the hardware component is calculated by the intelligent decision algorithm D k : , wherein the fuzzy membership function of the load value is:

[0085]

[0086] C 1k =100, C 1max =500, C 1min =50 into .

[0087] Assuming that the fuzzy membership values of other factors after calculation and mapping are: F2(C 2k ,B k )=0.5 (related to service demand prediction value), F3(C 3k ,B k )=0.4 (related to service type priority, corresponding to the priority weight mapping of ordinary web browsing service), F4(C 4k ,B k )=0.3 (related to service request frequency, assuming calculated according to recent service request times). ​

[0088] then .

[0089] preset dormancy threshold T s = 0.4, wake-up threshold T w = 0.6. Since D k = 0.4785 > 0.4 T s and D k < T w , the component does not currently meet the dormancy or wake-up conditions.

[0090] According to the output of the intelligent decision algorithm, a dormancy instruction is sent to the hardware component that meets the dormancy condition, along with an estimated dormancy duration value. In remote areas, due to the uncertainty of business demand, the estimated dormancy duration value can be set to a shorter interval, such as 1-2 minutes. Assuming that for this radio unit, based on the current business demand prediction and the characteristics of the hardware component, the estimated dormancy duration value is calculated to be 1.5 minutes (the calculation process is similar to that of Example One, and the relevant parameter values are assumed here), the dormancy instruction should include the dormancy mode of the hardware component and the wake-up condition, to ensure that the hardware component can maintain a low power consumption state during dormancy and quickly recover to the working state when needed. According to the urgency of business demand, set the wake-up priority identifier, for example, if there is a voice call service request, the wake-up priority identifier of the related hardware component is set to high, and when the business demand recovers, the hardware components are awakened in order according to the wake-up priority identifier, to ensure the continuity of critical business.

[0091] After implementing the energy-saving control method, the energy-saving effect and business continuity of the base station are periodically evaluated. Assuming that before implementing the energy-saving control method, the energy consumption of the remote area base station is 50 degrees per day, and the business interruption rate is 5%. After implementation, after a week of monitoring, the average daily energy consumption is reduced to 40 degrees, and the business interruption rate is reduced to 3%. By comparing the energy consumption data and business interruption rate indicators before and after implementation, the effectiveness of the energy-saving control method is evaluated, and the energy-saving control method is optimized and adjusted according to the evaluation results, such as adjusting the parameters of the business demand prediction model, optimizing the decision rules of the intelligent decision algorithm, and adjusting the dormancy and wake-up thresholds of the hardware components. Communicate with users in remote areas to collect feedback on the performance and energy-saving effect of the base station, in order to further improve and optimize the energy-saving control method.

[0092] It will be apparent to those skilled in the art that the application is not limited to the details of the above-exemplified embodiments and that the present application can be implemented in other particular forms without departing from the spirit or essential characteristics of the present application. The embodiments should therefore be considered in all respects as illustrative and not restrictive, the scope of the application being indicated by the appended claims rather than by the above description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. No reference signs in the claims should be considered as limiting the scope of the claims with respect to the figures of the patent document.

Claims

1. A 5G base station energy-saving control method, characterized in that: include: S1, real-time monitoring of communication data between 5G base stations and user equipment to obtain the current base station load and pre-process the collected data; S2, based on network service priorities, real-time service requests, and historical service patterns, predicts current and short-term service needs; S3, based on the results of step S1 and step S2, constructing an intelligent decision-making algorithm based on fuzzy logic and multi-objective optimization theory, and using the intelligent decision-making algorithm to determine whether various hardware components in the 5G base station meet the sleep condition or wake-up condition; The expression of the intelligent decision-making algorithm is: in, Indicates the k The decision value of a hardware component is used to determine whether the component meets the sleep or wake-up conditions. hour, T s is the preset sleep threshold, which determines that the component meets the sleep condition. hour, T w is the preset wake-up threshold, which determines whether the component meets the wake-up condition. Indicates the number of factors involved in the decision, including load-related factors and business demand-related factors, is the weight coefficient of each factor, Indicates the l Factors about k The fuzzy membership function of each hardware component is used to map the actual value of each factor to the interval [0, 1]; No. l Factors about k Fuzzy membership function of hardware components for: Fuzzy membership function as decision value D k The core input factors of C k The value of the influencing factor weight Dynamic allocation of D k The comparison with the threshold ultimately determines the sleep or wake-up state of the hardware component; S4, if it is determined that the component meets the sleep condition, a sleep instruction is sent to the component to put it into a sleep state; if it is determined that the component meets the wake-up condition, a wake-up instruction is sent to the component to restore it to a normal working state.

2. A 5G base station energy-saving control method according to claim 1, characterized in that: In step S1, a denoising algorithm based on wavelet transform is used to denoise the collected data, and then the denoised data is decomposed into multiple scales by setting the wavelet basis function and the number of decomposition layers.

3. A 5G base station energy-saving control method according to claim 1, characterized in that: The specific operations of step S2 include: S2.1: Prioritize different types of services based on their reliance on network resources, real-time requirements, and factors affecting user experience, from high to low: emergency calls (which are highly real-time and have a significant impact on user experience), high-definition video streaming services, and standard web browsing services. S2.2, captures the time, number, and type of requests for each type of service in real time, and records the number of requests for each type of service per unit time; S2.3, collect and store historical base station service data, build a service demand forecasting model based on the historical data, and obtain the predicted value of historical service patterns; S2.4, assign weighted coefficients of network service priority weight, number of real-time service requests, and historical service rule prediction value respectively through weighted summation to obtain a comprehensive prediction value of current and short-term service demand.

4. A 5G base station energy-saving control method according to claim 3, characterized in that: In step S2.1, if there is a virtual reality / augmented reality service, its priority weight value is set between the priority weight values ​​of the emergency call service and the high-definition video streaming service, which are extremely real-time and have a significant impact on user experience.

5. A 5G base station energy-saving control method according to claim 3, characterized in that: In step S2.2, the distribution of source IP addresses of business requests is analyzed. By performing cluster analysis on the source IP addresses of business requests, users are divided into different regional groups. The business request frequency and business type preference of each regional group are counted. Based on the statistical results, the business demand forecast value is segmented and defined.

6. A 5G base station energy-saving control method according to claim 3, characterized in that: In step S2.3, the business demand forecasting model is constructed using a time series decomposition and adaptive weighted forecasting algorithm or a grey forecasting model.

7. A 5G base station energy-saving control method according to claim 1, characterized in that: In step S4, when a sleep instruction is sent to the hardware component, an estimated sleep duration value is included in the sleep instruction. The estimated value is calculated using a linear regression model based on the current business demand forecast and the characteristics of the hardware component.

8. A 5G base station energy-saving control method according to claim 1, characterized in that: In step S4, when a wake-up instruction is sent to the hardware component, the wake-up instruction is accompanied by a wake-up priority identifier, which is determined according to the urgency of the business demand and the importance of the hardware component in meeting the business demand. By setting the wake-up priority identifier, the hardware component is awakened in order of priority.

9. A 5G base station energy-saving control system, characterized in that: include: Real-time load monitoring module, used to monitor the communication data between 5G base stations and user equipment in real time to obtain the current base station load and pre-process the collected data; The business demand analysis module is used to predict current and short-term business needs based on network business priorities, real-time business requests, and historical business patterns; The intelligent decision-making algorithm module is used to construct an intelligent decision-making algorithm based on fuzzy logic and multi-objective optimization theory according to the results of the real-time load monitoring module and the results of the business demand analysis module. This intelligent decision-making algorithm is used to determine whether various hardware components in the 5G base station meet the sleep conditions or wake-up conditions. The expression of the intelligent decision-making algorithm is: in, Indicates the k The decision value of a hardware component is used to determine whether the component meets the sleep or wake-up conditions. hour, T s is the preset sleep threshold, which determines that the component meets the sleep condition. hour, T w is the preset wake-up threshold, which determines whether the component meets the wake-up condition. Indicates the number of factors involved in the decision, including load-related factors and business demand-related factors, is the weight coefficient of each factor, Indicates the l Factors about k The fuzzy membership function of each hardware component is used to map the actual value of each factor to the interval [0, 1]; No. l Factors about k Fuzzy membership function of hardware components for: Fuzzy membership function as decision value D k The core input factors of C k The value of the influencing factor weight Dynamic allocation of D k The comparison with the threshold ultimately determines the sleep or wake-up state of the hardware component; The hardware component control module sends a sleep instruction to the component to put it into sleep mode if it determines that the component meets the sleep condition; and sends a wake-up instruction to the component to restore it to normal working mode if it determines that the component meets the wake-up condition.

Images