Dynamic load prediction method and apparatus, base station, and storage medium

Abstract

The application relates to a dynamic load prediction method and device, a base station and a storage medium. The method comprises the following steps: in a plurality of base station cells located on a preset load migration line, a cross-layer co-directional group of a target base station cell is determined, real-time load of the target base station cell at a current moment and historical predicted load corresponding to a previous moment are obtained; cross-layer adjacent cell load data transmitted by a left adjacent cell co-directional group and a right adjacent cell co-directional group of the cross-layer co-directional group of the target base station cell is received, the corresponding cross-layer adjacent cell load data comprises historical load data and real-time load data; a preset load prediction model is used to perform load prediction based on the real-time load, the historical predicted load and the cross-layer adjacent cell load data, forward predicted load and reverse predicted load corresponding to the target base station cell are obtained, and a corresponding load prediction result is determined based on the forward predicted load and the reverse predicted load.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of base station systems, and in particular to a dynamic load prediction method and device, a base station and a storage medium. BACKGROUND

[0002] In related technologies, the load of a mobile communication network usually refers to the number of users (terminals) accessing a network base station cell or the resource occupancy rate of the network base station cell. Network load is a data index that is highly concerned in mobile communication network operation and maintenance, is a key index for real-time monitoring of the network, and is directly related to the smooth operation of the network, user perception and energy saving. Therefore, it is a research hotspot and application focus of mobile communication networks. In related technologies, through load prediction, various energy-saving measures, timely awakening, and advance implementation of load balancing and admission control for high traffic special protection can be assisted and guided to ensure the smooth operation of the base station and the service perception of users.

[0003] In related technologies, for load prediction, an Auto Regressive Integrated Moving Average (ARIMA) model is commonly used. The model is trained and learned and regressed using historical load data of the network base station cell to predict the future load of each base station cell. A deep neural network model, such as a Recurrent Neural Network (RNN) model and an improved Long-short term Memory (LSTM), is also used to implement load prediction. However, the known prediction methods usually require a long enough time and a large enough amount of historical data to achieve effective training and learning and to make a relatively accurate prediction. In addition, the load prediction requires the load of the base station cell to have a periodic pattern. For a high-speed mobile communication scenario without a clear time period pattern, a complex network topology and user movement trajectory, or a relatively simple network topology but with high-speed user movement and unpredictable load burst, the existing load prediction methods cannot achieve accurate and effective load prediction.

[0004] There is no effective solution to the problem that the load prediction method in related technologies cannot be applied to a high-speed mobile communication scenario with high-speed user movement, load burst and no accurate time period pattern. SUMMARY

[0005] The present application provides a dynamic load prediction method, device, base station and storage medium to at least solve the problem that the load prediction method in related technologies cannot be applied to a high-speed mobile communication scenario with high-speed user movement, load burst and no accurate time period pattern.

[0006] In a first aspect, the application provides a dynamic load prediction method applied to a target base station cell under a high-speed private network, comprising: determining a cross-layer co-directional group in a plurality of base station cells located on a preset load migration line, wherein the base station cells located at both ends of the cross-layer co-directional group are K-layer main neighboring areas corresponding to the target base station cell, K being a preset value; acquiring real-time load at a current time and historical predicted load corresponding to a previous time, the historical predicted load being used to represent load migration change to be generated at the current time predicted at the previous time; receiving cross-layer neighboring area load data transmitted by a left neighboring area co-directional group and a right neighboring area co-directional group of the cross-layer co-directional group, wherein the corresponding cross-layer neighboring area load data each comprises historical load data and real-time load data, the historical load data being used to represent the K-layer main neighboring area possessed by the corresponding neighboring area co-directional group corresponding to the predicted load at the previous time, and the real-time load data being used to represent the K-layer main neighboring area possessed by the corresponding neighboring area co-directional group actual load at the current time; performing load prediction based on the real-time load, the historical predicted load, and the cross-layer neighboring area load data by using a preset load prediction model, to obtain forward predicted load and reverse predicted load corresponding to the target base station cell, and determining a corresponding load prediction result based on the forward predicted load and the reverse predicted load.

[0007] In a second aspect, the application provides a dynamic load prediction device applied to a target base station cell under a high-speed private network, comprising:

[0008] A determining module is configured to determine a cross-layer co-directional group in a plurality of base station cells located on a preset load migration line, wherein the base station cells located at both ends of the cross-layer co-directional group are K-layer main neighboring areas corresponding to the target base station cell, K being a preset value.

[0009] An acquiring module is configured to acquire real-time load at a current time and historical predicted load corresponding to a previous time, the historical predicted load being used to represent load migration change to be generated at the current time predicted at the previous time.

[0010] A receiving module is configured to receive cross-layer neighboring area load data transmitted by a left neighboring area co-directional group and a right neighboring area co-directional group of the cross-layer co-directional group, wherein the corresponding cross-layer neighboring area load data each comprises historical load data and real-time load data, the historical load data being used to represent the K-layer main neighboring area possessed by the corresponding neighboring area co-directional group corresponding to the predicted load at the previous time, and the real-time load data being used to represent the K-layer main neighboring area possessed by the corresponding neighboring area co-directional group actual load at the current time.

[0011] The prediction module is configured to perform load prediction based on the real-time load, the historical predicted load and the cross-layer neighbor cell load data by using a preset load prediction model, to obtain forward predicted load and reverse predicted load corresponding to the target base station cell, and to determine a corresponding load prediction result based on the forward predicted load and the reverse predicted load.

[0012] In a third aspect, a base station is provided, comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus.

[0013] The memory is configured to store a computer program.

[0014] The processor is configured to execute the program stored in the memory, and implement the steps of the dynamic load prediction method according to any one of the embodiments of the first aspect.

[0015] In a fourth aspect, a computer readable storage medium is provided, which stores a computer program. The computer program is executed by a processor to implement the steps of the dynamic load prediction method according to any one of the embodiments of the first aspect.

[0016] Compared with the related art, the dynamic load prediction method, device, base station and storage medium provided in the embodiments of the present application have the beneficial effects that the cross-layer co-directional group is determined in the multiple base station cells located on the preset load migration line, wherein the base station cells located at the two ends of the cross-layer co-directional group are both the K-layer main adjacent areas corresponding to the target base station cell, and K is a preset value; the real-time load at the current time and the historical predicted load corresponding to the last time are obtained, and the historical predicted load is used to represent the load migration change that will occur at the current time predicted at the last time; the cross-layer adjacent area load data transmitted by the left adjacent area co-directional group and the right adjacent area co-directional group of the cross-layer co-directional group are received, wherein the corresponding cross-layer adjacent area load data both include the historical load data and the real-time load data, the historical load data is used to represent the predicted load of the K-layer main adjacent area of the corresponding adjacent area co-directional group at the last time, and the real-time load data is used to represent the actual load of the K-layer main adjacent area of the corresponding adjacent area co-directional group at the current time; the load prediction is performed based on the real-time load, the historical predicted load and the cross-layer adjacent area load data by using the preset load prediction model, the forward predicted load and the reverse predicted load corresponding to the target base station cell are obtained, and the corresponding load prediction result is determined based on the forward predicted load and the reverse predicted load, thereby solving the problem that the load prediction method in the related art cannot be applied to the high-speed mobile communication scene without the accurate time period rule of the user high-speed movement and the load burst, and realizing the beneficial effects that the base station cell can detect and sense the user load and the change trend of the adjacent area coming from a farther distance at an accurate time point in advance, and the prediction load and the actual load of the base station cell are combined to provide effective reference and guidance for the energy saving, load control, traffic guarantee and other operations of the base station cell.

[0017] The details of one or more embodiments of the present application are presented in the following drawings and description to make other features, objects and advantages of the present application more apparent. BRIEF DESCRIPTION OF DRAWINGS

[0018] The accompanying drawings, which are incorporated into and form part of the specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the application.

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced as follows, and obviously, other drawings can also be obtained by those skilled in the art without any creative labor based on these drawings.

[0020] Figure 1 is a flowchart of the dynamic load prediction method provided in the embodiments of the present application;

[0021] Figure 2 is a structure diagram of the cross-layer co-directional group in the embodiments of the present application;

[0022] Figure 3 is a structural block diagram of a dynamic load prediction device provided by an embodiment of the present application;

[0023] Figure 4 is a structural schematic diagram of a base station of an embodiment of the present application. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work belong to the scope of protection of the present application.

[0025] Before the embodiments of the present application are described, the related technical names involved in the present application are described as follows:

[0026] Base Station (BS for short) is a basic unit of a mobile network.

[0027] Cell, the basic unit of a base station, under the conventional clover structure of a cellular base station, one base station has three 120° sector-shaped cells, and under a high-speed rail private network, one base station has two or more cells.

[0028] User Equipment (UE for short).

[0029] The Fifth-Generation mobile communication (5G for short).

[0030] The Fourth-Generation mobile communication (4G for short).

[0031] NEW Ratio (NR for short).

[0032] Long Term Evolution (LTE for short).

[0033] Auto Regressive Integrated Moving Average (ARIMA for short), a time series model based on the improvement of ARMA.

[0034] Recurrent Neural Network (RNN).

[0035] Long Short-Term Memory (LSTM) is an improved neural network model based on RNN.

[0036] Next, the technical solutions in the embodiments of the present application will be described in conjunction with the accompanying drawings in the embodiments of the present application.

[0037] Figure 1 It is a schematic flowchart of the dynamic load prediction method provided for the embodiments of the present application. As Figure 1 shown, the embodiments of the present application provide a dynamic load prediction method, which is applied to the target base station cell under the high-speed private network. The method includes the following steps:

[0038] Step S101: Among multiple base station cells located on the preset load migration line, determine the cross-layer same-direction group. Among them, the base station cells at both ends of the cross-layer same-direction group are the Kth-layer main neighboring cells corresponding to the target base station cell, and K is a preset value.

[0039] In the embodiments of the present application, the dynamic load prediction method is that the corresponding base station cell predicts the load to come at the next moment based on the relevant load at the current moment and the relevant load at the previous moment, and through the load prediction, to know the changes, trends, and change amounts of the load that will occur in the corresponding load of the base station cell in the future; in this embodiment, the load prediction is not mainly aimed at making the predicted load close to the actual load, that is, not aimed at prediction accuracy, but aimed at combining the predicted load and the actual load to obtain the load change trend and change value in the future corresponding to the base station cell, so as to provide effective reference and guidance for operations such as energy saving, load control, and traffic guarantee of the base station cell; in this embodiment, the load prediction is performed by the target base station cell, that is, one of the multiple base station cells in the load migration direction.

[0040] In this embodiment, the considered high-speed private network is the high-speed private network corresponding to the linear high-speed railway line. Therefore, for each base station cell among the multiple base station cells under the high-speed private network, there is only one corresponding cross-layer same-direction group; for the base station cell at the central intersection of the T-shaped or cross-shaped high-speed railway line, there are two cross-layer same-direction groups, for the base station cell at the central intersection of the "Ж"-shaped high-speed railway line, there are three cross-layer same-direction groups, and for the base station cell at the central intersection of the star-shaped high-speed railway line, there are four cross-layer same-direction groups.

[0041] In this embodiment, before determining the cross-layer co-directional group corresponding to the target base station cell, multiple base station cells C set up on a section of the high-speed private network corresponding to the high-speed rail line will be analyzed. i For each base station cell in (i = 1, 2, ..., N), the primary neighbor cells are determined. This involves filtering and determining the primary neighbor cells of each base station cell from its left and right neighbor cells based on the handover statistics of neighbor cell pairs. For example, for base station cell C... i Based on hourly statistics of the number of handovers between itself and its neighboring cells over a period of time (e.g., 168 hours in a week), including handovers from neighboring cells and handovers from neighboring cells to base station cell C. i Based on the number of handovers, neighboring cells are divided into two categories according to a clustering algorithm. The neighboring cell with more handovers is designated as base station cell C. i The primary neighboring cell; of course, in some alternative implementations, this can be done by network planning and optimization personnel based on base station cell C. i And network engineering parameter information of neighboring cells, such as installation location, coverage direction, etc., or by referring to handover statistics, manually filter out base station cell C. i The nearest neighboring cells on both the left and right sides along the high-speed rail line are designated as base station cell C. i The primary neighboring cell; then, the base station cell C i The nearest neighboring cells on the left and right sides are designated as base station cell C. i The corresponding in-direction group is then identified. Based on the association between the in-direction groups of multiple base station cells, the cross-layer in-direction group corresponding to each base station cell is determined. For example, if C2 is the right neighbor cell of the in-direction group C1, and C3 is the right neighbor cell of the in-direction group C2, then C3 can be identified as the right neighbor cell of C1 that crosses layer 1. Similarly, the right neighbor cell C1 that crosses layer K can be obtained. k {C1, C2, ... C} K This forms a cross-layer, same-direction group C1 that spans layer K.

[0042] In this embodiment of the application, the cross-layer co-directional group corresponding to the determined target base station cell includes two cross-layer co-directional groups, left and right, for example: left neighbor cell co-directional group and right neighbor cell co-directional group. Both the left neighbor cell co-directional group and the right neighbor cell co-directional group have K-layer primary neighbor cells, where K is a preset value. For example, target base station cell C... i If the goal is to predict potential future loads earlier and sense loads over greater distances, then K can be set very large. However, this would increase the overhead and difficulty of transmitting load information across layers, as well as increase random uncertainty. Therefore, it is necessary to consider both the target requirements of load prediction and the overhead costs when setting the value of K, for example: K≤9.

[0043] In step S102, the real-time load at the current moment and the historical predicted load corresponding to the last moment are obtained, and the historical predicted load is used to represent the load migration change predicted at the last moment to be generated at the current moment.

[0044] In the embodiment of the present application, when predicting the load at the next moment, the current moment and the last moment are considered from the time dimension, wherein the last moment, the current moment and the next moment are separated by a set time period T, the time period T is determined according to the typical base station cell average coverage distance in the high-speed rail private network and the average speed of the high-speed train, and the value range can be 10-60 seconds, and preferably T=30 seconds; from the load information dimension, the real-time load and the predicted load are considered, and the predicted load refers to the load that will occur at the current moment which has been predicted at the last moment; from the base station cell object dimension, the base station cells participating in the load prediction include the target base station cell and one cross-layer same-direction group in two dimensions.

[0045] In the embodiment, for the load related to the target base station cell, the real-time load at the current moment and the predicted load corresponding to the last moment are included, because the predicted load is the prediction of the load change that may occur in the future of the base station cell, by the predicted load corresponding to the last moment, the difference between the real-time load at the current moment and the predicted load is determined to represent whether the load change trend of the target base station cell itself matches the predicted change trend.

[0046] It should be noted that when considering the time dimension, the time period T needs to be set according to the time situation, for example, if T is too small, the load prediction and the load information transmission will be very large, and if T is too large, the high-speed train is likely to pass through the target base station cell and some of its associated neighboring cells, so that the corresponding base station cell cannot sample the real actual load, and the target base station cell cannot achieve effective load prediction; when considering the real-time load, the real-time load of the base station cell located in the forward direction of the load migration direction needs to be considered, so the real-time load of the base station cell is the real-time load checked minus the load migrated into the base station cell by the corresponding base station cell at the last moment, so that the real-time load at the current moment matches the specific load migration situation.

[0047] In step S103, the cross-layer neighboring cell load data transmitted by the left neighboring cell same-direction group and the right neighboring cell same-direction group of the cross-layer same-direction group are received, wherein the corresponding cross-layer neighboring cell load data includes historical load data and real-time load data, the historical load data is used to represent the predicted load of the K-layer main neighboring cell of the corresponding neighboring cell same-direction group at the last moment, and the real-time load data is used to represent the actual load of the K-layer main neighboring cell of the corresponding neighboring cell same-direction group at the current moment.

[0048] In this embodiment, when the target base station cell C i When located at the beginning or end of a high-speed private network, the target base station cell C i The corresponding cross-layer same-direction group has only one of the left neighbor cell same-direction group and the right neighbor cell same-direction group. At this time, the target base station cell C i Only forward load forecasting or reverse load forecasting is performed.

[0049] Step S104: Using a preset load prediction model, load prediction is performed based on real-time load, historical predicted load, and cross-layer neighbor cell load data to obtain the forward predicted load and reverse predicted load corresponding to the target base station cell, and the corresponding load prediction result is determined based on the forward predicted load and reverse predicted load.

[0050] In this embodiment of the application, the preset load prediction model sets a corresponding operation function, and based on the operation function, performs forward load prediction and reverse load prediction on the real-time load, historical predicted load and corresponding cross-layer neighbor cell load data of the target base station cell, respectively. After obtaining the corresponding forward predicted load and reverse predicted load, the total predicted load is calculated, and then the corresponding prediction result is determined based on the total predicted load.

[0051] Through the steps S101 to S104, in the plurality of base station cells located on the preset load migration line, the cross-layer co-directional group is determined, the base station cells at both ends of the cross-layer co-directional group are the Kth layer main adjacent areas corresponding to the target base station cell, K is a preset value; the real-time load at the current time and the historical predicted load corresponding to the last time are obtained, the historical predicted load is used to represent the load migration change predicted at the last time at the current time; the cross-layer adjacent area load data transmitted by the left adjacent area co-directional group and the right adjacent area co-directional group of the cross-layer co-directional group is received, the corresponding cross-layer adjacent area load data includes the historical load data and the real-time load data, the historical load data is used to represent the predicted load of the Kth layer main adjacent area of the corresponding adjacent area co-directional group at the last time, and the real-time load data is used to represent the actual load of the Kth layer main adjacent area of the corresponding adjacent area co-directional group at the current time; the preset load prediction model is used to perform load prediction based on the real-time load, the historical predicted load and the cross-layer adjacent area load data, to obtain the forward predicted load and the reverse predicted load corresponding to the target base station cell, and to determine the corresponding load prediction result based on the forward predicted load and the reverse predicted load, thereby solving the problem that the load prediction method in the related art cannot be applied to the high-speed mobile communication scene without accurate time period rules, such as user high-speed movement, load burst, and the like, and achieving the beneficial effects that the base station cell can detect and sense the user load and the change trend of the adjacent area coming from a farther distance at an accurate time point in advance, and the prediction load and the actual load of the base station cell are combined to provide effective reference and guidance for the energy saving, load control, traffic guarantee and the like of the base station cell.

[0052] In some embodiments, the preset load prediction model is used to perform load prediction based on the real-time load, the historical predicted load and the cross-layer adjacent area load data, to obtain the forward predicted load and the reverse predicted load corresponding to the target base station cell, and the following steps are used to achieve the above:

[0053] Step 21, the load prediction model is used to perform forward load calculation on the real-time load, the historical predicted load, the historical load data and the real-time load data corresponding to the left adjacent area co-directional group, to obtain the forward predicted load.

[0054] In the embodiment, it is defined that from left to right is forward and from right to left is reverse on the corresponding load migration line, and when performing the forward load calculation, the load transferred from the first layer main adjacent area in the right adjacent area co-directional group to the target base station cell at the last time also needs to be referred to, so as to determine the accurate real-time load of the target base station cell, so that the predicted load is closer to the change trend of the load, that is, the change trend of the target base station cell is better reflected.

[0055] Step 22, using the load prediction model, performing reverse load calculation on the real-time load, the historical predicted load, and the historical load data corresponding to the real-time load data of the same direction group of the right adjacent area, to obtain the reverse predicted load.

[0056] In this embodiment, when performing reverse load calculation, the load transferred by the first layer main adjacent area in the left adjacent area same direction group to the target base station cell at the previous time is also needed to be referred to, to determine the accurate real-time load of the target base station cell, so that the predicted load is closer to the trend of load change, that is, the change trend of the target base station cell is better reflected.

[0057] In order to obtain the accurate real-time load of each base station cell, in some embodiments, the following steps are also implemented:

[0058] Step 31, obtaining the first actual load of each base station cell located on the preset load migration line at the current time.

[0059] In this embodiment, the first actual load includes a part of the load migrated to the target base station cell at the previous time and a part of the load for corresponding load prediction; in order to make the predicted load more matched with the expected load prediction and the predicted load change trend more accurate, in this embodiment, the part of the load migrated to the target base station cell at the previous time is excluded, and then the corresponding prediction is performed. Step 32, determining the corresponding nearest left adjacent area and nearest right adjacent area of each base station cell, and respectively obtaining the corresponding actual load of the nearest right adjacent area and the nearest left adjacent area of the corresponding base station cell at the previous time, wherein the left end to the right end of the preset load migration line is set as a positive direction, the nearest left adjacent area is used to represent the first base station cell located on the left side of the corresponding base station cell in the positive direction, and the nearest right adjacent area is used to represent the first base station cell located on the right side of the corresponding base station cell in the positive direction.

[0060] In this embodiment, the nearest left adjacent area is the first base station cell located on the left side (rear) of the corresponding base station cell in the corresponding set positive direction, and the nearest right adjacent area is the first base station cell located on the right side (that is, the front) of the corresponding base station cell in the corresponding set positive direction. The nearest left adjacent area receives the reverse migrated load of the base station cell, and the nearest right adjacent area receives the forward migrated load of the base station cell.

[0061] Step 33, determining the first migration load of the nearest right adjacent area migrated to the base station cell at the previous time according to the first preset proportion based on the corresponding actual load of the nearest right adjacent area at the previous time, and determining the second migration load of the nearest left adjacent area migrated to the base station cell at the previous time according to the second preset proportion based on the corresponding actual load of the nearest left adjacent area at the previous time.

[0062] Step 34, determining the real-time load corresponding to the forward prediction of the base station cell at the current time according to the difference between the first actual load and the first migration load, and determining the real-time load corresponding to the reverse prediction of the base station cell at the current time according to the difference between the first actual load and the second migration load.

[0063] In the embodiment, the real load corresponding to the forward prediction or reverse prediction of the base station cell at the current time is determined based on the difference between the first actual load and the first migration load and the second migration load, respectively. In this way, the load predicted by the load prediction operation can more accurately reflect the trend of the load of the base station cell, for example, more accurately predict the time of arrival of the corresponding load of the base station cell.

[0064] In the above steps, the first actual load of each base station cell at the current time located on the preset load migration line is obtained; the nearest left neighbor area and the nearest right neighbor area corresponding to each base station cell are determined, and the actual load corresponding to the nearest right neighbor area and the nearest left neighbor area of the corresponding base station cell at the previous time is obtained respectively; the first migration load of the nearest right neighbor area to the base station cell at the previous time according to the first preset proportion is determined based on the actual load corresponding to the nearest right neighbor area at the previous time, and the second migration load of the nearest left neighbor area to the base station cell at the previous time according to the second preset proportion is determined based on the actual load corresponding to the nearest left neighbor area at the previous time; the real-time load corresponding to the forward prediction of the base station cell at the current time is determined according to the difference between the first actual load and the first migration load, and the real-time load corresponding to the reverse prediction of the base station cell at the current time is determined according to the difference between the first actual load and the second migration load, thereby realizing the accurate real-time load of each base station cell, and further providing effective reference and guidance for energy saving, load control, traffic protection and other operations of the base station cell.

[0065] In some embodiments, the load prediction model includes a forward prediction load calculation formula and a reverse prediction load calculation formula

[0066]

[0067]

[0068] wherein m represents the current time, m-1 represents the previous time, → represents the forward direction, ← represents the reverse direction, K represents the number of layers of the cross-layer same-direction group, p is the index of the target base station cell C i L is the index of one of the P cross-layer same-direction groups corresponding to the target base station cell C i (m) is the actual load of the target base station cell C i at the current time, respectively represent the target base station cell C i the forward predicted load corresponding to the current time and the last time, respectively represent the target base station cell C i the first layer main neighbor cell in the corresponding right neighbor co-site the actual load at the last time, and γ is a first preset proportion, the j-i layer main neighbor cell in the corresponding left neighbor co-site the actual load at the current time, the j-i layer main neighbor cell in the corresponding left neighbor co-site the corresponding nearest right neighbor cell the actual load at the last time, the j-i layer main neighbor cell in the corresponding left neighbor co-site the forward predicted load corresponding to the last time, respectively represent the target base station cell C i the backward predicted load at the current time and the last time, respectively represent the target base station cell C i the first layer main neighbor cell in the corresponding left neighbor co-site the actual load at the last time, the j-i layer main neighbor cell in the corresponding right neighbor co-site the actual load at the current time, the j-i layer main neighbor cell in the corresponding right neighbor co-site the corresponding nearest left neighbor cell the actual load at the last time, the j-i layer main neighbor cell in the corresponding right neighbor co-site the backward predicted load corresponding to the last time; α, β, γ are weight factors between [0, 1], α is the weight of the self load of the target base station cell C i , (1-α) is the weight of the corresponding load of the neighbor co-site, β is the weight of the real-time load of each base station cell, (1-β) is the weight of the predicted load of each base station cell, γ is a first preset proportion, and γ is the load ratio of the load of the corresponding nearest right neighbor cell migrated to the corresponding base station cell at the current time to the actual load of the corresponding nearest right neighbor cell at the last time in the forward load prediction, and 1-γ is a second preset proportion, and 1-γ is the load ratio of the load of the corresponding nearest left neighbor cell migrated to the corresponding base station cell at the current time to the actual load of the corresponding nearest left neighbor cell at the last time in the backward load prediction.

[0069] In the embodiment, the base station cell corresponds to the base station cell of a one-dimensional high-speed rail network, and the number P of the corresponding cross-layer co-site is 1; in the embodiment, the default values of α, β and γ are all 0.5.

[0070] In the present embodiment, the corresponding forward load prediction direction and reverse load prediction direction are represented by "→" and "←" respectively, thus the corresponding adjacent zone number of the main adjacent zone in the left adjacent zone co-directional group and right adjacent zone co-directional group for the forward load prediction and reverse load prediction respectively is also corresponding with directionality, through the adjacent zone number with direction data to indicate that the main adjacent zone of the cross-layer co-directional group is from the first layer main adjacent zone to the Kth layer main adjacent zone in the corresponding load prediction, for example: when j = i + 2, C represents the target base station cell i the 2nd layer main adjacent zone in the corresponding left adjacent zone co-directional group, and for example: when j = i + K, C represents the target base station cell i the Kth layer main adjacent zone in the corresponding right adjacent zone co-directional group, j only represents the label of the corresponding base station cell and does not represent the parameter size participating in the load prediction calculation.

[0071] It is to be noted that the time interval between the current time and the last time is the period T of load prediction and load information transmission, T is determined according to the average coverage distance of the typical base station cell under the high-speed rail private network and the average speed of the high-speed train, the value range of T is approximately 10-60 seconds, and preferably T = 30 seconds; it is to be noted that if T is too small, the overhead of load prediction and load information transmission will be large, and if T is too large, the high-speed train is likely to pass through the target base station cell and part of the main adjacent zones associated with the target base station cell, resulting in that the corresponding base station cell cannot sample the real-time load, and the target base station cell cannot realize effective load prediction.

[0072] It is to be further noted that in the case that T is set to match the base station cell coverage distance and the high-speed train speed, if α and β are larger, that is, the target base station cell self-load weight and the actual load weight are larger, the prediction load accuracy is higher, but the load detection distance is shorter, which is not conducive to sensing the load at a farther distance and predicting the load arrival earlier; if α and β are smaller, that is, the target base station cell self-load weight and the actual load weight are smaller, the load detection distance is longer, but the prediction load accuracy is lower; in the present embodiment, the purpose of load prediction is not to expect the predicted load to be as close to the actual load as possible, that is, the core target is not to improve the prediction accuracy, but to combine the predicted load and the actual load for comprehensive consideration to provide effective reference and guidance for the energy saving, load control, traffic guarantee and other operations of the base station cell.

[0073] In some embodiments, the corresponding load prediction result is determined based on the forward predicted load and the reverse predicted load, which is realized through the following steps:

[0074] Step 41, sum the forward predicted load and the reverse predicted load to obtain a first predicted total load corresponding to the target base station on the preset load migration line.

[0075] In the embodiment, the first predicted total load is calculated by using the following formula:

[0076]

[0077] wherein p is the index of a cross-layer homing group in P cross-layer homing groups corresponding to the target base station cell, represents the first predicted total load, represents the forward predicted load, represents the reverse predicted load.

[0078] Step 42, accumulate the first predicted total loads corresponding to multiple preset load migration lines to generate a total predicted load, wherein the load prediction result comprises the total predicted load.

[0079] In the embodiment, the total predicted load is calculated by using the following formula:

[0080]

[0081] p is the index of a cross-layer homing group in P cross-layer homing groups corresponding to the target base station cell, is the total predicted load, represents the first predicted total load.

[0082] In the embodiment, for a one-letter-shaped high-speed rail network, the total predicted load is equal to the first predicted total load, and for a cross-shaped or a Chinese character-shaped high-speed rail network, the total predicted load corresponding to the target base station cell located at the intersection of multiple load migration lines is the sum of multiple first predicted total loads.

[0083] By summing the forward predicted load and the reverse predicted load in the above steps, the first predicted total load corresponding to the target base station on the preset load migration line is obtained; by accumulating the first predicted total loads corresponding to multiple preset load migration lines, the total predicted load is generated, wherein the load prediction result comprises the total predicted load, and the total load predicted by the target base station cell at the next time is realized.

[0084] In some embodiments, after determining the corresponding load prediction result, the following steps are further implemented:

[0085] Step 51, respectively acquire the total predicted load corresponding to the current time and the real-time load corresponding to the current time, and respectively acquire the historical total predicted load corresponding to the last time and the historical actual load corresponding to the last time.

[0086] In the embodiment, the predicted load of the target base station cell at the current time is obtained by aggregating the forward predicted load and the reverse predicted load and summing the predicted total loads corresponding to the plurality of load migration lines.

[0087] In step 52, a difference function of the load prediction is generated based on at least two of the actual load corresponding to the current time, the total predicted load corresponding to the current time, the historical total predicted load corresponding to the last time, and the historical actual load corresponding to the last time, and a calculation value corresponding to the difference function is determined, wherein the difference function is used to represent the prediction performance of the load prediction.

[0088] In the embodiment, the performance of the different load prediction sensitivity parameters is determined by a preset calculation function for optimizing the parameters of the load prediction model, for example, the difference between the real-time load corresponding to the current time and the total predicted load corresponding to the current time can be used as the difference function.

[0089] In the embodiment, the generation of the difference function of the load prediction refers to defining the preset difference function by using at least two of the actual load corresponding to the current time, the total predicted load corresponding to the current time, the historical total predicted load corresponding to the last time, and the historical actual load corresponding to the last time as corresponding parameters, and the determination of the calculation value corresponding to the difference function refers to the function value obtained by inputting the above four parameters into the corresponding defined difference function, which is the corresponding calculation value.

[0090] In step 53, it is judged whether the calculation value corresponding to the difference function is greater than a preset threshold, and the weight factor corresponding to the load prediction model is adjusted according to the judgment result, wherein the weight factor corresponding to the load prediction model includes the weight α of the target base station cell C i own load, the weight β of the actual load of each base station cell, and the first preset proportion γ.

[0091] In the embodiment, whether the load prediction model is sensitive to the load is determined by judging whether the calculation value corresponding to the difference function is greater than the preset threshold (that is, the corresponding threshold), specifically, when the difference function value is positive and exceeds the preset threshold (the upper limit of the set threshold), it means that the prediction is too fast and sensitive, and α and β are increased by a preset step size to reduce the weights of the neighbor load and the predicted load, on the contrary, when the difference function value is negative and lower than the preset threshold (the lower limit of the set threshold), it means that the prediction is too slow and sluggish, and α and β are decreased by a preset step size to increase the weights of the neighbor load and the predicted load; in the embodiment, the default value of the preset step size is 0.05.

[0092] The total predicted load corresponding to the current moment and the real-time load corresponding to the current moment are respectively obtained in the above steps, and the historical total predicted load corresponding to the previous moment and the historical actual load corresponding to the previous moment are respectively obtained; based on at least two of the actual load corresponding to the current moment, the total predicted load corresponding to the current moment, the historical total predicted load corresponding to the previous moment, and the historical actual load corresponding to the previous moment, a difference function for load prediction is generated, and a calculation value corresponding to the difference function is determined, wherein the difference function is used to represent the prediction performance of the load prediction; it is judged whether the calculation value corresponding to the difference function is greater than a preset threshold, and the weight factor corresponding to the load prediction model is adjusted according to the judgment result, which realizes real-time adjustment of the model parameters of the load prediction model according to the prediction result, so as to realize the accuracy of the load prediction, and prolong the detection sensing distance corresponding to the load prediction.

[0093] In some optional embodiments, the difference function includes one of the following: a difference value between the actual load corresponding to the current moment and the historical total predicted load corresponding to the previous moment, a ratio of the difference value and the actual load corresponding to the current moment, a difference value between the historical total predicted load corresponding to the previous moment and the total predicted load corresponding to the current moment, and a load difference value generated by weighting the difference value between the total predicted load corresponding to the current moment and the historical total predicted load corresponding to the previous moment, and the difference value between the actual load corresponding to the current moment and the historical actual load corresponding to the previous moment.

[0094] In the embodiment, each base station cell supports optimization of the corresponding weight factor after determining the corresponding load prediction result, that is, the model parameters of the load prediction model are maintained and optimized, and a personalized dynamic load prediction model suitable for itself is formed. In some optional embodiments, the corresponding model parameters are optimized in the following manner, that is, the first difference function is defined and generated in the following manner:

[0095] Method one: define the difference function dif(m) of the following formula:

[0096]

[0097] Or define the difference function dif(m) of the following formula:

[0098]

[0099] Wherein, L i (m) is the target base station cell C i The actual load corresponding to the current moment, L

[0100] Mode two

[0101] The difference function dif(m) is defined as follows

[0102]

[0103] wherein, L represents the total predicted load corresponding to the previous moment,

[0104] Or the difference function dif(m) is calculated as follows

[0105] Dif(m)=L i (m)-L i (m-1)

[0106] wherein, L i (m) is the actual load corresponding to the current moment, L i (m-1) represents the historical actual load corresponding to the previous moment.

[0107] Or the difference function dif(m) is further defined as follows

[0108]

[0109] wherein, L represents the total predicted load corresponding to the previous moment, L i (m) is the actual load corresponding to the current moment, L i (m-1) represents the historical actual load corresponding to the previous moment, μ is the weight of the predicted load of the current moment and the previous moment, and the value range of μ is [0, 1], and the default value is 0.5.

[0110] In mode two, the model parameters are adjusted in real time according to the changes of the actual load and the predicted load, the prediction accuracy and the longer detection sensing distance are considered, and when it is learned that the predicted load is increasing, i.e., the train load is approaching, α and β are increased by a preset step, and the target base station cell itself load and the actual load are further focused on, so as to improve the prediction accuracy, and when the predicted load is decreasing, i.e., the train load is away, then α and β are reduced by a preset step, and the adjacent zone load and the predicted load are further focused on, so as to expand the distance of the detection sensing base station adjacent zone load.

[0111] Mode three

[0112] The center of gravity CoL(m,p) of the overall load of the preset high-speed private network section is calculated according to the following formula:

[0113]

[0114] The high-speed private network section is composed of the target base station cell C i and the 2K-layer main adjacent area of the corresponding K-layer same-direction group, CoL(m,p) represents the center of gravity of the overall load of the corresponding high-speed private network section at m moment, i represents the real-time load of each base station cell C j at m moment, and p is the label of the K-layer same-direction group corresponding to the target base station cell C i ; then, the center of gravity of the overall load is taken as a difference variable function.

[0115] In the embodiment, the center of gravity of the overall load of the high-speed private network section is determined by using the actual load of each base station adjacent area in the corresponding K-layer same-direction group of the high-speed private network section, the moving direction of the train on the corresponding high-speed rail line of the high-speed private network and the corresponding load moving direction are determined by the change of the position of the center of gravity of the overall load, when it is determined that the position of the center of gravity of the overall load of the high-speed private network section moves from left to right in a positive direction, then γ is reduced by a preset step, otherwise, γ is increased, in the embodiment, the preset step can be 0.05, and the adjustment of γ cannot exceed the preset value.

[0116] In some embodiments, the K-layer adjacent area load data transmitted by the left adjacent area same-direction group and the right adjacent area same-direction group of the K-layer same-direction group is received, and the following steps are implemented:

[0117] Step 61, after detecting the first target load data from the first adjacent area group load data received by the nearest left adjacent area, the first target load data and the actual load corresponding to the current moment and the historical predicted load corresponding to the last moment of the nearest left adjacent area are synthesized into the first K-layer adjacent area load data according to a preset format, and the first K-layer adjacent area load data is transmitted to the corresponding target base station cell, wherein the first adjacent area group load data is the K-layer adjacent area load data corresponding to the left adjacent area same-direction group of the nearest left adjacent area, and the first target load data includes the K-layer adjacent area load data corresponding to all the main adjacent areas excluding the K-layer main adjacent area of the nearest left adjacent area in the left adjacent area same-direction group of the nearest left adjacent area.

[0118] ​In some optional embodiments, the target base station cell is set as C8, the left-neighbor same-direction group corresponding to the target base station cell C8 is set as {C2, C3, C4, C5, C6, C7}, the nearest left-neighbor of the target base station cell C8 is C7, and the process of receiving the cross-layer neighbor load data transferred by the left-neighbor same-direction group of the target base station cell C8 includes the following steps: the nearest left-neighbor C7 receives the cross-layer neighbor load data M0 of its own left-neighbor same-direction group {C1, C2, C3, C4, C5, C6}, M0={L(1), L(2), L(3), L(4), L(5), L(6)}, the nearest left-neighbor C7 deletes the load L(1) corresponding to the Kth layer main neighbor C1 in its own left-neighbor same-direction group from M0, adds the load L(7) corresponding to the target base station cell C8 at the current time, and forms the corresponding first cross-layer neighbor load data, i.e., M1={L(2), L(3), L(4), L(5), L(6), L(7)}, and then transmits the first cross-layer neighbor load data to the target base station cell C8.

[0119] Step 62, after detecting the second target load data from the second neighbor group load data received by the nearest right-neighbor, the nearest right-neighbor synthesizes the second cross-layer neighbor load data from the second target load data, the actual load corresponding to the current time, and the historical predicted load corresponding to the last time according to a preset format, and transmits the second cross-layer neighbor load data to the corresponding target base station cell, wherein the second neighbor group load data includes the cross-layer neighbor load data corresponding to the right-neighbor same-direction group of the nearest right-neighbor, and the second target load data includes the cross-layer neighbor load data corresponding to all the main neighbors in the right-neighbor same-direction group of the nearest right-neighbor except the Kth layer main neighbor;

[0120] In some optional embodiments, the target base station cell is set as C8, the right-neighbor same-direction group corresponding to the target base station cell C8 is set as {C9, C 10 , C 11 , C 12 , C 13 , C 14}, the nearest right-neighbor of the target base station cell C8 is C9, and the process of receiving the cross-layer neighbor load data transferred by the right-neighbor same-direction group of the target base station cell C8 includes the following steps: the nearest right-neighbor C9 receives the cross-layer neighbor load data M2 of its own left-neighbor same-direction group {C 10 , C 11 , C 12 , C 13 , C 14 , C 15}, M2={L(10), L(11), L(12), L(13), L(14), L(15)}, the nearest right-neighbor C9 deletes the load L(10) corresponding to the Kth layer main neighbor C 15After the corresponding load L(15) is obtained, the corresponding load L(9) of the target base station cell C8 at the current time is added to form the corresponding second cross-layer neighboring cell load data, i.e., M3={L(9), L(10), L(11), L(12), L(13), L(14)}, and is transmitted to the target base station cell C8.

[0121] In the embodiment, the nearest left neighboring cell and the nearest right neighboring cell form the left and right main neighboring cells of the corresponding base station cell; in the embodiment, the load information transmitted by the base station cell is the real-time load at the current time, the historical predicted load at the previous time, and the load information of the main neighboring cells in the corresponding cross-layer co-directional group.

[0122] In the embodiment, before the corresponding load data is prepared to be transmitted to the corresponding target base station cell, the nearest left neighboring cell first receives the corresponding cross-layer neighboring cell load data transmitted by the corresponding left neighboring cell co-directional group of the nearest left neighboring cell; at this time, the Kth layer main neighboring cell in the corresponding left neighboring cell co-directional group of the nearest left neighboring cell does not belong to the main neighboring cells in the corresponding left neighboring cell co-directional group of the target base station cell, and thus the load data transmitted by the nearest left neighboring cell does not include the load data of the Kth layer main neighboring cell in the corresponding left neighboring cell co-directional group of the nearest left neighboring cell; before the corresponding load data is prepared to be transmitted to the corresponding target base station cell, the nearest right neighboring cell first receives the corresponding cross-layer neighboring cell load data transmitted by the corresponding right neighboring cell co-directional group of the nearest right neighboring cell; at this time, the Kth layer main neighboring cell in the corresponding right neighboring cell co-directional group of the nearest right neighboring cell does not belong to the main neighboring cells in the corresponding right neighboring cell co-directional group of the target base station cell, and thus the load data transmitted by the target base station cell does not include the load data of the Kth layer main neighboring cell in the corresponding right neighboring cell co-directional group of the nearest right neighboring cell.

[0123] In the embodiment, after the load of the Kth layer main neighboring cell in the corresponding left neighboring cell co-directional group of the nearest left neighboring cell / the Kth layer main neighboring cell in the corresponding right neighboring cell co-directional group of the nearest right neighboring cell is deleted from the neighboring cell group load data stored in the preset format, the real-time load at the current time and the historical predicted load at the previous time of the nearest left neighboring cell / the nearest right neighboring cell are supplemented, and the corresponding load data in the preset format is generated and transmitted to the target base station cell.

[0124] The base station cell transmits the load information in the manner of directional relay through the above steps.

[0125] In some embodiments, among a plurality of base station cells located on a preset load migration line, the cross-layer co-directional group is determined, and the following steps are implemented:

[0126] Step 71, obtain the handover frequency granularity data corresponding to each base station cell on the preset load migration line, and select the neighbor area with the maximum handover frequency granularity data as the main neighbor area of the base station cell; wherein the handover frequency granularity data is used to represent the number of user handovers between the base station cell and the neighbor area corresponding to the base station cell within a preset time.

[0127] Step 72, select the left main neighbor area and the right main neighbor area corresponding to each base station cell as the base station cell corresponding to the base station cell co-directional group.

[0128] Step 73, take the target base station cell as the starting point, select K continuous base station cells corresponding to the base station cell co-directional group along the forward direction and the reverse direction of the preset load migration line, and perform duplicate removal processing on the base station cells in the selected base station cell co-directional group to obtain a cross-layer co-directional group.

[0129] In this embodiment, the base station cell C i According to the hour granularity statistical data of the "neighbor area pair" handover times of the base station cell C i and its neighbor areas in a period of time (such as 168 hours corresponding to 1 week), including the number of handovers from the base station cell C i to the neighbor area and the number of handovers from the neighbor area to the base station cell C i , the handover times are divided into two categories according to the clustering algorithm, and the neighbor area with more handover times is selected as the main neighbor area of the base station cell C i ; of course, the network planning and optimization personnel can also manually select the left and right nearest neighbor areas of the base station cell C i along the high-speed rail line direction as the main neighbor area of the base station cell C i based on the network engineering parameter information of the base station cell C i and the neighbor area, such as installation location, coverage direction, etc., or by referring to the handover statistical data; then, the left and right nearest neighbor areas of the base station cell C k are selected as the co-directional group corresponding to the base station cell C K , and then the association relationship between the co-directional groups corresponding to multiple base station cells is determined to determine the cross-layer co-directional group corresponding to each base station cell, for example, if C2 is the right neighbor area of the co-directional group of C1, and C3 is the right neighbor area of the co-directional group of C2, then C3 can be associated and confirmed as the right neighbor area of C1 across 1 layer, and so on. The right neighbor area C k of C1 across K layers can be obtained. {C1, C2,... C K} forms a cross-layer co-directional group of C1 across K layers.

[0130] In this embodiment, the main neighbor areas of the base station cell are matched and selected as the left main neighbor area and the right main neighbor area according to the load migration space-time feature pattern on the same direction line, and are divided into the same direction line group, that is, the co-directional group.

[0131] Specifically, for the target base station cell Ci For its primary and neighboring cells, select the time periods with the highest number of handovers from the handover statistics at the hourly granularity of one week. For example, select two peak time periods with the highest number of handovers each day, resulting in 14 hours per week. Obtain user count statistics D2 at the granularity of 10-60 seconds (e.g., 30 seconds) within these time periods. Use a clustering algorithm to divide D2 into two categories, setting the load of the category with more users to 1 and the load of the category with fewer users to 0. Then, search and filter for primary neighboring cells 1 and 2 that meet the load migration spatiotemporal characteristics shown in Tables 1 and 2 (meeting one of them is sufficient). Thus, the selected primary neighboring cells 1 and 2 are used as the target base station cell C. i The left and right neighboring areas.

[0132] Load Primary neighbor 1 Target base station cell Ci Primary neighbor 2 Time 1 1 0 0 Time 2 0 1 0 Time 3 0 0 1

[0133] Table 1

[0134] Load Primary neighbor 1 Target base station cell Ci Primary neighbor 2 Time 1 0 0 1 Time 2 0 1 0 Time 3 1 0 0

[0135] Table 2

[0136] In this embodiment, the cross-layer co-directional groups of each base station cell are determined based on multiple associated co-directional groups. Specifically, if C2 is the right neighbor cell of the C1 co-directional group and C3 is the right neighbor cell of the C2 co-directional group, then C3 can be associated and confirmed as the right neighbor cell of C1 that crosses layer 1. Similarly, the right neighbor cell C1 that crosses layer K can be obtained. k {C1, C2, ... C K This forms a cross-layer unidirectional group C1 spanning layer K; in this embodiment of the application, the cross-layer unidirectional group corresponding to the determined target base station cell includes two cross-layer unidirectional groups, for example: a left neighbor cell unidirectional group and a right neighbor cell unidirectional group, wherein both the left neighbor cell unidirectional group and the right neighbor cell unidirectional group have a primary neighbor cell at layer K, where K is a preset value, for example: target base station cell C i To predict potential future loads earlier and sense loads over longer distances, K can be set very large. However, this increases the overhead and difficulty of transmitting load information across layers, and also increases random uncertainty. Therefore, the value of K needs to be set while balancing the target requirements of load forecasting with the overhead costs. For example, K ≤ 9. Figure 2 As shown, {C1, C2, C3, C4, C5, C6} are the left-neighbor regions of C7 that span 6 layers in the same direction, and {C8, C9, C6} are the left-neighbor regions of C7 that span 6 layers in the same direction. 10 C 11 C 12 C 13} This means that C7 spans 6 layers and is located in the same direction to the right neighboring area. Load information can be relayed to the left or right across 6 layers.

[0137] The following gives examples 1 to 6 to illustrate the load prediction of the embodiments of the present application as follows:

[0138] Before a specific description, a segment containing 20 base station cells connected to form a continuous coverage of a linear high-speed rail private network is given, adjacent base station cells are mutually primary neighbor cells, and are mutually left or right neighbors, and form a cross-layer same-direction group; the actual load conditions are shown in the corresponding graphs of each embodiment, and the values represent the load of the number of users; the embodiments set several typical load distribution, migration, and change scenarios; in the embodiments 1 to 6 of the present application, K = 6, α = 0.5, β = 0.5, γ = 0.5, and the optimization of the model parameters of the load prediction model adopts the following formula: And μ = 0.5.

[0139] Example 1

[0140] The single-column high-speed rail is simulated to move forward from left to right, and the actual load and the predicted load are shown in Table 3 and Table 4 respectively. From the results of the predicted load, the prediction model can detect and sense the upcoming load and change trend in advance, and combined with the actual load and the predicted load, the base station cell can prepare energy saving or load control operations in advance.

[0141]

[0142]

[0143] Table 3

[0144] T [C1] [C2] [C3] [C4] [C5] [C6] [C7] [C8] [C9] [C 10 ]]> [C 11 ]]> [C 12 ]]> [C 13 ]]> [C 14 ]]> [C 15 ]]> T0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 T1 73 18 11 7 5 3 2 2 1 1 0 0 0 0 0 T2 25 63 24 16 11 7 5 3 2 2 1 0 0 0 0 T3 12 16 60 23 16 11 7 5 3 2 2 1 0 0 0 T4 10 11 15 60 23 16 11 8 5 4 3 2 1 0 0 T5 8 9 11 15 60 24 17 12 8 6 4 3 2 2 0 T6 6 8 9 11 15 60 24 17 12 9 6 4 3 3 2 T7 5 6 8 9 11 15 60 24 17 12 9 6 5 3 3 T8 5 5 6 8 9 11 15 60 24 17 12 9 6 5 3 T9 4 5 5 6 8 9 11 15 60 24 17 13 9 7 5 T10 3 4 5 5 6 8 9 11 15 60 24 17 13 9 7 T11 2 3 4 5 5 6 8 9 11 15 60 24 17 13 9 T12 2 2 3 4 5 5 6 8 9 11 15 60 24 17 13 T13 2 2 2 3 4 5 5 6 8 9 11 15 60 24 17 T14 2 2 2 2 3 4 5 5 6 8 9 11 15 60 24 T15 1 2 2 2 2 3 4 5 5 6 8 9 11 15 60

[0145] Table 4

[0146] Example 2

[0147] The single-column high-speed rail is simulated to move backward from right to left. The actual load and the predicted load are shown in Table 5 and Table 6 respectively. From the results of the predicted load, the prediction model can detect and sense the upcoming load and change trend in advance, and combined with the actual load and the predicted load, the base station cell can prepare energy saving or load control operations in advance.

[0148]

[0149]

[0150] Table 5

[0151] T [C1] [C2] [C3] [C4] [C5] [C6] [C7] [C8] [C9] [C 10 ]]> [C 11 ]]> [C 12 ]]> [C 13 ]]> C 14 ]]> C 15 ]]> T0 0 0 0 0 0 0 2 2 3 4 6 9 12 17 25 T1 0 0 0 0 0 2 3 3 4 6 9 13 18 25 60 T2 0 0 0 0 2 3 3 5 7 9 13 18 25 60 14 T3 0 0 1 2 3 4 5 7 9 13 18 25 60 14 9 T4 0 1 2 3 4 5 7 10 13 18 25 60 14 9 7 T5 1 2 3 4 5 7 10 13 18 25 60 14 9 7 6 T6 2 3 4 5 7 10 13 18 26 60 14 9 7 6 5 T7 3 4 5 7 10 13 19 26 60 14 9 7 6 5 4 T8 4 5 7 10 14 19 26 60 14 9 7 6 5 4 3 T9 5 7 10 14 19 26 60 14 9 7 6 5 4 3 3 T10 7 10 14 19 26 60 14 9 7 6 5 4 3 3 2 T11 10 14 19 26 60 14 9 7 6 5 4 3 3 2 2 T12 14 19 26 60 14 9 7 6 5 4 3 3 2 2 1 T13 19 26 60 14 9 7 6 5 4 3 3 2 2 1 1 T14 26 60 14 9 7 6 5 4 3 3 2 2 1 1 1 T15 60 14 9 7 6 5 4 3 3 2 2 1 1 1 1

[0152] Table 6

[0153] Example 3

[0154] The double high-speed rails are simulated to move from left to right and from right to left respectively. The actual load and the predicted load are shown in Table 7 and Table 8 respectively. From the results of the predicted load, the prediction model can detect and sense the upcoming load and change trend in advance. In combination with the actual load and the predicted load, the base station cell can prepare energy saving or load control operation in advance.

[0155]

[0156]

[0157] Table 7

[0158] T [C1] [C2] [C3] [C4] [C5] [C6] [C7] [C8] [C9] C 10 ]]> C 11 ]]> C 12 ]]> C 13 ]]> C 14 ]]> C 15 ]]> T0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 T1 11 7 5 3 2 2 1 1 1 1 2 2 3 5 7 T2 24 16 11 7 5 3 3 3 3 3 3 5 7 11 16 T3 60 23 16 11 7 5 5 4 4 5 5 7 11 16 24 T4 15 60 23 16 12 9 8 6 6 8 9 12 17 24 59 T5 11 15 60 24 18 14 11 10 10 11 15 19 25 60 14 T6 9 12 16 61 26 20 17 15 15 17 21 28 61 14 9 T7 8 10 13 18 63 29 24 22 22 24 30 63 16 11 8 T8 7 10 12 15 20 67 33 30 31 34 66 18 12 10 8 T9 7 9 11 14 18 25 73 42 43 72 23 16 12 9 8 T10 8 9 12 15 19 25 34 85 84 31 22 17 13 10 8 T11 8 10 12 16 21 28 37 73 73 33 25 19 14 11 8 T12 8 11 14 19 25 33 69 15 13 67 30 22 17 13 9 T13 10 13 17 23 31 66 20 9 9 19 64 28 21 15 12 T14 12 16 22 29 64 18 10 13 13 11 17 63 26 20 14 T15 15 20 28 63 16 10 11 12 12 12 11 16 62 25 19

[0159] Table 8

[0160] Example 4

[0161] The double high-speed rails are simulated to move from left to right and from right to left respectively. The actual load and the predicted load are shown in Table 7 and Table 8 respectively. From the results of the predicted load, the prediction model can detect and sense the upcoming load and change trend in advance. In combination with the actual load and the predicted load, the base station cell can prepare energy saving or load control operation in advance.

[0162]

[0163]

[0164] Table 9

[0165] T [C1] [C2] [C3] [C4] [C5] [C6] [C7] [C8] [C9] [C 10 ]]> [C 11 ]]> [C 12 ]]> [C 13 ]]> [C 14 ]]> [C 15 ]]> T0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 T1 74 19 13 10 8 8 10 13 19 74 18 11 7 5 3 T2 25 65 26 18 14 12 12 14 14 19 58 20 13 9 6 T3 13 17 62 26 20 16 14 14 13 12 16 60 22 15 10 T4 10 12 17 62 27 21 17 15 13 13 11 16 60 23 16 T5 9 11 13 18 63 28 21 17 15 13 13 11 16 61 23 T6 7 9 11 13 18 63 28 22 18 15 14 13 11 16 61 T7 6 8 9 11 13 18 63 28 22 18 15 15 13 11 16 T8 6 7 8 9 11 13 18 63 28 22 18 15 14 13 11 T9 5 6 7 8 9 11 13 18 63 28 22 18 15 14 13 T10 4 5 6 7 8 9 11 13 18 63 28 22 18 16 14 T11 3 4 5 6 7 8 9 11 13 18 63 28 22 18 16 T12 3 3 4 5 6 7 8 9 11 13 18 63 27 21 17 T13 2 3 3 4 5 6 7 8 9 11 13 17 62 27 20 T14 2 2 3 3 4 5 6 7 8 9 11 13 17 62 26 T15 2 2 2 3 3 4 5 6 7 8 9 11 13 16 61

[0166] Table 10

[0167] Example 5

[0168] The double high-speed rails are simulated to move from left to right and from right to left respectively. The actual load and the predicted load are shown in Table 7 and Table 8 respectively. From the results of the predicted load, the prediction model can detect and sense the upcoming load and change trend in advance. In combination with the actual load and the predicted load, the base station cell can prepare energy saving or load control operation in advance.

[0169] T [C1] [C2] [C3] [C4] [C5] [C6] [C7] [C8] [C9] [C 10 ]]> [C 11 ]]> [C 12 ]]> [C 13 ]]> [C 14 ]]> [C 15 ]]> T0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 T1 100 0 0 0 0 0 0 0 0 0 0 0 0 0 0 T2 0 100 0 0 0 0 0 0 0 0 0 0 0 0 0 T3 0 0 100 0 0 0 0 0 0 0 0 0 0 0 0 T4 0 0 0 100 0 0 0 0 0 0 100 0 0 0 0 T5 0 0 0 0 100 0 0 0 0 100 0 0 0 0 0 T6 0 0 0 0 0 100 0 0 100 0 0 0 0 0 0 T7 0 0 0 0 0 0 200 0 0 0 0 0 0 0 0 T8 0 0 0 0 0 100 0 100 0 0 0 0 0 0 0 T9 0 0 0 0 100 0 0 0 100 0 0 0 0 0 0 T10 0 0 0 100 0 0 0 0 0 100 0 0 0 0 0 T11 0 0 100 0 0 0 0 0 0 0 100 0 0 0 0 T12 0 100 0 0 0 0 0 0 0 0 0 100 0 0 0 T13 100 0 0 0 0 0 0 0 0 0 0 0 100 0 0 T14 0 0 0 0 0 0 0 0 0 0 0 0 0 100 0 T15 0 0 0 0 0 0 0 0 0 0 0 0 0 0 100

[0170] Table 11

[0171]

[0172]

[0173] Table 12

[0174] Embodiment 6

[0175] In the process of simulating the forward movement of the single-track high-speed rail from left to right, the number of users is randomly increased or decreased to simulate the change in the number of users caused by the passengers getting on and off the train at some intermediate cells. The actual load and the predicted load are shown in Table 13 and Table 14, respectively. From the results of the predicted load, the prediction model can detect and sense the upcoming load and the trend of change in advance. In combination with the actual load and the predicted load, the cell can prepare for energy saving or load control operations in advance.

[0176] T [C1] [C2] [C3] [C4] [C5] [C6] [C7] [C8] [C9] [C 10 ]]> [C 11 ]]> [C 12 ]]> [C 13 ]]> [C 14 ]]> [C 15 ]]> T0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 T1 100 0 0 0 0 0 0 0 0 0 0 0 0 0 0 T2 0 100 0 0 0 0 0 0 0 0 0 0 0 0 0 T3 0 0 100 0 0 0 0 0 0 0 0 0 0 0 0 T4 0 0 0 100 0 0 0 0 0 0 0 0 0 0 0 T5 0 0 0 0 100 0 0 0 0 0 0 0 0 0 0 T6 0 0 0 0 0 50 0 0 0 0 0 0 0 0 0 T7 0 0 0 0 0 0 50 0 0 0 0 0 0 0 0 T8 0 0 0 0 0 0 0 50 0 0 0 0 0 0 0 T9 0 0 0 0 0 0 0 0 50 0 0 0 0 0 0 T10 0 0 0 0 0 0 0 0 0 200 0 0 0 0 0 T11 0 0 0 0 0 0 0 0 0 0 200 0 0 0 0 T12 0 0 0 0 0 0 0 0 0 0 0 200 0 0 0 T13 0 0 0 0 0 0 0 0 0 0 0 0 200 0 0 T14 0 0 0 0 0 0 0 0 0 0 0 0 0 100 0 T15 0 0 0 0 0 0 0 0 0 0 0 0 0 0 100

[0177] Table 13

[0178]

[0179]

[0180] Table 14

[0181] The embodiments of the present application also provide a dynamic load prediction device, which is used to implement the above-mentioned embodiments and preferred embodiments, and will not be described again. The terms "module", "unit", "sub-unit" and the like used below can be a combination of software and / or hardware that can implement a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and is conceived.

[0182] Figure 3 The structure block diagram of the dynamic load prediction device provided by the embodiments of the present application is shown in FIG. 1, which comprises: Figure 3

[0183] The determining module 31 is used to determine the cross-layer same-direction group in the multiple base station cells located on the preset load migration route, wherein the base station cells located at both ends of the cross-layer same-direction group are the Kth layer main adjacent areas corresponding to the target base station cell, and K is a preset value.

[0184] The acquiring module 32 is coupled with the determining module 31 and is used to acquire the real-time load at the current time and the historical predicted load corresponding to the last time, wherein the historical predicted load is used to represent the load migration change predicted at the last time.

[0185] ​The receiving module 33 is coupled to the obtaining module 32 and is configured to receive cross-layer adjacent zone load data transmitted by the left adjacent zone co-directional group and the right adjacent zone co-directional group of the cross-layer co-directional group, wherein the corresponding cross-layer adjacent zone load data each includes historical load data and real-time load data, the historical load data is used to represent the predicted load of the K-layer main adjacent zone of the corresponding adjacent zone co-directional group at the previous time, and the real-time load data is used to represent the actual load of the K-layer main adjacent zone of the corresponding adjacent zone co-directional group at the current time.

[0186] The prediction module 34 is coupled to the receiving module 33 and is configured to perform load prediction based on the real-time load, the historical predicted load, and the cross-layer adjacent zone load data by using a preset load prediction model, to obtain the forward predicted load and the reverse predicted load corresponding to the target base station cell, and to determine the corresponding load prediction result based on the forward predicted load and the reverse predicted load.

[0187] By using the dynamic load prediction device provided in the embodiments of the present application, the cross-layer co-directional group is determined in a plurality of base station cells located on a preset load migration line, the base station cells at both ends of the cross-layer co-directional group are the K-layer main adjacent zones corresponding to the target base station cell, K is a preset value; the real-time load at the current time and the historical predicted load corresponding to the previous time are obtained, the historical predicted load is used to represent the load migration change that will occur at the current time predicted at the previous time; the cross-layer adjacent zone load data transmitted by the left adjacent zone co-directional group and the right adjacent zone co-directional group of the cross-layer co-directional group is received, the corresponding cross-layer adjacent zone load data each includes historical load data and real-time load data, the historical load data is used to represent the predicted load of the K-layer main adjacent zone of the corresponding adjacent zone co-directional group at the previous time, and the real-time load data is used to represent the actual load of the K-layer main adjacent zone of the corresponding adjacent zone co-directional group at the current time; the load prediction is performed based on the real-time load, the historical predicted load, and the cross-layer adjacent zone load data by using a preset load prediction model, to obtain the forward predicted load and the reverse predicted load corresponding to the target base station cell, and to determine the corresponding load prediction result based on the forward predicted load and the reverse predicted load, thereby solving the problem that the load prediction method in the related art cannot be applied to the high-speed mobile communication scenario without accurate time period rules, such as user high-speed movement and load burst, and achieving the beneficial effects that the base station cell can detect and sense the user load and the change trend of the adjacent zone coming from a farther distance at an accurate time point in advance, and the predicted load and the actual load of the base station cell are combined to provide effective reference and guidance for the energy saving, load control, and traffic guarantee of the base station cell.

[0188] In some embodiments, the prediction module 34 is further configured to utilize the load prediction model to perform forward load calculation on the real-time load, the historical predicted load, and historical load data and real-time load data corresponding to the left-neighbor same-direction group to obtain a forward predicted load; and utilize the load prediction model to perform reverse load calculation on the real-time load, the historical predicted load, and historical load data and real-time load data corresponding to the right-neighbor same-direction group to obtain a reverse predicted load.

[0189] In some embodiments, the apparatus is further configured to obtain a first actual load of each base station cell located on the preset load migration line at a current time; determine a nearest left-neighbor and a nearest right-neighbor corresponding to each base station cell, and obtain actual loads corresponding to the nearest right-neighbor and the nearest left-neighbor corresponding to the base station cell at a previous time, respectively, wherein a positive direction is set from a left end to a right end of the preset load migration line, the nearest left-neighbor is used to represent a first base station cell located on a left side of the corresponding base station cell in the positive direction, and the nearest right-neighbor is used to represent a first base station cell located on a right side of the corresponding base station cell in the positive direction; determine a first migration load of the nearest right-neighbor migrated to the base station cell at the previous time according to a first preset proportion based on the actual load of the nearest right-neighbor at the previous time, and determine a second migration load of the nearest left-neighbor migrated to the base station cell at the previous time according to a second preset proportion based on the actual load of the nearest left-neighbor at the previous time; determine a real-time load corresponding to the base station cell when performing forward prediction at the current time according to a difference between the first actual load and the first migration load, and determine a real-time load corresponding to the base station cell when performing reverse prediction at the current time according to a difference between the first actual load and the second migration load.

[0190] In some embodiments, the load prediction model comprises a forward predicted load calculation formula and a reverse predicted load calculation formula

[0191]

[0192]

[0193] wherein m represents a current time, m-1 represents a previous time, → represents a positive direction, ← represents a reverse direction, K represents a number of layers of a cross-layer same-direction group, p is a target base station cell C i an index of a cross-layer same-direction group in the corresponding P cross-layer same-direction groups, L i (m) of the target base station cell C i an actual load of the target base station cell C at the current time, i forward predicted loads corresponding to the target base station cell C the target base station cell C i the first layer main neighbor cell in the corresponding right neighbor co-site group actual load at the last time, γ is a first preset proportion, the j-i layer main neighbor cell in the left neighbor co-site group actual load at the current time, the j-i layer main neighbor cell in the left neighbor co-site group the corresponding nearest right neighbor cell actual load at the last time, the j-i layer main neighbor cell in the left neighbor co-site group corresponding forward predicted load at the last time, respectively represent the target base station cell C i backward predicted load at the current time and the last time, the target base station cell C i the first layer main neighbor cell in the corresponding left neighbor co-site group actual load at the last time, the j-i layer main neighbor cell in the right neighbor co-site group actual load at the current time, the j-i layer main neighbor cell in the right neighbor co-site group the corresponding nearest left neighbor cell actual load at the last time, the j-i layer main neighbor cell in the right neighbor co-site group corresponding backward predicted load at the last time; α, β, γ are weight factors between [0, 1], α is the weight of the load of the target base station cell C i itself, (1-α) is the weight of the corresponding load of the neighbor co-site group, β is the weight of the real-time load of each base station cell, (1-β) is the weight of the predicted load of each base station cell, γ is a first preset proportion, and 1-γ is a second preset proportion.

[0194] In some embodiments, the prediction module 34 is further configured to sum the forward predicted load and the backward predicted load to obtain a first predicted total load corresponding to the target base station on the preset load migration line; accumulate the first predicted total loads corresponding to a plurality of preset load migration lines to generate a total predicted load, wherein the load prediction result comprises the total predicted load.

[0195] In some embodiments, after determining the corresponding load prediction result, the apparatus is further configured to obtain a total predicted load corresponding to the current time and an actual load corresponding to the current time, respectively, and obtain a historical total predicted load corresponding to the previous time and a historical actual load corresponding to the previous time, respectively; generate a difference function of load prediction based on at least two of the actual load corresponding to the current time, the total predicted load corresponding to the current time, the historical total predicted load corresponding to the previous time, and the historical actual load corresponding to the previous time, and determine a calculation value corresponding to the difference function of load prediction, wherein the difference function of load prediction is used to represent a prediction performance of the load prediction; determine whether the calculation value of the difference function of load prediction is greater than a preset threshold, and adjust a weight factor corresponding to the load prediction model according to the determination result, wherein the weight factor corresponding to the load prediction model includes a weight α of the target base station cell itself, a weight β of the actual load of each base station cell, and a first preset proportion γ. i

[0196] In some embodiments, the difference function of load prediction includes one of the following: a difference between the actual load corresponding to the current time and the historical total predicted load corresponding to the previous time, a ratio of the difference between the actual load corresponding to the current time and the historical total predicted load corresponding to the previous time and the actual load corresponding to the current time, a difference between the historical total predicted load corresponding to the previous time and the total predicted load corresponding to the current time, and a difference between the actual load corresponding to the current time and the historical actual load corresponding to the previous time and a load difference generated by weighting the difference between the total predicted load corresponding to the current time and the historical total predicted load corresponding to the previous time, and the difference between the actual load corresponding to the current time and the historical actual load corresponding to the previous time.

[0197] In some embodiments, the obtaining module 31 is further configured to obtain handover frequency granularity data corresponding to each base station cell on the preset load migration line, and select a neighbor area with the maximum handover frequency granularity data as a main neighbor area of the base station cell; wherein the handover frequency granularity data is used to represent the number of user handovers between the base station cell and the neighbor area corresponding to the base station cell within a preset time; take the selected left main neighbor area and right main neighbor area corresponding to each base station cell as a base station cell co-directional group corresponding to the base station cell; take the target base station cell as a starting point, select K continuous base station cell co-directional groups corresponding to the base station cell along the forward direction and the reverse direction of the preset load migration line, respectively, and perform duplicate removal processing on the base station cells in the selected base station cell co-directional groups to obtain a cross-layer co-directional group.

[0198] Figure 4 is a structural schematic diagram of a base station of an embodiment of the present application, as Figure 4 ​As shown in the figure, this application embodiment provides a base station, including a processor 41, a communication interface 42, a memory 43, and a communication bus 44, wherein the processor 41, the communication interface 42, and the memory 43 communicate with each other through the communication bus 44.

[0199] Memory 43 is used to store computer programs;

[0200] When processor 41 executes the program stored in memory 43, it implements... Figure 1 The methods and steps in the text.

[0201] The processing implementation in this base station Figure 1 The method steps described above, and the resulting technical effects, are the same as those achieved in the embodiments described above. Figure 1 The technical effects of the dynamic load forecasting method are consistent with those in the text, and will not be elaborated further here.

[0202] The communication bus mentioned in the base station above can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 4 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0203] The communication interface is used for communication between the aforementioned terminal and other devices.

[0204] The memory may include random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0205] The processor described above can be a general processor, including a central processing unit (CPU), a network processor (NP), etc.; can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component.

[0206] The embodiment of the present application further provides a computer readable storage medium, which has a computer program stored thereon, and the computer program is executed by a processor to implement the steps of the dynamic load prediction method provided by any one of the preceding method embodiments.

[0207] In another embodiment provided by the present application, a computer program product containing instructions, which, when executed on a computer, causes the computer to perform the steps of the dynamic load prediction method described in any one of the preceding embodiments.

[0208] It should be noted that, in the present document, relational terms such as "first" and "second", and the like, are used solely to distinguish one entity or action from another entity or action, without necessarily requiring or implying any actual such relationship or order between or among the entities or actions. Moreover, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises... a" does not, without more constraints, exclude the existence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0209] The above description is merely one specific implementation of the application. Many modifications and other embodiments of the application set forth herein will come to mind to one skilled in the art to which the application pertains having the benefit of the teachings presented in the foregoing description. Therefore, it is to be understood that the application is not to be limited to the specific embodiments disclosed and that modifications and other embodiments are intended to be included within the scope of the appended claims. Without intent to limit the scope of the application, examples of modifications and other embodiments of the application in addition to those of the specific embodiment described herein will be readily apparent to one skilled in the art from the teachings set forth in the foregoing description. The goals, objectives and advantages of the application will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.

Claims

1. A dynamic load prediction method applied to a target base station cell under a high-speed private network, characterized in that, The method comprises the steps of: determining a cross-layer co-directional group in a plurality of base station cells located on a preset load migration line, wherein the base station cells located at both ends of the cross-layer co-directional group are K-layer main adjacent areas corresponding to the target base station cell, K being a preset value; obtaining real-time load at the current time and historical predicted load corresponding to the last time, the historical predicted load being used to represent the load migration change predicted at the last time to be generated at the current time; receiving cross-layer adjacent area load data transmitted by left adjacent area co-directional groups and right adjacent area co-directional groups of the cross-layer co-directional group, wherein the corresponding cross-layer adjacent area load data each comprises historical load data and real-time load data, the historical load data being used to represent the predicted load at the last time of the K-layer main adjacent area possessed by the corresponding adjacent area co-directional group, and the real-time load data being used to represent the actual load at the current time of the K-layer main adjacent area possessed by the corresponding adjacent area co-directional group; performing load prediction based on the real-time load, the historical predicted load and the cross-layer adjacent area load data by using a preset load prediction model, obtaining forward predicted load and reverse predicted load corresponding to the target base station cell, and determining a corresponding load prediction result based on the forward predicted load and the reverse predicted load.

2. The method of claim 1, wherein, The method comprises the steps of: performing load prediction based on the real-time load, the historical predicted load and the cross-layer adjacent area load data by using a preset load prediction model, obtaining forward predicted load and reverse predicted load corresponding to the target base station cell, and determining a corresponding load prediction result based on the forward predicted load and the reverse predicted load. The method comprises the steps of:

3. The method of claim 2, wherein, performing forward load calculation on the real-time load, the historical predicted load and the historical load data and the real-time load data corresponding to the left adjacent area co-directional group by using the load prediction model, to obtain the forward predicted load; performing reverse load calculation on the real-time load, the historical predicted load and the historical load data and the real-time load data corresponding to the right adjacent area co-directional group by using the load prediction model, to obtain the reverse predicted load. The method further comprises the steps of: obtaining a first actual load of each of the base station cells located on the preset load migration line at the current time; determining a nearest left adjacent area and a nearest right adjacent area corresponding to each of the base station cells, and obtaining actual load corresponding to the nearest right adjacent area and the nearest left adjacent area corresponding to each of the base station cells at the last time, wherein a positive direction is defined from the left end to the right end of the preset load migration line, the nearest left adjacent area is used to represent the first base station cell located on the left side of the corresponding base station cell in the positive direction, and the nearest right adjacent area is used to represent the first base station cell located on the right side of the corresponding base station cell in the positive direction; determining a first migration load of the nearest right adjacent area migrated to the base station cell at the last time by a first preset proportion based on the actual load of the nearest right adjacent area at the last time, and determining a second migration load of the nearest left adjacent area migrated to the base station cell at the last time by a second preset proportion based on the actual load of the nearest left adjacent area at the last time; The real-time load corresponding to the base station cell when the forward prediction is performed at the current time is determined according to a difference between the first actual load and the first migration load, and the real-time load corresponding to the base station cell when the reverse prediction is performed at the current time is determined according to a difference between the first actual load and the second migration load.

4. The method of claim 3, wherein, The load prediction model comprises a forward prediction load calculation formula and a backward prediction load calculation formula Where m represents the current time, m-1 represents the previous time, → represents the forward direction, ← represents the reverse direction, K represents the layer number of the cross-layer same-direction group, and p is the target base station cell C. i The label of one of the P cross-layer in-direction groups, L i (m) represents the target base station cell C. i The actual load at the current moment, Representing target base station cell C respectively i The positive forecast load corresponding to the current time and the previous time. For target base station cell C i The corresponding first-layer primary neighbor in the right neighbor group. In the previous moment, the actual load was γ, which was the first preset ratio. The ji-th primary neighbor in the left neighbor group is the same direction as the ji-th layer. The actual load at the current moment, The ji-th primary neighbor in the left neighbor group is the same direction as the ji-th layer. The corresponding nearest right neighbor The actual load at the previous moment, The ji-th primary neighbor in the left neighbor group is the same direction as the ji-th layer. The positive forecast load corresponding to the previous time step Representing target base station cell C respectively i The reverse predicted load at the current time and the previous time, For target base station cell C i The corresponding first-layer primary neighbor in the left neighbor group. The actual load at the previous moment, The ji-th main neighboring region in the same direction group of the right neighboring regions The actual load at the current moment, The ji-th main neighboring region in the same direction group of the right neighboring regions The corresponding nearest left neighbor The actual load at the previous moment, The ji-th main neighboring region in the same direction group of the right neighboring regions The reverse predicted load corresponding to the previous time step; α, β, and γ are weighting factors between [0, 1], and α is the target base station cell C. i The weight of its own load, (1-α) is the weight of the load corresponding to the neighboring cell in the same direction group, β is the weight of the real-time load of each base station cell, (1-β) is the weight of the predicted load of each base station cell, γ is the first preset ratio, and 1-γ is the second preset ratio.

5. The method of claim 4, wherein, The corresponding load prediction result is determined based on the forward prediction load and the reverse prediction load, and the corresponding load prediction result comprises: The first prediction total load corresponding to the target base station on the preset load migration line is obtained by summing the forward prediction load and the reverse prediction load; The first prediction total loads corresponding to a plurality of preset load migration lines are accumulated to generate a total prediction load, and the load prediction result comprises the total prediction load.

6. The method of claim 5, wherein, After the corresponding load prediction result is determined, the method further comprises: The total prediction load corresponding to the current time and the actual load corresponding to the current time are respectively acquired, and the historical total prediction load corresponding to the previous time and the historical actual load corresponding to the previous time are respectively acquired; A difference function of load prediction is generated based on at least two of the actual load corresponding to the current time, the total prediction load corresponding to the current time, the historical total prediction load corresponding to the previous time and the historical actual load corresponding to the previous time, and a calculation value corresponding to the difference function of load prediction is determined, wherein the difference function of load prediction is used to represent the prediction performance of the load prediction. determining whether the calculated value corresponding to the difference variation function is greater than a preset threshold, and adjusting a weight factor corresponding to the load prediction model according to a determination result, wherein the weight factor corresponding to the load prediction model comprises a weight α of a self-load, a weight β of an actual load of each base station cell, and the first preset proportion γ. i a weight α of a self-load, a weight β of an actual load of each base station cell, and the first preset proportion γ.

7. The method of claim 6, wherein, The difference function of load prediction comprises one of a difference between the actual load corresponding to the current time and the historical total prediction load corresponding to the previous time, a ratio of the difference between the actual load corresponding to the current time and the historical total prediction load corresponding to the previous time and the actual load corresponding to the current time, a difference between the historical total prediction load corresponding to the previous time and the total prediction load corresponding to the current time, a difference between the actual load corresponding to the current time and the historical actual load corresponding to the previous time, and a load difference value generated by weighting the difference between the total prediction load corresponding to the current time and the historical total prediction load corresponding to the previous time, and the difference between the actual load corresponding to the current time and the historical actual load corresponding to the previous time.

8. The method of claim 6, wherein, The method further comprises: The center of gravity CoL(m, p) of the overall load of the preset high-speed private network section is calculated according to the following formula: Wherein, the high-speed private network segment is composed of the target base station cell C i and its corresponding one 2K-layer main adjacent zone of the cross-K-layer co-directional group, CoL(m, p) represents the target base station cell C i the center of gravity of the overall load of the corresponding high-speed private network segment, represents the actual load of each base station cell C j at m moment, p is the target base station cell C i the label of one cross-layer co-directional group in the corresponding P cross-layer co-directional groups; The difference function of load prediction comprises the center of gravity of the overall load.

9. The method of claim 3, wherein, The cross-layer adjacent cell load data transmitted by the left adjacent zone same direction group and the right adjacent zone same direction group of the cross-layer same direction group is received, and the cross-layer adjacent cell load data comprises: The nearest left neighbor zone synthesizes first cross-layer neighbor zone load data according to the first target load data, actual load corresponding to the current moment and historical predicted load corresponding to the last moment, and delivers the first cross-layer neighbor zone load data to the corresponding target base station cell, wherein the first neighbor zone group load data comprises cross-layer neighbor zone load data corresponding to the same direction group of the left neighbor zone of the nearest left neighbor zone, and the first target load data comprises cross-layer neighbor zone load data corresponding to all the main neighbor zones except the Kth layer main neighbor zone in the same direction group of the left neighbor zone of the nearest left neighbor zone; The nearest right neighbor zone synthesizes second cross-layer neighbor zone load data according to the second target load data, actual load corresponding to the current moment and historical predicted load corresponding to the last moment, and delivers the second cross-layer neighbor zone load data to the corresponding target base station cell, wherein the second neighbor zone group load data comprises cross-layer neighbor zone load data corresponding to the same direction group of the right neighbor zone of the nearest right neighbor zone, and the second target load data comprises cross-layer neighbor zone load data corresponding to all the main neighbor zones except the Kth layer main neighbor zone in the same direction group of the right neighbor zone of the nearest right neighbor zone.

10. The method of claim 1, wherein, In a plurality of base station cells located on a preset load migration line, a cross-layer same direction group is determined, comprising: Obtaining handover frequency granularity data corresponding to each of the base station cells on the preset load migration line, and selecting a neighbor zone with the largest handover frequency granularity data as a main neighbor zone of the base station cell; wherein the handover frequency granularity data is used to represent the number of user handovers between the base station cell and the neighbor zone corresponding to the base station cell within a preset time; Selecting the left main neighbor zone and the right main neighbor zone corresponding to each of the base station cells as the base station cell same direction group corresponding to the base station cell; Selecting K continuous base station cell same direction groups corresponding to the base station cells along the forward direction and the reverse direction of the preset load migration line respectively from the target base station cell as a starting point, and performing deduplication processing on the base station cells in the selected base station cell same direction groups to obtain the cross-layer same direction group.

11. A dynamic load prediction device applied to a target base station cell under a high-speed private network, characterized in that, Comprising: A determination module for determining a cross-layer same direction group in a plurality of base station cells located on a preset load migration line, wherein the base station cells at both ends of the cross-layer same direction group are the Kth layer main neighbor zones corresponding to the target base station cell, and K is a preset value; An acquisition module for acquiring real-time load at the current moment and historical predicted load corresponding to the last moment, wherein the historical predicted load is used to represent the load migration change that will occur at the current moment predicted at the last moment; The receiving module is configured to receive cross-layer adjacent cluster load data transmitted by left and right adjacent cluster co-directional groups of the cross-layer co-directional group, wherein the cross-layer adjacent cluster load data each includes historical load data and real-time load data, the historical load data is used to represent a predicted load of a K-layer main cluster of a corresponding adjacent cluster co-directional group at a previous time, and the real-time load data is used to represent an actual load of the K-layer main cluster of the corresponding adjacent cluster co-directional group at a current time. The prediction module is configured to perform load prediction based on the real-time load, the historical predicted load, and the cross-layer adjacent cluster load data by using a preset load prediction model, to obtain forward predicted load and reverse predicted load corresponding to the target base station cell, and to determine a corresponding load prediction result based on the forward predicted load and the reverse predicted load.

12. A base station, characterized by The apparatus includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus. The memory is configured to store a computer program. The processor is configured to execute the program stored in the memory to implement the steps of the dynamic load prediction method according to any one of claims 1-10.

13. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the dynamic load prediction method according to any one of claims 1-10.

Images