Method and apparatus for adjusting power of long term evolution (LTE) base station

Abstract

The application provides a training method and device of a power prediction model, and a long term evolution (LTE) base station power adjustment method and device. The training method comprises: obtaining historical characteristic data of an LTE base station; calculating target power corresponding to the historical characteristic data by using an energy efficiency reasonable rule algorithm; and training a power prediction model based on the historical characteristic data and the target power. The adjustment method comprises: obtaining characteristic data of the LTE base station according to a preset trigger period; inputting the characteristic data into the trained power prediction model to obtain predicted power; determining a to-be-adjusted power list by comparing the predicted power with actual power of the LTE base station; and dynamically adjusting power according to the to-be-adjusted power list. Thus, accurate prediction of the LTE base station power can be efficiently realized to guide actual power adjustment.

Description

TECHNICAL FIELD

[0001] The one or more embodiments of the present specification relate to the field of mobile communication technology, in particular to a training method and device of a power prediction model, and a method and device for adjusting power of a long term evolution (LTE) base station. BACKGROUND

[0002] At present, the long term evolution (LTE) base station is a radio transceiver station in a certain radio coverage area, which transmits information between the mobile communication switching center and the mobile phone terminal. It can achieve tens of times of time division-synchronous code division multiple access (TD-SCDMA) technology, and makes it possible to achieve ubiquitous high-speed Internet access.

[0003] However, in the service idle time, the utilization rate of the physical resource block (PRB) of the LTE cell is relatively low, and the number of connected users is also relatively small. In this case, it is obviously wasteful for the operator to provide the same base station transmission power as in the busy time. Therefore, there is an urgent need to provide a power adjustment method to monitor the cell service in real time, reduce the base station transmission power as much as possible under the condition that the user coverage perception remains unchanged, and achieve the overall energy saving goal of the communication operator, so as to realize sustainable development. In the prior art, there are many judgment conditions, complex operation, numerical deviation problems and low batch operation efficiency. SUMMARY

[0004] The present application describes a training method and device of a power prediction model, and a method and device for adjusting power of a long term evolution (LTE) base station, which can solve the above technical problems. By collecting and aggregating the LTE base station power and related data, performing data analysis, screening important features, and calculating reasonable power values as target data using energy efficiency reasonable rules, the target data is input into the extreme gradient boosting (XGBoost) algorithm to obtain a power prediction model for batch prediction of the optimal power value of the LTE base station, and to efficiently improve the energy efficiency of the base station.

[0005] According to a first aspect, a training method of a power prediction model is provided. The method comprises:

[0006] obtaining historical feature data of a long term evolution (LTE) base station; calculating target power corresponding to the historical feature data using an energy efficiency reasonable rule algorithm; and training a power prediction model based on the historical feature data and the target power.

[0007] In some embodiments, the acquiring the historical feature data of the long term evolution (LTE) base station comprises at least one of: uplink physical resource block (PRB) utilization, maximum number of users, average timing advance (TA), handover success rate, uplink available PRB number, downlink cell average throughput, uplink and downlink total throughput, average uplink channel quality indicator (CQI), radio resource control (RRC) connection establishment completion number, call establishment success rate, uplink interference average, drop rate, average number of users, and wireless connection rate.

[0008] In some embodiments, the energy efficiency reasonable rule algorithm comprises the following calculation formula:

[0009]

[0010] wherein, W represents the target power, W t represents the maximum power, W m represents the coverage-aware invariant minimum power, e represents the natural constant 1.4, p represents the average of uplink utilization and downlink utilization, n represents the number of existing users, N represents the maximum number of users, q represents the natural constant 0.98, and w represents the buffer power value.

[0011] In some embodiments, the training of the power prediction model based on the historical feature data and the target power comprises:

[0012] By performing correlation analysis on the multiple features included in the historical feature data, the importance of each feature is determined, and target features with an importance within a preset range are determined. The target features are input into the power prediction model to obtain a predicted power, and the model parameters of the power prediction model are updated based on the predicted power and the target power.

[0013] In some more specific embodiments, by performing correlation analysis on the multiple features included in the historical feature data, the importance of each feature is determined, comprising:

[0014] The historical feature data is preprocessed, and the preprocessing method comprises at least one of the following: removing duplicate data, removing undefined data, and removing null data; the correlation analysis is performed on the multiple features after the preprocessing to determine the importance of each feature.

[0015] In some more specific embodiments, the updating of the model parameters of the power prediction model based on the predicted power and the target power comprises:

[0016] Based on the predicted power and the target power, the model parameters of the power prediction model are updated in a grid search manner.

[0017] In some embodiments, the power prediction model is implemented using an extreme gradient boosting (XGBoost) algorithm.

[0018] According to a second aspect, a method for adjusting power of a long term evolution (LTE) base station is provided. The method comprises:

[0019] According to a preset trigger period, feature data of the LTE base station is obtained; the feature data is input into the power prediction model trained by the method provided in the first aspect to obtain predicted power; by comparing the predicted power with actual power of the LTE base station, a list of power to be adjusted is determined; and according to the list of power to be adjusted, dynamic adjustment of power is performed.

[0020] According to a third aspect, a device for training a power prediction model is provided. The device comprises:

[0021] A feature data obtaining module is configured to obtain historical feature data of a long term evolution (LTE) base station; a target power calculating module is configured to calculate target power corresponding to the historical feature data using an energy efficiency reasonable rule algorithm; and a training module is configured to train a power prediction model based on the historical feature data and the target power.

[0022] In some embodiments, the training module comprises: a correlation analysis unit configured to determine importance of each feature in a plurality of features included in the historical feature data by performing correlation analysis on the plurality of features; a target feature determining unit configured to determine target features having importance within a preset range in the plurality of features; a power prediction unit configured to input the target features into the power prediction model to obtain predicted power; and a parameter updating unit configured to update model parameters of the power prediction model based on the predicted power and the target power.

[0023] In some more specific embodiments, the correlation analysis unit is specifically configured to: pre-process the historical feature data, the pre-processing manner comprising at least one of the following: removing duplicate data, removing undefined data, and removing null data; and perform the correlation analysis on the plurality of features after the pre-processing to determine importance of each feature.

[0024] In some embodiments, the training module is specifically configured to update the model parameters of the power prediction model in a grid search manner based on the predicted power and the target power.

[0025] In some embodiments, the energy efficiency reasonable rule algorithm comprises the following calculation formula:

[0026]

[0027] wherein, W represents the target power, W trepresents the maximum power, W m represents the coverage perception invariant minimum power, e represents a natural constant 1.4, p represents the average of the uplink utilization rate and the downlink utilization rate, n represents the number of existing users, N represents the maximum number of users, q represents a natural constant 0.98, and w represents the buffer power value.

[0028] In some embodiments, the historical feature data includes at least one of the following: uplink physical resource block (PRB) utilization rate, maximum number of users, average timing advance (TA), handover success rate, number of uplink available physical resource blocks (PRBs), downlink cell average throughput rate, total uplink and downlink throughput, average uplink channel quality indicator (CQI), number of radio resource control (RRC) connection establishment completion, call establishment success rate, average uplink interference, drop rate, average number of users, and wireless connection rate.

[0029] According to a fourth aspect, the embodiments of the present application also provide a device for adjusting the power of a long term evolution (LTE) base station, which comprises:

[0030] a feature data acquisition module configured to acquire feature data of the LTE base station according to a preset trigger period; a power prediction module configured to input the feature data into a power prediction model trained by the device provided in the third aspect to obtain a predicted power; a power comparison module configured to determine a list of powers to be adjusted by comparing the predicted power with an actual power of the LTE base station; and a power adjustment module configured to perform dynamic adjustment of the powers according to the list of powers to be adjusted.

[0031] In the above method and device provided by the embodiments of the present application, power data obtained based on energy efficiency reasonable rules is used as target data, important features are selected from feature data such as work parameters (KPIs), and are input into an XGBoost algorithm to train a power prediction model, which is used to predict the power of an LTE base station and guide the adjustment of the actual power of the LTE base station. The calculation efficiency and accuracy of the power are effectively improved. BRIEF DESCRIPTION OF DRAWINGS

[0032] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description are briefly introduced as follows. 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 effort on the basis of these drawings.

[0033] Figure 1 a flowchart of a training method of a power prediction model provided by the embodiments of the present application is shown;

[0034] Figure 2A flowchart illustrating a method for adjusting power of a long term evolution (LTE) base station is shown.

[0035] Figure 3 A structural diagram of a device for training a power prediction model is shown.

[0036] Figure 4 A structural diagram of a device for adjusting power of a long term evolution (LTE) base station is shown. DETAILED DESCRIPTION

[0037] The schemes provided by the present specification will be described below with reference to the accompanying drawings.

[0038] In order to make the purposes, technical schemes and advantages of the embodiments of the present application clearer, the technical schemes in the embodiments of the present application will be described below with reference to the accompanying drawings.

[0039] In the description of the embodiments of the present application, the words "exemplary", "for example", or "e.g." are used to mean serving as an example, instance, or illustration. Any embodiment or design presented as "exemplary", "for example", or "e.g." in the embodiments of the present application should not be construed as being more preferred or advantageous than other embodiments or design schemes. In fact, the words "exemplary", "for example", or "e.g." are used to present relevant concepts in a specific manner.

[0040] In the description of the embodiments of the present application, the term "and / or" is merely used to describe the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B, which means that there are three cases of A alone, B alone, and A and B together. In addition, unless otherwise specified, the term "multiple" means two or more.

[0041] In addition, the terms "first", "second", etc. are used only for descriptive purposes, and should not be construed as indicating or implying relative importance or implicitly indicating the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more features. The terms "include", "contain", "have" and their variants mean "include but are not limited to", unless otherwise specifically emphasized.

[0042] Figure 1 A flowchart illustrating a method for training a power prediction model is shown. Figure 1 As shown, the method comprises the following steps:

[0043] In step S110, historical feature data of a long term evolution (LTE) base station is acquired.

[0044] In some embodiments, the historical feature data comprises at least one of: uplink physical resource block (PRB) utilization, maximum number of users, average timing advance (TA), handover success rate, number of uplink available PRBs, downlink cell average throughput, total uplink and downlink throughput, average uplink channel quality indicator (CQI), number of radio resource control (RRC) connection setup completion, call setup success rate, average uplink interference, drop rate, average number of users, and wireless connection rate.

[0045] In one embodiment, a plurality of samples can be obtained in this step, wherein each sample comprises historical feature data of a corresponding LTE base station.

[0046] In step S120, an energy efficiency reasonable rule algorithm is used to calculate a target power corresponding to the historical feature data.

[0047] In some embodiments, the energy efficiency reasonable rule algorithm in this step is calculated using the following formula (1):

[0048]

[0049] wherein W represents the target power, W t represents the maximum power, W m represents the coverage-aware invariant minimum power, e represents a natural constant 1.4, p represents the average of uplink utilization and downlink utilization, n represents the number of existing users, N represents the maximum number of users, q represents a natural constant 0.98, and w represents a buffer power value.

[0050] In some embodiments, the historical feature data in each of the above samples can be substituted into formula (1) to obtain a corresponding target power value. Thus, a plurality of target power values corresponding to the plurality of samples can be obtained.

[0051] In step S130, a power prediction model is trained based on the historical feature data and the target power.

[0052] Specifically, input features are determined based on the historical feature data, and the power prediction model is used to process the input features to obtain corresponding predicted powers. Thus, based on the predicted powers and the corresponding target powers, model parameters in the power prediction model are updated.

[0053] For determination of the input features, in some embodiments, the historical feature data can be directly used as the input features of the power prediction model.

[0054] In some embodiments, the important features can be selected from the historical feature data as the input of the power prediction model. In some specific embodiments, the importance of each feature included in the historical feature data is determined by performing a correlation analysis on the features, and then a number of features whose importance is within a preset range are determined as the important features. For example, the preset range can be the top 10% or the importance greater than 0.6. It should be understood that the number herein refers to one or more.

[0055] In some embodiments, the historical feature data can be preprocessed, and then the feature data obtained after the preprocessing is used as the input feature. For example, the preprocessing can include removing duplicate data, removing undefined data, removing values outside the valid range, removing null data, and filling missing values.

[0056] It should be noted that the selection of the important features and the preprocessing can be implemented separately or in combination.

[0057] For the updating of the model parameters, in some embodiments, the model parameters of the power prediction model can be updated by using a grid search method. In some other embodiments, the implementation of this step includes updating the model parameters of the power prediction model by using a back propagation and gradient descent method.

[0058] It should be understood that the implementation of the power prediction model can be based on a tree model such as a Gradient Boosting Decision Tree (GBDT) or an eXtreme Gradient Boosting (XGBoost), or a Deep Neural Networks (DNN).

[0059] The above is a description of the training method of the power prediction model. By using the method shown in the above embodiment, the power prediction model can be trained. Figure 1 After the power prediction model is trained by using the method shown in the above embodiment for multiple iterations, a trained power prediction model can be obtained. It should be understood that the training of the power prediction model can be performed by using a training set, and the determination of whether the power prediction model is trained can be performed by using a validation set. For example, if the accuracy of the trained power prediction model on the validation set reaches a preset threshold (such as 0.8), it is considered that the model has been trained and the training is ended. In addition, the training set and the validation set are usually mutually exclusive, but the structures of the samples are completely the same.

[0060] Further, the trained power prediction model can be used to guide the adjustment of the power of the LTE base station. Figure 2 A flowchart of a method for adjusting the power of a Long Term Evolution (LTE) base station is shown. As shown in FIG. 1, the method includes the following steps. Figure 2As shown, the method comprises the following steps:

[0061] Step S210: According to a preset trigger period, the characteristic data of the LTE base station is obtained.

[0062] In some embodiments, the preset trigger period is artificially set according to expert experience. For example, the current trigger period is one hour.

[0063] It should be understood that the characteristic items included in the characteristic data are consistent with the characteristic items input into the power prediction model in the training phase.

[0064] Step S220: The characteristic data is input into the trained power prediction model to obtain the predicted power.

[0065] Step S230: The list of power to be adjusted is determined by comparing the predicted power and the actual power of the LTE base station.

[0066] In some embodiments, the list of power to be adjusted is obtained according to the difference between the actual power of the base station and the predicted power. It should be understood that the actual power of the base station and the predicted power have no difference, and then the power to be adjusted is not entered into the list.

[0067] Step S240: According to the list of power to be adjusted, the dynamic adjustment of the power is performed.

[0068] In some embodiments, the dynamic adjustment of the power is performed under the condition that the coverage perception is unchanged.

[0069] In some specific embodiments, the unchanged coverage perception means that in the case of unchanged antenna angle and reduced power, the signal strength perceived by the user in the coverage area is unchanged. Here, there is a minimum power value, and the actual power needs to be higher than the minimum value.

[0070] Corresponding to the training method of the power prediction model provided by the present application, the present application also provides a training device. Figure 3 The structure schematic diagram of the training device of the power prediction model provided by the embodiment of the present application is shown. As shown in the figure, Figure 3 The device 300 comprises:

[0071] The characteristic data acquisition module 310 is configured to acquire the historical characteristic data of the long term evolution (LTE) base station.

[0072] The target power calculation module 320 is configured to calculate the target power corresponding to the historical characteristic data by using an energy efficiency reasonable rule algorithm.

[0073] The training module 330 is configured to train the power prediction model based on the historical characteristic data and the target power.

[0074] In some embodiments, the training module 330 specifically comprises: a correlation analysis unit configured to determine the importance of each feature in the historical feature data by performing correlation analysis on the multiple features included in the historical feature data; a target feature determination unit configured to determine a target feature in the multiple features whose importance is within a preset range; a power prediction unit configured to input the target feature into the power prediction model to obtain a predicted power; and a parameter updating unit configured to update the model parameters of the power prediction model based on the predicted power and a target power.

[0075] In some specific embodiments, the correlation analysis unit is specifically configured to: pre-process the historical feature data, and the pre-processing manner includes at least one of the following: removing duplicate data, removing undefined data, and removing null data; and perform the correlation analysis on the multiple features after the pre-processing to determine the importance of each feature.

[0076] In some embodiments, the training module 330 is specifically configured to update the model parameters of the power prediction model in a grid search manner based on the predicted power and the target power.

[0077] In some embodiments, the energy efficiency reasonable rule algorithm includes the above formula (1).

[0078] In some embodiments, the historical feature data includes at least one of the following: uplink physical resource block (PRB) utilization, maximum number of users, average timing advance (TA), handover success rate, number of available uplink PRBs, downlink cell average throughput, total uplink and downlink throughput, average uplink channel quality indicator (CQI), number of radio resource control (RRC) connection establishment completion, call establishment success rate, average uplink interference, drop rate, average number of users, and wireless connection rate.

[0079] It should be noted that the description of the device in Figure 3 may also refer to the description of the foregoing method.

[0080] Corresponding to the method for adjusting the power of a long term evolution (LTE) base station provided by the present application, the present application further provides an adjusting device. Figure 4 A structural schematic diagram of an adjusting device for the power of a long term evolution (LTE) base station provided by an embodiment of the present specification is shown. As shown in Figure 4 , the device 400 comprises:

[0081] A feature data acquisition module 410 configured to acquire feature data of the LTE base station according to a preset trigger period.

[0082] The power prediction module 420 is configured to input the feature data into the power prediction model trained by the device S300 to obtain predicted power.

[0083] The power comparison module 430 is configured to determine a list of power to be adjusted by comparing the predicted power with actual power of the LTE base station.

[0084] The power adjustment module 440 is configured to perform dynamic adjustment of the power according to the list of power to be adjusted.

[0085] It should be noted that the description of the device in the above description can also refer to the description of the foregoing method. Figure 4

[0086] According to another aspect of the embodiments, a computer readable storage medium is also provided, which stores a computer program, when the computer program is executed in a computer, the computer is caused to perform the method described in combination with Figure 1 and Figure 2 According to another aspect of the embodiments, a computer readable storage medium is also provided, which stores a computer program, when the computer program is executed in a computer, the computer is caused to perform the method described in combination with and

[0087] According to another aspect of the embodiments, a computer readable storage medium is also provided, which stores a computer program, when the computer program is executed in a computer, the computer is caused to perform the method described in combination with Figure 1 and Figure 2 Those skilled in the art should be aware that the functions described in the above one or more examples can be implemented in hardware, software, firmware or any combination thereof. When implemented in software, the functions can be stored in a computer readable medium or transmitted as one or more instructions or codes on a computer readable medium.

[0088] The above detailed description of the specific implementation of the present application further explains the purpose, technical solutions and beneficial effects of the present application. It should be understood that the above description is only a specific implementation of the present application and is not intended to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made on the basis of the technical solutions of the present application should be included in the protection scope of the present application.

Claims

1. A training method for a power prediction model, characterized in that, The method comprises: acquiring historical feature data of a long term evolution (LTE) base station; calculating target power corresponding to the historical feature data by using an energy efficiency reasonable rule algorithm; the energy efficiency reasonable rule algorithm comprises the following calculation formula: wherein W represents a target power, W t represents a maximum power, W m represents a coverage-aware invariant minimum power, e represents a natural constant 1.4, p represents an average of uplink utilization and downlink utilization, n represents a number of existing users, N represents a maximum number of users, q represents a natural constant 0.98, and w represents a buffer power value; training a power prediction model based on the historical feature data and the target power.

2. The method of claim 1, wherein, The historical feature data comprises at least one of the following: uplink physical resource block (PRB) utilization, maximum number of users, average timing advance (TA), handover success rate, uplink available PRB number, downlink cell average throughput, uplink and downlink total throughput, average uplink channel quality indicator (CQI), radio resource control (RRC) connection establishment completion number, call establishment success rate, uplink interference average, drop rate, average number of users, and wireless connection rate.

3. The method of claim 1, wherein, Training a power prediction model based on the historical feature data and the target power comprises: determining the importance of each feature in the historical feature data by performing correlation analysis on the multiple features included in the historical feature data; determining target features with an importance within a preset range from the multiple features; inputting the target features into the power prediction model to obtain predicted power; updating model parameters of the power prediction model based on the predicted power and the target power.

4. The method of claim 3, wherein, Determining the importance of each feature in the historical feature data by performing correlation analysis on the multiple features included in the historical feature data comprises: performing preprocessing on the historical feature data, wherein the preprocessing mode comprises at least one of the following: removing duplicate data, removing undefined data, and removing null data; performing correlation analysis on the multiple features after the preprocessing to determine the importance of each feature.

5. The method of claim 3, wherein, Updating model parameters of the power prediction model based on the predicted power and the target power comprises: updating model parameters of the power prediction model by using a grid search based on the predicted power and the target power.

6. The method according to any one of claims 1-5, characterized in that, The power prediction model is implemented by using an extreme gradient boosting (XGBoost) algorithm.

7. A method for adjusting power of a long term evolution (LTE) base station, characterized in that, The method comprises: acquiring feature data of the LTE base station according to a preset trigger period; inputting the feature data into the power prediction model trained by the method of claim 1 to obtain predicted power; determining a to-be-adjusted power list by comparing the predicted power with actual power of the LTE base station; and performing dynamic adjustment of power according to the to-be-adjusted power list.

8. A training device for a power prediction model, characterized in that, The device comprises: a feature data acquisition module configured to acquire historical feature data of a long term evolution (LTE) base station; a target power calculation module configured to calculate target power corresponding to the historical feature data by using an energy efficiency reasonable rule algorithm; the energy efficiency reasonable rule algorithm comprises the following calculation formula: wherein W represents a target power, W t represents a maximum power, W m represents a coverage-aware invariant minimum power, e represents a natural constant 1.4, p represents an average of uplink utilization and downlink utilization, n represents a number of existing users, N represents a maximum number of users, q represents a natural constant 0.98, and w represents a buffer power value; a training module configured to train a power prediction model based on the historical feature data and the target power.

9. A device for adjusting the power of a Long Term Evolution (LTE) base station, characterized in that, The device comprises: a feature data acquisition module configured to acquire feature data of the LTE base station according to a preset trigger period; a power prediction module configured to input the feature data into the power prediction model trained by the device of claim 8 to obtain predicted power. a power ratio module, configured to determine a power list to be adjusted by comparing the predicted power with actual power of the LTE base station; a power adjustment module, configured to dynamically adjust the power according to the power list to be adjusted.

Images