JP7638330B2 - Method for implementing network side device in wireless communication network and network side device - Google Patents

Description

The present disclosure relates to the field of wireless communications, and more specifically to a method performed by a network side device in a wireless communications network, and a corresponding network side device.

Wireless communication network systems are widely deployed to provide various types of communication content, such as voice, video, packet data, message transmission, broadcast, etc. A typical wireless communication network system may employ multiple access technologies that can communicate with multiple communication devices by employing sharable system resources (e.g., bandwidth, transmission power, etc.). Examples of such multiple access technologies include Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Frequency Division Multiple Access (FDMA), Orthogonal Frequency Division Multiple Access (OFDMA), Single Carrier Frequency Division Multiple Access (SC-FDMA), Time Division Synchronous Code Division Multiple Access (TD-SCDMA), and Long Term Evolution (LTE). LTE/LTE-Advanced is a set of mobile specification extensions to the Universal Mobile Telecommunications System (UMTS) delivered by the Third Generation Partnership Project (3GPP).

A wireless communication network system may include multiple network side devices, for example multiple base stations. Each base station can simultaneously support communication with multiple communication devices, which are called user equipment (UE). The UE can communicate with the base station via downlink and uplink. The downlink refers to the communication link from the base station to the UE, and the uplink refers to the communication link from the UE to the base station.

Because base stations need to communicate with multiple UEs, they are forced to consume large amounts of power. In today's carbon neutral environment, reducing power consumption at base stations is an urgent problem to be solved.

TW 201528203A, "Behavioral Screen Automated Signal Line Traffic Growth Prediction System"

In response to the above problem, according to one aspect of the present disclosure, a network side device in a wireless communication network is provided, the network side device in the wireless communication network includes an acquisition unit configured to acquire information about traffic in a past time period, time indicator information indicating whether the traffic in the past time period is traffic in a first time period or traffic in a second time period, and sector direction indication information indicating the sector direction of a cell corresponding to the network side device, and a processing unit configured to predict traffic information in a future time period using a neural network model based on the information about traffic in the past time period, the time indicator information, and the sector direction indication information.

According to one aspect of the present disclosure, a wireless communication method for a network side device is provided, which includes acquiring information about traffic in a past time period, time indicator information indicating whether the traffic in the past time period is traffic in a first time period or traffic in a second time period, and sector direction indication information indicating the sector direction of a cell corresponding to the network side device, and predicting traffic information in a future time period using a neural network model based on the information about traffic in the past time period, the time indicator information, and the sector direction indication information.

The network side device and method according to the above aspect of the present disclosure can predict with high accuracy the traffic that a base station should process in a future time period, so that the base station can turn off other resources and provide only a portion of the resources that can process the above traffic, thereby saving power consumption at the base station.

The above and other objects, features and advantages of the present disclosure will become more apparent by describing the embodiments of the present disclosure in more detail with reference to the drawings. The drawings are used to provide a further understanding of the embodiments of the present disclosure, constitute a part of the specification, and are used to explain the present disclosure together with the embodiments of the present disclosure, and are not intended to limit the present disclosure. In the drawings, the same reference numerals generally represent the same components or steps.
FIG. 1 shows a flowchart of a method for wireless communication in a network side device according to an embodiment of the present disclosure. FIG. 2A shows a schematic diagram of traffic forecasting according to one embodiment of the present disclosure. FIG. 2B shows a schematic diagram of traffic forecasting according to another embodiment of the present disclosure. FIG. 3 is a schematic diagram showing the variation of the values of the coefficients α, β, and γ throughout the day. 4A to 4E are diagrams illustrating a method for determining a first coefficient based on traffic for the past four days according to an embodiment of the present disclosure. FIG. 5 shows a flowchart of a method for wireless communication in a network side device according to an embodiment of the present disclosure. FIG. 6A illustrates some training data used in training a neural network model according to an embodiment of the present disclosure. FIG. 6B shows a schematic diagram of training a neural network model according to an embodiment of the present disclosure. 7A to 7C show schematic diagrams of power saving level information added when training a neural network according to an embodiment of the present disclosure. FIG. 8A shows a schematic diagram of a loss function before refinement according to an embodiment of the present disclosure. FIG. 8B shows a schematic diagram of the improved loss function according to an embodiment of the present disclosure. 9A to 9C show schematic diagrams for obtaining fusion weights according to an embodiment of the present disclosure. 10A-10B show a schematic diagram of a data normalization process when training a neural network according to an embodiment of the present disclosure. FIG. 11 shows a flowchart of another method of wireless communication in a network side device according to an embodiment of the present disclosure. 12A and 12B show schematic diagrams of resources used by a network side device according to one embodiment of the present disclosure. 13A to 13C show other schematic diagrams of resources used by a network side device according to an embodiment of the present disclosure. FIG. 14 shows a schematic diagram of data throughput degradation rate and power consumption according to an embodiment of the present disclosure. FIG. 15 is a block diagram of a network side device 1500 in a wireless communication network according to an embodiment of the present disclosure. FIG. 16 is a block diagram of a network side device 1600 in a wireless communication network according to an embodiment of the present disclosure. FIG. 17 is a block diagram of a network side device 1700 in a wireless communication network according to an embodiment of the present disclosure. FIG. 18 is a diagram illustrating an example of a hardware configuration of a network side device according to an embodiment of the present disclosure.

In order to make the objectives, technical solutions and advantages of the present disclosure clearer, exemplary embodiments of the present disclosure are described in detail below with reference to the drawings. In the drawings, the same reference numerals always refer to the same elements. It should be noted that the described embodiments of the present disclosure are merely illustrative and should not be construed as limiting the scope of the present disclosure.

Because the base station needs to communicate with a large number of UEs, it is inevitable that the base station will consume a large amount of power. In order to save the power consumption of the base station, some power saving methods for base stations are provided. For example, a method for node control based on different time periods is provided. For example, during a certain low traffic period from 0:00 to 6:00 a.m., some or all cells in the base station are turned off to achieve the purpose of power saving. However, this method cannot save power during the entire period, and the power saving effect is low, and when the traffic changes (for example, the number of UEs accessing from 0:00 to 6:00 a.m. increases sharply), the base station is still in power saving mode, and the data throughput of the base station is reduced, which affects the network experience of the user.

Also, a method for node control based on real-time traffic is provided. For example, based on observed real-time traffic, if the observed real-time traffic is lower than a preset traffic threshold, some or all cells in the base station are turned off to achieve the purpose of power saving. However, for a scenario in which a Radio Intelligent Controller (RIC) is deployed, there is a delay in the controller (e.g., RIC) acquiring real-time traffic (e.g., 15 to 30 minutes), so the base station cannot accurately select the power saving mode.

Also, the power saving objective can be achieved by predicting traffic for a future time period using historical traffic data and turning off some or all cells in the base station based on the predicted traffic. However, since such an embodiment is performed by considering only historical traffic, the rough prediction error is large and the base station cannot accurately select the power saving mode, resulting in low power saving effect or reduced data throughput.

According to one aspect of the present disclosure, a method for wireless communication in a network side device of a base station is provided, and the method can predict with high accuracy the traffic that the network side device needs to process in a future time period, so that the network side device performs more accurate power saving operations based on the prediction results. The method for wireless communication in a network side device and the corresponding network side device provided by one aspect of the present disclosure will be described in detail below with reference to the drawings.

Figure 1 shows a flowchart of a method for wireless communication in a network side device according to an embodiment of the present disclosure. The method shown in Figure 1 may be performed by the network side device. As an example, the network side device may be a base station. Or, the network side device may be an operation administration and maintenance (OAM) module or a service management and orchestration (SMO) module that interacts with the outside of the base station and the base station. When the network side device is the OAM or SMO module, the OAM or SMO module needs to transmit traffic information for a future time period predicted by executing the method provided by the present disclosure to, for example, the base station, and the base station performs a corresponding operation.

As shown in FIG. 1, in step S110, the network side device can obtain information about traffic in a historical time period and information about traffic in an adjacent time period, where the adjacent time period is closer to a future time period than the historical time period.

For example, the information about the traffic of the historical time period and the information about the traffic of the adjacent time period can be stored in an internal memory of the network side device, such as a memory, such as a random access memory (RAM), a read-only memory (ROM), and an electrically erasable programmable ROM (EEPROM). Alternatively, the information about the traffic of the historical time period and the information about the traffic of the adjacent time period can be stored in an external memory or cloud that interacts with the network side device.

As an example, a time period may be a time in hours or minutes. For example, a time period may be 1 hour or 2 hours. Also, a time period may be 15 minutes or 30 minutes.

In an embodiment of the present disclosure, the future time period may be a time period in the predicted future. For example, the future time period may be the time period immediately following the current time, e.g., if the current time is 12:00, the future time period may be 13:00-14:00. Also, for example, the future time period may be a time period separated from the current time by a predetermined time, e.g., if the current time is 12:00, the future time period may be 15:00-16:00, which is a time period that is two hours or less apart.

The historical time band may be a time band of days prior to the current day on which the future time band exists. For example, the historical time band may be a time band of days or days prior to the day on which the future time slot exists. Accordingly, the traffic of the historical time band may be traffic generated within the time band of days prior to the day on which the future time slot exists. For example, if the time band is one hour, the traffic of the historical time band may be traffic generated within each hour during the day prior to the day on which the future time slot exists.

The adjacent time period may also be a time period prior to a future time period within the current day. For example, the adjacent time period may be one or more time periods prior to a future time period within the current day. For example, if the future time period is from 4:00 pm to 5:00 pm on the current day, the adjacent time period may be from 3:00 pm to 4:00 pm on the current day, or from 10:00 am to 11:00 am on the current day, etc. Accordingly, the traffic of the adjacent time period may be traffic generated within a time period prior to the future time period within the current day. For example, if the time period is one hour, the traffic of the adjacent time period may be traffic generated within a time period prior to the future time period within the current day.

In step S120, the network side device can predict traffic for a future time period based on information about traffic for the historical time period and information about traffic for the adjacent time period, or on information about traffic for multiple stored adjacent time periods.

For example, the traffic information for the historical time period and the traffic information for the adjacent time period can be acquired in real time by the network side device. Alternatively, the traffic information for the historical time period and the traffic information for the adjacent time period can be acquired by the network side device at a preset time according to actual needs.

According to one embodiment of the present disclosure, in step S120, traffic for the future time period can be predicted based on information about traffic for a time period in the history time period that corresponds to the future time period, information about traffic for a time period in the history time period that corresponds to a specific adjacent time period, and information about traffic for the specific adjacent time period.

2A and 2B are schematic diagrams showing predicted traffic according to an embodiment of the present disclosure. In the example shown in FIG. 2A and FIG. 2B, an example will be described in which one hour is one time period.

FIG. 2A shows a schematic diagram of traffic prediction according to one embodiment of the present disclosure. In the example shown in FIG. 2A, the future time period in which traffic is predicted is preferably the time period from 12:00 to 13:00 on the current day, and the traffic in that time period is denoted as x. As shown in FIG. 2A, the time period corresponding to the future time period in the history time period is the time period from 12:00 to 13:00 on the previous day, and the traffic in that time period is denoted as a. The specific adjacent time period may be any time period in which traffic has already been stored on the current day. For example, the time period immediately before the future time period, i.e., the time period from 11:00 to 12:00, and the traffic in that time period is denoted as e. Also, the time period corresponding to the specific adjacent time period in the history time period is the time period from 11:00 to 12:00 on the previous day, and the traffic in that time period is denoted as d. In step S120, the network side device can predict x based on the above a, d, and e.

In the example shown in FIG. 2A, it is preferable that the future time slot in which traffic is predicted is adjacent to a specific adjacent time slot for predicting it. Or, there may be one or more interval time slots (also called observation delays in this case) between the future time slot and the current time. FIG. 2B shows a schematic diagram of traffic prediction according to another embodiment of the present disclosure. As shown in FIG. 2B, the future time slot in which traffic is predicted is preferably a time slot from 12:00 to 13:00 on the current day, and the traffic of the time slot is denoted as x. The time slot corresponding to the future time slot in the history time slot is a time slot from 12:00 to 13:00 on the previous day, and the traffic of the time slot is denoted as a. If the current time is 11:00, as shown in FIG. 2, the current time is 11:00. The specific adjacent time slot may be a time slot from 8:00 to 9:00 on the current day, and the traffic of the time slot is denoted as f. The historical time period that corresponds to a particular adjacent time period is the time period from 8:00 to 9:00 on the previous day, and the traffic in that time period is denoted as c. In step S120, the network side device can predict x based on the above a, c, and f.

According to another embodiment of the present disclosure, the historical time band and the adjacent time band include a related time band of the future time band and a reference time band corresponding to the future time band and the related time band, respectively, or the adjacent time band includes a related time band of the future time band and a reference time band corresponding to the future time band and the related time band, respectively. In this case, in step S120, the traffic of the future time band can be predicted based on a first coefficient indicating a correlation between the traffic of the future time band and the traffic of the related time band and the traffic of the reference time band.

For example, referring to FIG. 2A, the historical time period may be the 10:00-11:00, 11:00-12:00, or 12:00-13:00 time period of the previous day. The adjacent time period may be the 10:00-11:00, or 11:00-12:00 time period of the current day.

As an example, the future time period may be the 12:00-13:00 time period of the current day, and the related time period of the future time period included in the historical time period and the adjacent time period may be the 12:00-13:00 time period of the previous day. The reference time periods corresponding to the future time period and the related time period included in the historical time period and the adjacent time period may be the 11:00-12:00 time period of the current day and the 11:00-12:00 time period of the previous day, respectively. The first coefficient α may indicate the correlation between the traffic of the 12:00-13:00 time period of the current day and the 12:00-13:00 time period of the previous day, and the traffic of the 11:00-12:00 time period of the current day and the 11:00-12:00 time period of the previous day.

As another example, the future time slot may be the 12:00-13:00 time slot of the current day, and the related time slot of the future time slot included in the historical time slot and the adjacent time slot may be the 11:00-12:00 time slot of the current day. The reference time slots corresponding to the future time slot and the related time slot included in the historical time slot and the adjacent time slot may be the 12:00-13:00 time slot of the previous day and the 11:00-12:00 time slot of the previous day, respectively. The first coefficient may indicate a correlation between the traffic of the 12:00-13:00 time slot of the current day and the traffic of the 11:00-12:00 time slot of the current day and the traffic of the 12:00-13:00 time slot of the current day and the traffic of the 11:00-12:00 time slot of the previous day. For example, in this case, the first coefficient β can indicate the correlation between the difference between the traffic during the 12:00-13:00 time slot of the current day and the traffic during the 11:00-12:00 time slot of the current day, and the difference between the traffic during the 12:00-13:00 time slot of the previous day and the traffic during the 11:00-12:00 time slot of the previous day. Alternatively, the first coefficient γ can indicate the correlation between the ratio between the traffic during the 12:00-13:00 time slot of the current day and the traffic during the 11:00-12:00 time slot of the current day, and the ratio between the traffic during the 12:00-13:00 time slot of the previous day and the traffic during the 11:00-12:00 time slot of the previous day.

As another example, the future time period may be the 12:00-13:00 time period of the current day, and the related time period of the future time period included in the adjacent time period may be the 11:00-12:00 time period of the current day. The reference time periods corresponding to the future time period and the related time period included in the adjacent time period may be the 11:00-12:00 time period of the current day and the 10:00-11:00 time period of the current day, respectively. The first coefficient δ may indicate the correlation between the traffic in the 12:00-13:00 time period of the current day and the 11:00-12:00 time period of the current day, and between the traffic in the 11:00-12:00 time period of the current day and the 10:00-11:00 time period of the current day.

Here, the formulas corresponding to α, β, γ, and δ are as shown in Table 1 below.

Note that in the example of the first coefficient described above with reference to FIG. 2A and Table 1, the day before the current day is used as an example, but the above method can also be applied to an example where a certain day in the past that is one or more days away from the current day is used as an example.

The first coefficient may be obtained by statistically analyzing historical traffic data. This will be further described below with reference to FIGS. 4A to 4E. Alternatively, the first coefficient may be preset as a fixed value according to actual needs. A larger absolute value of the first coefficient may indicate a higher relevance, or a smaller absolute value of the first coefficient may indicate a lower relevance. For example, a value of the first coefficient of +1 or -1 indicates a high correlation, and a value of the first coefficient of 0 indicates a low correlation. The value of the first coefficient may be different for different time periods throughout the day. In addition, the values of the first coefficients (e.g., the above formulas α, β, γ, δ) obtained based on different methods may be different for a certain time period. Therefore, according to another embodiment of the present disclosure, when actually predicting traffic for a future time period, any of the candidate coefficients obtained based on different methods according to the actual situation may be selected as the first coefficient. For example, the network side device may select a candidate coefficient having the highest correlation with the future time period from among a plurality of candidate coefficients as the first coefficient.

Figure 3 shows the change in the values of coefficients α, β, and γ throughout the day, taking coefficients α, β, and γ as examples. As shown in Figure 3, between 5:00 and 6:00 a.m., the absolute value of coefficient α is clearly smaller than the absolute value of coefficients β or γ, and in other time periods, coefficients β and γ are clearly larger than the absolute value of coefficient α. In such a case, in step S120, the coefficient with the maximum correlation value may be selected as the first coefficient. For example, α, β, γ, and δ may be candidate coefficients for the first coefficient. In step S120, the maximum coefficient from the candidate coefficients α, β, γ, and δ may be selected as the first coefficient for the future time period for which the traffic is to be predicted.

For example, in the example shown in FIG. 3, between 5:00 a.m. and 6:00 a.m., traffic for a future time period can be predicted by determining the candidate coefficient corresponding to method 2 or 3 in Table 1 as the first coefficient, and for other time periods, traffic for a future time period can be predicted by determining the candidate coefficient corresponding to method 1 in Table 1 as the first coefficient.

Thus, the method provided by the present disclosure can predict traffic for future time periods based on different candidate coefficients, making the method provided by the present disclosure flexible and versatile.

In addition, the above method provided by the present disclosure allows for more accurate prediction of traffic changes in future time periods (e.g., increases or decreases in traffic in future time periods) and traffic in future time periods by selecting the most relevant candidate coefficients.

As can be seen from the detailed description of the method provided by the present disclosure above, the method provided by the present disclosure can accurately predict traffic changes in future time periods by using the first coefficient, but there are occasional situations in which the prediction of the magnitude of the change is inaccurate. For such situations, the present disclosure further provides a further improvement to the above method.

According to yet another embodiment of the present disclosure, in step S120, the method may further include predicting traffic for the future time period based on a second coefficient indicating a correlation between the magnitude of traffic for the relevant time period and the magnitude of traffic for a reference time period for the relevant time period.

For example, the relationship between the traffic magnitude of the relevant time band and the traffic magnitude of a reference time band for the relevant time band may be the ratio of the traffic magnitude of the relevant time band to the traffic magnitude of a reference time band for the relevant time band, or may be the ratio of the average traffic values or the ratio of the maximum traffic values therebetween.

For example, in the example shown in Table 2 below, the average value of traffic in a future time slot (e.g., 13:00-14:00) on the current day is denoted as x. The relevant time slot is the time slot of the previous day that corresponds to the future time slot, and its average value of traffic is denoted as a, with a magnitude of 10. The reference time slot of the relevant time slot is the time slot of the previous day that corresponds to the adjacent time slot, and its average value of traffic is denoted as d, with a magnitude of 3. Let the second coefficient be L. In such a case, the second coefficient L may be 10/3, i.e., the second coefficient L is equal to 3, in which case the formula of method 1 in Table 1 above can be transformed to x = a + L * α * (e - d). If α is 1, the traffic x predicted for the future time slot in this case is 30. The predicted x = a + α * (e - d) before transforming method 1, i.e., x is 16. Then, the actual traffic in the future time slot is 30.

The relevant time bin may further be a time bin corresponding to a future time bin in the past two days, the traffic average value of which is 20, and the reference time bin of the relevant time bin may be a time bin corresponding to an adjacent time bin in the past two days, the traffic average value of which is 6. In such a case, the second coefficient L may be 20/6, i.e., the second coefficient L is equal to 3. In this case, the disclosure of method 1 in Table 1 above can be transformed to x=a+L*α*(e-d). If α is 1, the traffic x of the future time slot predicted in this case is 30. Before transforming method 1, the unpredicted x=a+α*(e-d), i.e., x is 16. Then, the actual traffic in the future time bin is 30.

Thus, by introducing the second coefficient, the method provided by the present disclosure can not only accurately predict traffic changes in future time periods, but also more accurately predict the magnitude of traffic changes in future time periods, thereby making the predicted traffic more accurate.

After accurately predicting the traffic of the future time band, the network side device saves power consumption in the network side device by only providing some resources that can handle the predicted traffic and turning off other resources. For example, if the network side device can handle the traffic of the future time band with only three antennas, the network side device can turn off the remaining antennas and not operate the remaining antennas, thereby saving power consumption in the network side device. Furthermore, for example, if the network side device can handle the traffic of the future time band with only two component carriers, the network side device does not transmit the remaining component carriers (CCs), thereby saving power consumption in the network side device.

A method that can be executed by a network side device in a wireless communication network provided by the present disclosure has been described above with reference to Figures 1 to 3. The method provided by the present disclosure can more accurately predict changes in traffic in future time periods based on the method provided by the present disclosure by using a first coefficient. The method provided by the present disclosure can also more accurately predict changes in the amount of traffic in future time periods by using the second coefficient. As a result, the method provided by the present disclosure can predict with high accuracy the traffic that the network side device needs to process in future time periods, and the network side device saves power consumption in the network side device by only providing some resources that can process the traffic and turning off other resources.

As described above in the method provided by the present disclosure, the first coefficient can be obtained by statistically analyzing historical traffic data, which is described below in the form of an example.

Figures 4A to 4E are diagrams illustrating how a first coefficient is calculated based on traffic over the past four days according to an embodiment of the present disclosure.

Referring to FIG. 4A, select the traffic for the past four days enclosed in a gray dotted frame in FIG. 4A. The traffic for the adjacent period from 11:00 to 12:00 for the past four days in FIG. 4A and the traffic for the historical period corresponding to the future period from 12:00 to 13:00 are as shown in FIG. 4B. In this case, the following description will be given using an example in which the candidate coefficient α is the first coefficient. Note that the methods for calculating the other three candidate coefficients are similar, so repeated description will be omitted.

First, use a 2*2 sliding window of "2 days x 2 periods" to extract three 2*2 data points as shown in Figure 4C.

Next, for data within the same time period, calculate the difference between adjacent days, i.e. 11-8=3, 9-6=3, 8-7=1, 6-5=1, 7-8=-1, 5-5=0, and obtain the data shown in Figure 4D.

Next, the data shown in Figure 4D is averaged over the same period, i.e. (3+1+(-1))/3=1, (3+1+0)/3=4/3, and the data shown in Figure 4E is obtained.

Finally, the value of the coefficient α can be obtained by substituting the above data into the following equation (1):

Here, (x _i -x ^- ) indicates that the average value corresponding to each data in the same time period (i.e., 11:00-12:00) in Fig. 4D is subtracted, and (y _i -y ^- ) indicates that the average value corresponding to each data in the same time period (i.e., 12:00-13:00) in Fig. 4D is subtracted, where i is a positive integer.

After substituting the above data into equation (1) above, we obtain the following results:

The traffic predicted for the future time period can be obtained by substituting the calculated α value into the formula of Method 1 in Table 1. For example, in Fig. 4A, if the value of d is 5, the value of a is 4, and the value of e is 6, the traffic as shown below is calculated and obtained by Method 1.

In addition to the above-mentioned method for predicting traffic for future time periods, the present disclosure also provides a method for predicting traffic for future time periods, which is described below with reference to Figures 5-10B.

Figure 5 shows a flowchart of a method for wireless communication in a network side device according to an embodiment of the present disclosure. The method shown in Figure 5 can be executed by the network side device. As an example, the network side device may be a base station. Or, the network side device may be an operation administration and maintenance (OAM) module or a service management and orchestration (SMO) module that interacts with the outside of the base station and the base station. When the network side device is the OAM or SMO module, the OAM or SMO module needs to transmit traffic information for a future time period predicted by executing the method provided by the present disclosure to, for example, the base station, and the base station performs a corresponding operation.

As shown in FIG. 5, in step S510, the network side device can obtain information on traffic in a past time period, time indicator information, and sector direction indication information, where the time indicator information can indicate whether the traffic in the past time period is a first time period or a second time period, and the sector direction indication information can indicate the sector direction of the cell corresponding to the network side device.

As an example, the information about the traffic of the past time period may be the information about the traffic of the historical time period described in FIG. 1 and the information about the traffic of the adjacent time period. The information about the traffic of the past time period may be stored in an internal memory of the network side device, for example a memory, such as a random access memory (RAM), a read-only memory (ROM) and an electrically erasable programmable ROM (EEPROM). Alternatively, the information about the traffic of the past time period may be stored in an external memory or cloud that interacts with the network side device.

For example, the first period may be working days, e.g., every day or every day from Monday to Friday. The second period may be non-working days, e.g., every day or every day from Saturday to Sunday.

For example, the sector direction refers to the sector direction corresponding to each angle when combining uniform sectors as described below in FIG. 12A, or the sector direction corresponding to each angle when combining non-uniform sectors as described below in FIG. 13A. The traffic transmitted by the sector directions corresponding to different angles may be the same or different. Therefore, predicting future time zone traffic information based on the sector direction helps to make the predicted future time zone traffic information more accurate.

In step S520, the network side device can predict traffic information for a future time period using a neural network model based on the information about traffic for the past time period, the time sign information, and the sector direction indication information.

According to one embodiment of the present disclosure, the Neural Network (NN) model is a pre-trained neural network model. The Neural Network model may be any suitable conventional neural network model, such as a Convoluting Neural Network (CNN) model, a Recurrent Neural Network (RNN) model, etc.

The neural network model may be trained based on information about traffic over past time periods, time sign information, and sector direction indication information.

As an example, the information on traffic in the past time period may be traffic generated during each time of each day for the past 100 days. The time indicator information indicates whether the traffic generated per day for the past 100 days is weekday traffic or holiday traffic. The sector direction indication information can indicate the sector direction of the cell corresponding to the network side device when generating the above traffic, and will be described below in combination with the sector direction corresponding to 60 degrees shown in FIG. 12A.

As shown in FIG. 6A, FIG. 6A shows some training data when training a neural network model according to an embodiment of the present disclosure. The training data is to predict traffic (T ^- ₅ ) at the fifth hour on the fourth day (day 4). In this case, traffic four hours before the fourth day (i.e., _T1 to _T4 ) and traffic for each hour within three days before the fourth day (i.e., T1 to _T24 on day ₁ , _T1 to _T24 on day 2, and _T1 to _T24 on day 3) may be selected as traffic for the past period. Traffic for each hour within any one day or multiple days before the fourth day may also be selected.

Referring next to FIG. 6A, in this case, the time indicator information may be 0011, which is a total of four bits of information. The first bit "0" in "0011" may represent that all traffic generated on day 1 is holiday traffic. The second bit "0" in "0011" may represent that all traffic generated on day 2 is holiday traffic. The third bit "1" in "0011" may represent that all traffic generated on day 3 is working day traffic. The fourth bit "1" in "0011" may represent that all traffic generated on day 4 is working day traffic.

The sector direction indication information may be 6 bits of information, 010000. From left to right, this 6-bit information may indicate directions corresponding to 30 degrees, 90 degrees, 150 degrees, 210 degrees, 270 degrees, and 330 degrees, respectively, as described in FIG. 12A. When traffic is transmitted in the corresponding direction, the corresponding bit in the 6-bit information may be set to 1. When traffic is not transmitted in the corresponding direction, the corresponding bit in the 6-bit information may be set to 0. In this case, the "010000" may indicate that traffic is transmitted only in the sector direction corresponding to 90 degrees.

6A, after inputting the one training data into the neural network, the neural network obtains one predicted traffic T ^- ₅ , and according to the predicted traffic T ^- ₅ and the actual traffic _T5 at the time, the weight of the neural network is adjusted according to the loss function, so that the training of the neural network can be completed.

Or, after inputting the one training data into the neural network, the neural network obtains the predicted traffic T ^- ₅ , the predicted time indicator information and the predicted sector direction information, in this case, the neural network weights can be adjusted according to the predicted traffic T ^- ₅ and the actual traffic _T5 at the time, the predicted time indicator information and the actual time indicator information at the time, and the predicted sector direction information and the actual sector direction information at the time according to the loss function, thereby completing the training of the neural network.

In addition, as shown in FIG. 6B, in order to improve the prediction effect of the trained neural network, in the actual training, a large amount of data similar to the above one training data within many days (shown in Table 3 below) is first selected and input into the neural network for training.

In Table 3, the training data is used to train the neural network. The neural network is trained multiple times, and the validation data is used to validate the trained neural network each time. During the training process, only the neural network with the highest probability among the validation data is reserved. For example, if the current neural network has been trained 50 times, the one with the highest probability in the validation data is the neural network trained 25th time. After the 51st training, if the probability of the neural network trained 51st time in the validation data is lower than the probability of the neural network trained 25th time, the neural network trained 51st time is not reserved and training is continued. If the probability of the neural network trained 51st time in the validation data is higher than the probability of the neural network trained 25th time, the neural network trained 51st time is reserved, that is, the neural network trained 25th time is deleted and only the neural network trained 51st time is left, and training is continued until a neural network that meets the predetermined requirements is trained. The test data is finally used to test the trained neural network.

Based on the traffic T ^- ₅ predicted based on each training data and the actual traffic _T5 at the time, the weights of the neural network are adjusted based on a loss function to complete the training of the neural network. The loss function may be any suitable loss function, such as a mean square error (MSE) loss function, a cross entropy loss function (e.g. softmax), etc.

Finally, after the above training is completed, each weight of the neural network is fixed. That is, after the above process, the neural network is trained and can predict traffic information for future time periods.

As an example, the predicted future time slot traffic information may be one or more of the following: traffic for the future time slot, time marker for the future time slot, and sector direction for the future time slot.

The above describes in detail the wireless communication method in a network side device provided by the present disclosure. In addition, since the method provided by the present disclosure uses a neural network trained through a large amount of data, the method provided by the present disclosure can accurately predict traffic information for future time periods. Compared with the method described with reference to FIG. 1 above, the method described with reference to FIG. 5 is more convenient and quicker.

Furthermore, the traffic information for the future time period predicted by the method described with reference to FIG. 5 above can be further evaluated to determine the power saving level to be used for the network side device in the future time period, and further a corresponding power saving operation can be performed for the network side device. For example, the traffic for the future time period predicted by the method described with reference to FIG. 5 can be combined with the method described with reference to FIG. 11 below to determine the power saving level to be used for the network side device in the future time period. In some cases, the method for determining the power saving level may not be appropriate, for example, not very accurate in terms of the vicinity of the boundary value of the power saving level. Based on this, the present disclosure provides a further improvement of the method described with reference to FIG. 5.

According to an embodiment of the present disclosure, the method described with reference to FIG. 5 further obtains power saving level information for the network side device corresponding to the traffic in the past time period, and information on the traffic in the past time period, the time indicator information, the sector direction indication information, and the power saving level information, and predicts traffic information for a future time period using the neural network model.

As an example, the power saving level information may be predetermined Energy Saving (ES) level information, for example, as shown in Table 4 below, where n and N are positive integers.

In one embodiment of the present disclosure, the power saving level information can also be used as input data to train a neural network. When the trained neural network is actually used, the traffic information for the predicted future time period includes the power saving level information corresponding to the future time period, thereby performing a corresponding power saving operation on the network side device.

As an example, power saving level information corresponding to each traffic can be added to the data used in training the neural network. As shown in FIG. 7A, traffic T ₀ corresponds to ES level ES ₀ in Table 4. Traffic T ₁ corresponds to ES level ES ₁ in Table 4. Traffic T _n corresponds to ES level ES _n in Table 4. The ES ₀ to ES _n are also used as data to train the neural network. Thus, when the trained neural network is actually used, the traffic information for the predicted future time slot may include power saving level information corresponding to the future time slot.

As another example, in some cases, it may be necessary to know only whether the current power saving level has decreased or increased compared to the power saving level corresponding to the future time period. In a scene where only the change in the power saving level is required, as shown in FIG. 7B, the neural network can be trained with the power saving level change information used in training the neural network. Specifically, assume that the ES level corresponding to traffic T ₀ changes to 0. The ES level corresponding to traffic T ₁ is (ES ₁ -ES ₀ ). The ES level corresponding to traffic T _n is (ES _n -ES _n-1 ). The 0, (ES ₁ -ES ₀ ) ... (ES _n -ES _n-1 ) are also used as data used in training the neural network to train the neural network. Therefore, when the trained neural network is actually used, the traffic information of the predicted future time period may include the power saving level change information corresponding to the future time period.

As another example, in some cases, it may be desired to know only the boundary value of the power saving level corresponding to the traffic in the future time period. In a scene where only the boundary value of the power saving level is desired to be known, as shown in FIG. 7C, the neural network can be trained using the boundary value of the power saving level corresponding to each traffic as data used in training the neural network. Specifically, the boundary value of the power saving level corresponding to the traffic T ₀ is [Boundary 0, Boundary 1] (e.g., [0, 100] in Table 4 above). The boundary value of the power saving level corresponding to the traffic T ₁ is [Boundary 1, Boundary 2] (e.g., [100, 200] in Table 4 above). The boundary value of the power saving level corresponding to the traffic T _n is [Boundary n-1, Boundary n] (e.g., [N-1, N] in Table 4 above). The boundary value corresponding to each traffic is also used as data used in training the neural network to train the neural network. Therefore, when the trained neural network is actually used, the traffic information of the predicted future time period may include the boundary value of the power saving level corresponding to the traffic in the future time period.

In the above three embodiments, when training the neural network model, the neural network model is pre-trained by adjusting the parameters of the neural network model based on the model prediction loss (e.g., the above loss function) of the neural network model calculated using the power saving level information, and the traffic information for the future time period predicted by the trained neural network during actual use includes power saving level information corresponding to the future time period, so that the network side device performs the corresponding power saving operation.

In all three of the above examples, the neural network model is pre-trained by increasing the amount of data used to train the neural network, and adjusting the parameters of the neural network model based on the model prediction loss of the neural network model calculated using the power saving level information.

By improving the model prediction loss without increasing the data used to train the neural network, traffic information for future time periods predicted when the trained neural network is actually used includes power saving level information corresponding to the future time periods, thereby allowing the network side device to perform corresponding power saving operations.

In another embodiment of the present disclosure, the model prediction loss is calculated by calculating the model prediction loss of the neural network model based on traffic range information corresponding to the power saving level information and traffic information predicted by the neural network model.

As an example, the model prediction loss is an MSE loss function as shown in FIG. 8A and FIG. 8B. FIG. 8A shows a schematic diagram of the loss function before improvement according to an embodiment of the present disclosure. FIG. 8B shows a schematic diagram of the loss function after improvement according to an embodiment of the present disclosure.

Referring to FIG. 8A, when training the neural network before being refined, the MSELOSS of the neural network completes the training of the neural network by comparing the predicted traffic T' _n (n=0, 1...m, where n and m are both integers) with the actual traffic T _n (n=0, 1...m, where n and m are both integers) and adjusting the weights of the neural network based on a loss function.

Referring to FIG. 8B, when training the neural network after it has been improved, first determine the corresponding power saving level of the actual traffic corresponding to each predicted traffic, for example, the power saving level corresponding to the actual traffic T _n (n=0, 1...m, where n and m are all integers) is ES _n (n=0, 1...m, where n and m are all integers).

Then, the traffic range corresponding to the power saving level is determined; for example, the traffic range corresponding to the determined ES2 is 100 to 200 (as shown in Table 4 above).

A representative value within the range is then selected as the traffic value corresponding to the power saving level. For example, a median value mid _n (n=0, 1...m, where n and m are both integers) within the range is selected as the traffic value corresponding to the power saving level (as shown in FIG. 8B). Alternatively, an extreme value within the range may be selected as the traffic value corresponding to the power saving level. For example, if the range is 0-100, the minimum value 0 may be selected as the traffic value corresponding to the power saving level. If the range is 100-200, the minimum value 100 may be selected as the traffic value corresponding to the power saving level. Furthermore, for example, if the range is 0-100, the maximum value 100 may be selected as the traffic value corresponding to the power saving level. If the range is 100-200, the maximum value 200 may be selected as the traffic value corresponding to the power saving level.

Finally, the MSELOSS of the neural network completes the training of the neural network by comparing the predicted traffic (n=0, 1...m, where n and m are both integers) with the traffic value corresponding to the selected power saving level (e.g., mid _n (n=0, 1...m, where n and m are both integers)) and adjusting the weights of the neural network based on the loss function.

With the above improvements, when the trained neural network model is actually used, the traffic information for the predicted future time periods can include power saving level information corresponding to the future time periods, thereby performing corresponding power saving operations on the network side devices.

Note that when training a neural network, a large amount of data is used for training, and the above-mentioned training neural network embodiment can well mine rules based on the data, but cannot combine human experiential knowledge, thereby further improving the trained neural network to obtain more favorable results. For the above case, the present disclosure provides a further improvement of the method described with reference to FIG. 5.

Specifically, when training the neural network, the neural network is trained based on information on traffic in the historical time period and information on traffic in the adjacent time period used in the method described with reference to FIG. 1 above, and at least one of the first coefficients (e.g., α, β, γ, δ described above), and a small amount of traffic information in the past time period used in the manner described with reference to FIG. 5 above, predicted time marker information, and predicted sector direction information, for example, a manner in which a neural network is adopted and trained in combination can be adopted.

When training the neural network, the small amount of past traffic information, the predicted time sign information, and the predicted sector direction information are input to the first neural network for training, and the weight used for training using the first neural network is _αm (m=0, 1, 2...n, where m and n are both integers).

At least one of the information regarding traffic in the historical time period and the information regarding traffic in the adjacent time period used in the method described with reference to Figure 1 above, and the first coefficients (e.g., the above-mentioned α, β, γ, δ) is given a predetermined weight, for example β _m (m = 0, 1, 2...n, where m and n are both integers).

The two weights α _m and β _m are input to a second neural network for training to obtain weights α _m and β _m (denoted as p). A fusion weight of the whole neural network including the first neural network and the second neural network can be obtained based on the weights α _m , β _m and p. The second neural network may be any suitable conventional neural network, such as a Convoluted Neural Network (CNN), a Recurrent Neural Network (RNN), etc.

Finally, the loss function is minimized by adjusting the fusion weights of the entire neural network including the first neural network and the second neural network, thereby completing the training of the neural network.

According to an embodiment of the present disclosure, fusion weights can be obtained according to different needs.

Figures 9A-9C show schematic diagrams for obtaining fusion weights according to an embodiment of the present disclosure.

For example, as shown in Figure 9A, the weights α _m and β _m are input into the second neural network, and then the weight p is obtained. Then, the fusion weight is obtained based on α _m , β _m , and p.

As another example, as shown in Fig. 9B, after inputting the weights _αm and _βm into the second neural network, we obtain a set of corresponding weights _pn (m = 0, 1, 2...n, where m and n are both integers), and then obtain the fusion weights based on _pn .

As another example, as shown in FIG. 9C, after inputting the weights α _m and β _m into the second neural network, the fusion weight w _n (m=0, 1, 2...n, where m and n are both integers) is directly obtained.

According to an embodiment of the present disclosure, the information on traffic in the past time periods may include multiple sets of traffic data for the past time periods, and the method described with reference to FIG. 5 may further include normalizing the traffic data for each past time period in the multiple sets by dividing the traffic data for each past time period by the maximum value in the multiple sets, and predicting traffic information for a future time period using the neural network model based on the normalized traffic data. Alternatively, the traffic data in each set may be normalized by dividing the traffic data for each set by the maximum value in the respective set, and predicting traffic information for a future time period using the neural network model based on the normalized traffic data.

In the method and improved method described with reference to FIG. 5 above, when training a neural network, the neural network performs a normalization process on the data in batches, and each batch of data is composed of several random data. For example, the data in each batch may be n pieces (n is an integer) of traffic for a past time period as shown in FIG. 6A. That is, the information on traffic for a past time period for training each batch may include multiple sets of traffic data for a past time period.

Figures 10A-10B show schematic diagrams of the data normalization process when training a neural network according to an embodiment of the present disclosure.

10A, the batch of data input to the neural network for training includes n sets (i.e., 1-n, where n is an integer). For the i-th set, the data of the set may include traffic in the past 72 time periods, i.e., χ _0,0 to χ _0,71 . Before improving the aspect of the normalization process, the neural network divides all the traffic data in each time period in the batch of data by the sum of all the traffic data in the batch of data to achieve the purpose of the normalization process, as shown in FIG. 10A.

The value of each traffic data is relatively large (e.g., reaching hundreds of thousands), so the sum of the traffic data of the batch is large. The traffic data in each time period in the batch data is divided by the sum of all the traffic data in the batch data, and the resulting data is very small. Therefore, although the error generated in the subsequent training process is small, the predicted traffic error due to the small error caused by multiplying the sum is large.

In contrast, the improved method of the present disclosure achieves the purpose of normalization by dividing all traffic data in each time period in the batch of data by the maximum value (max) among all traffic data in the batch of data, as shown in FIG. 10A. Because the difference between the traffic data value in each time period in the batch of data and the maximum value is small, the trained neural network is good and fast.

In addition, the information on traffic in the past time period of each batch may include multiple sets of traffic data in the past time period. For example, the first batch of data is composed of three sets of data, 1, 3, and 5. The second batch of data is composed of three sets of data, 1, 2, and 4. When performing normalization processing based on the above improvement method, it is necessary to divide the first set of data in the first batch of data by the maximum value of the data in the batch, max 1, and to divide the first set of data in the second batch of data by the maximum value of the data in the batch, max 2. Since max 1 and max 2 are always different, when a neural network is trained, a similar set of data is divided by different values when normalized, which may reduce the prediction accuracy of the trained neural network.

For the above case, another improvement method of the present disclosure is to achieve the purpose of normalization by dividing all traffic data in each time period in each set of data in the batch of data by the maximum value (max) of all traffic data in the set of data, so that the trained neural network is faster and more accurate, as shown in FIG. 10B.

Note that the different improvements to the method described with reference to FIG. 5 above can be performed simultaneously or selectively, so that the resulting improved method includes at least one of the improvements described above.

A method that can be performed by a network side device in a wireless communication network provided by the present disclosure has been described above with reference to Figs. 5 to 10. The method provided by the present disclosure can quickly and accurately predict traffic information for a future time period based on the method provided by the present disclosure by using a neural network. The method provided by the present disclosure further introduces power saving level information when training the neural network, so that the method provided by the present disclosure can accurately predict power saving level information corresponding to traffic for a future time period, so that the network side device can more conveniently perform corresponding power saving operations. The method provided by the present disclosure further incorporates the use of weights, so that the method provided by the present disclosure can further improve the accuracy of the predicted traffic information for a future time period. The method provided by the present disclosure further improves the normalization aspect, so that the method provided by the present disclosure can more quickly and accurately predict traffic information for a future time period. Furthermore, the network side device only provides some resources that can process the traffic and turns off other resources, thereby saving power consumption in the network side device.

The network side device and the corresponding method described with reference to the above drawings can accurately predict traffic in a future time period, so that the network side device only provides some resources that can handle the traffic and turns off other resources, thereby achieving power consumption savings in the network side device for the entire time period.

In addition, although the above method provided by the present disclosure can accurately predict traffic in a future time period, there are cases where the network side device can only perform a relatively simple and rough power saving control operation. In such a situation, according to another aspect of the present disclosure, a method for determining a power saving level to be used in a network side device in a future time period is provided, so that the network side device can accurately select a power saving level, thereby further improving the power saving effect or data throughput. This will be described in detail below with reference to Figures 11 to 14.

Figure 11 shows a flowchart of another method of wireless communication in a network side device according to an embodiment of the present disclosure. The method shown in Figure 11 can be executed by the network side device. As an example, the network side device may be a base station. Or, the network side device may be the OAM module or SMO module that interacts with the outside of the base station and the base station. When the network side device is the OAM or SMO module, the OAM or SMO module needs to transmit information on the power saving level obtained by executing the method provided by the present disclosure, traffic information for the predicted future time period, and resource information corresponding to the power saving level, for example, to the base station, and the base station performs the corresponding power saving operation.

As shown in FIG. 11, in step S1110, the network side device can determine a power saving level to be used in the network side device in a future time period based on traffic in the future time period and a predetermined power saving level table, where the power saving level table may include a plurality of power saving levels corresponding to the wireless transmission capability of the network side device and the resources used when the network side device transmits wirelessly.

As an example, traffic for a future time period can be acquired based on the method described with reference to Figures 1 to 4E above or the method described with reference to Figure 5 above. The network side device can acquire traffic for the future time period in real time or at a preset time.

In the embodiment of the present disclosure, the resources used by the network side device for wireless transmission may be at least one of time domain resources, frequency domain resources, spatial domain resources, and power domain resources, or a combination thereof. The time domain resources may include resources such as subframes, slots, microslots, and symbols. The frequency domain resources may include resources such as component carriers (CCs), a portion of bandwidth (BWP), subbands, reference signal groups, and subcarrier groups. The spatial domain resources may include resources such as antenna elements, transmit/receive units (TxRUs), and virtual antenna ports. The power domain resources may include resources required to adjust at least one of the maximum transmit power, power spectral density (PSD), etc. of one or more channels of transmitted data (Data), control (Control), reference signal (RS), synchronization signal block (SSB), etc.

For example, the predetermined energy saving level table may be determined based on the wireless transmission capability of the network side device and the resources used when the network side device performs wireless transmission. For example, the energy saving (ES) level table shown in Table 5 below may be determined in advance.

Here, the ES level is divided into four levels, namely ES1 to ES4. The ES level may be divided into any number of levels according to actual needs.

In Table 5, the power saving levels ES1 to ES4 are sorted from highest to lowest as follows: ES1, ES2, ES3, ES4. The power saving level ES1 may indicate the lowest energy consumption in the network side device. The power saving level ES4 may indicate the highest energy consumption in the network side device. The power saving levels ES2 and ES3 indicate that the power consumption in the network side device is between ES1 and ES4. Other manners may be designed to indicate high or low energy consumption in the network side device.

The resources in Table 5 may indicate resources used by the network side device during wireless transmission. For example, at least one of time domain resources, frequency domain resources, spatial domain resources, and power domain resources, or a combination thereof.

The radio transmission capacity in Table 5 can indicate the radio transmission capacity of the network side device when using the corresponding resource. As an example, the transmission capacity of the network side device may be the maximum value, average value, weighted average value, etc. of the transmission data amount of the network side device in a certain time period in the past. For example, the transmission capacity of the network side device may be the maximum value or weighted average value of the data amount in a unit time period of a component carrier in a past time period. Also, for example, the transmission capacity of the network side device may be the product of the average value of all UE transmission rates on a specific component carrier in a certain time period and the bandwidth of the specific component carrier. Also, for example, the transmission capacity of the network side device may be the maximum rate of a component carrier for a specific modulation and coding scheme (MCS) or rank number. Also, for example, the transmission capacity of the network side device may be the ratio between the actual transmission data amount and the resource usage rate.

When the resources include multiple component carriers available to the network side device, according to one embodiment of the present disclosure, each component carrier can have a priority corresponding to the component carrier, where in the power saving level table, the higher the energy consumption, the more low priority component carriers the power saving level turns on.

12A to 12B show schematic diagrams of resources used by a network side device according to one embodiment of the present disclosure. The resources may also include four component carriers available to the network side device. As shown in FIG. 12A, the four component carriers are denoted as CC1, CC2, CC3, and CC4, respectively. CC1 has a center frequency of 800 M, a beam width of 55 M, and a beam coverage radius of 902 meters (m). CC2 has a center frequency of 1.5 G, a beam width of 50 M, and a beam coverage radius of 654 m. CC3 has a center frequency of 1.7 G, a beam width of 45 M, and a beam coverage radius of 613 m. CC4 has a center frequency of 2.0 G, a beam width of 45 M, and a beam coverage radius of 564 m. The lower the center frequency, the lower the priority of the component carrier. In other words, the priority of the four component carriers is CC4, CC3, CC2, and CC1, in descending order. The priority of the component carriers may also be determined based on factors such as the coverage range of the component carriers, the traffic permitted for transmission, and the power consumed for traffic transmission.

In this case, the contents of the resource column in the above Table 5, specifically, for the above component carriers CC1 to CC4, are as shown in the following Table 6.

In the power saving level table 6 above, the power saving level ES4 with the highest energy consumption turns on all four component carriers. The power saving level ES1 with the lowest energy consumption activates only component carrier CC1 among them. The number of component carriers turned on by ES2 and ES3, which have energy consumption between ES4 and ES1, is between ES4 and ES1. Therefore, a power saving level with high energy consumption activates more low priority component carriers than a power saving level with low energy consumption.

The wireless communication network may also include one or more sectors, which are also resources available to the network side device. The sectors may be uniform, in which case the size of the coverage area of each sector is the same. The sectors may be non-uniform, in which case the size of the coverage area of each sector may differ. For the case of uniform sectors, the determination of the radio transmission capability of the base station in Tables 5 and 6 above is described below in combination with Figures 12A and 12B. For the case of non-uniform sectors, the determination of the radio transmission capability of the base station in Tables 5 and 6 above is described below with reference to Figures 13A to 13C.

Next, referring to FIG. 12A, in the schematic diagram shown in FIG. 12A, the network side device has uniform sectors for each component carrier. That is, as shown in FIG. 12A, the network side device has six uniform sectors for each component carrier, which correspond to beams in the directions of 30 degrees (deg), 90 degrees, 150 degrees, 210 degrees, 270 degrees, and 330 degrees, respectively. In this case, the wireless transmission capacity of the base station in Table 3 or 4 above can be determined by selecting the amount of transmission data in the sector corresponding to any one of the angles and indicating the amount of transmission data for the remaining other angles.

The following explains the amount of transmission data corresponding to 30 degrees as an example.

The maximum amount of data transmitted per hour on a component carrier during the previous day is regarded as the wireless transmission capacity of the network side device.

First, it is necessary to statistic the maximum transmission data amount of each component carrier within one hour within a predetermined time (for example, the previous day). The statistical results are shown in Table 7 below. Table 7 shows the transmission data amount at 30 degrees for each component carrier that has been statistically collected. Table 7 also shows the transmission data amount at 90 degrees for each component carrier that has been statistically collected. It is possible to statistic the transmission data amount of all component carriers according to actual needs. It is possible for the transmission data amount to be an empirical value determined according to actual conditions.

Next, the following allocation operation is performed based on the priority of the component carrier. Specifically, CC1 is allocated to correspond to power level ES1, which indicates the lowest energy consumption, and in this case, the data transmission amount corresponding to CC1 (i.e., 5842) is determined as the wireless transmission capacity of the network side device. CC1 and CC2 are both allocated to correspond to power saving level ES2, and the sum of the data transmission amount corresponding to CC1 (i.e., 5842) and the data transmission amount corresponding to CC2 (i.e., 6786) (i.e., 5842 + 6786 = 12628) is determined as the wireless transmission capacity of the network side device. CC1 to CC3 are all assigned to correspond to power saving level ES3, and the sum of the data transmission amount corresponding to CC1 (i.e., 5842), the data transmission amount corresponding to CC2 (i.e., 6786), and the data transmission amount corresponding to CC3 (i.e., 10739) (i.e., 5842+6786+10739=23367) is determined as the wireless transmission capacity of the network side device. CC1 to CC4 are all assigned to correspond to power saving level ES4, and the sum of the data transmission amount corresponding to CC1 (i.e., 5842), the data transmission amount corresponding to CC2 (i.e., 6786), the data transmission amount corresponding to CC3 (i.e., 10739), and the data transmission amount corresponding to CC4 (i.e., 9148) (i.e., 5842+6786+10739+9148=32515) is determined as the wireless transmission capacity of the network side device. The wireless transmission capabilities of the network side devices determined above are recorded as the data corresponding to Tables 5 and 6 above.

In this way, the above operation allows a power saving level table for uniform sectors to be pre-determined. In this case, a columnar schematic diagram of the power saving levels is as shown in 12B. As shown in Figs. 12A and 12B, at power saving level ES1, only CC1 is turned on. At power saving level ES2, CC1 and CC2 are turned on. At power saving level ES3, CC1 to CC3 are turned on. At power saving level ES4, all component carriers are turned on, i.e., CC1 to CC4. Thus, at power saving level ES4, the energy consumption of the network side device is the highest. At power saving level ES1, the energy consumption of the network side device is the lowest.

The above describes a method for determining the wireless transmission capability of a network side device using uniform sectors as an example. However, in some cases, the sector distribution of a base station may be non-uniform. In this case, the present invention further provides a method for determining the wireless transmission capability of a network side device for a network side device having non-uniform sectors. This will be described with reference to Figures 13A to 13C.

As shown in FIG. 13A, five component carriers are used in the network side device, which are denoted as CC1, CC2, CC3, CC4, and CC5. CC1 has a center frequency of 700 M, a beam coverage radius of 966 meters (m), and is transmitted in one sector corresponding to 0 degrees. CC2 has a center frequency of 800 M, a beam coverage radius of 902 m, and is transmitted in sectors of 0 degrees, 150 degrees, and 270 degrees, respectively. CC3 has a center frequency of 1.5 G, a beam coverage radius of 654 m, and is transmitted in sectors corresponding to 0 degrees, 30 degrees, 60 degrees, 150 degrees, and 270 degrees, respectively. CC4 has a center frequency of 1.7 G, a beam coverage radius of 613 m, and is transmitted in sectors corresponding to 0 degrees, 60 degrees, 150 degrees, and 270 degrees, respectively. CC5 has a center frequency of 2.0 G, a beam coverage radius of 564 m, and is transmitted in sectors corresponding to 0 degrees, 30 degrees, 60 degrees, 150 degrees, and 270 degrees, respectively. Also, as shown in FIG. 7A, the center points of the sectors corresponding to CC5 and CC2 are offset 39 m to the right compared to the center points of the sectors corresponding to CC1, CC3, and CC4. Also, as shown in FIG. 7A, CC5 corresponds to five sectors, CC4 corresponds to four sectors, CC3 corresponds to five sectors, CC2 corresponds to three sectors, and CC1 corresponds to one sector.

In the situation where the network side device shown in FIG. 13A has uneven sectors, the amount of transmission data corresponding to the uneven adjacent sectors can be integrated, and the power saving performance of the network side device can be improved by bringing different component carriers closer to the coverage in the sector direction after integration.

For example, first, it is necessary to count the maximum transmission data amount of each component carrier in one hour for a predetermined time (for example, the past 5 hours). The statistical result is as shown in Table 8 below, where ● indicates that the frequency CC is covered by a sector in the direction corresponding to the angle, and data can be counted. ○ indicates that the frequency CC is not covered by a sector in the direction corresponding to the angle, but may be covered by a sector in the direction corresponding to the adjacent angle. The transmission data amount may be an empirical value determined according to the actual situation.

The data for 0 degrees, 30 degrees and 60 degrees are then combined to obtain the results shown in Table 9 below.

The effect of the combined beam direction is shown in Figure 13B.

Finally, the wireless transmission capability of the network side device is determined based on the combined data. The process of determining the wireless transmission capability of the network side device is similar to the process described with reference to Figures 12A to 12B above, and further description is omitted. The bar graph of the power saving level in this case can be represented as shown in Figure 13C.

In yet another embodiment of the present disclosure, step S1110 may further include a step of adjusting the traffic of the future time period according to one or more of a forecast error, a target power saving amount, or a target data throughput, and determining a power saving level to be used in the network side device in the future time period based on the adjusted traffic of the future time period and a predetermined power saving level table.

In one embodiment of the present disclosure, there may be a certain error between the traffic in the future time slot and the actual traffic. Although the error is small, the traffic in the future time slot can be adjusted to more accurately determine the power saving level of the network side device in the future time slot. For example, it is determined based on experience or statistics that the relative error between the traffic in the future time slot and the actual traffic is x%, in which case the traffic in the future time slot can be adjusted according to the following formula:

b=a/(1-x%) (4)
Here, a is the traffic in the future time slot before adjustment, and b is the traffic in the future time slot after adjustment.

In another embodiment of the present disclosure, the higher the power saving level of the network side device in the determined future time period, the fewer resources will be used by the network side device, and thus the data throughput will be reduced, and a scaling factor can be determined based on at least one of the target power saving amount and the target data throughput when adjusting the traffic for the future time period. For example, statistical data indicates that when the scale factor x=1.1, the degradation of the data throughput in the network side device can be controlled within 2%, so that the service impact to the UE is within a controllable range when adjusting the power saving level. In this case, the traffic for the future time period can be adjusted according to the following formula:

b = a * x (5)
Here, a is the traffic in the future time slot before adjustment, and b is the traffic in the future time slot after adjustment.

In yet another embodiment of the present disclosure, the scaling coefficient can be further obtained based on a target scene. For example, the target scene may include scenes such as urban macro base station (Umi, Urban Microcell), urban macro base station (Uma, Urban Macrocell), rural (Rma, Rural Macrocell), indoor hotspot (InH, Indoor Hotxt), etc. A scaling coefficient mode or a scaling coefficient table may be designed in advance based on the different scenes and experiences, and then a specific scaling coefficient may be selected based on the target data throughput and target scene requirements that are achieved according to actual commercial needs. By considering one or more of the prediction error, the target power saving amount, the target data throughput, or the target scene, the minimum power saving level required for the network side device when transmitting all traffic based on the traffic of the future time period may be selected as much as possible.

As shown in FIG. 14, as the scaling factor increases, the traffic in the future time period is constantly amplified, so that a lower power saving level is selected. That is, as the scaling factor increases, the network side device uses more resources to transmit data, so that the data throughput reduction rate in the network side device is continuously reduced, and the power consumption in the video unit (AAU) in the network side device is increased accordingly. For example, when the scale factor is 1.1, the data throughput in the network side device is reduced by 1.7%, and the power consumption in the video unit (AAU) in the network side device is increased to 53.9%.

Returning to FIG. 11, in step S1120, the network side device can perform power saving operations based on the determined power saving level.

As an example, the power saving operation performed by the network side device may include closing resources that do not need to continue to be used by waiting for shutdown. For example, if it is determined that the power saving level of the network side device in the future time period is ES3 in Table 6 above, the network side device will turn off the previously activated CC4 in the future time period. When specifically turning off CC4, no new UE accesses the CC4 to be turned off, and CC4 is turned off after transmitting all the data transmitted by CC4, for example, when the amount of data transmitted by CC4 is small (there is a certain waiting delay in such a situation).

As another example, the power saving operation performed by the network side device may include turning off resources that do not need to continue to be used by an uninstallation method. For example, if it is determined that the power saving level of the network side device in a future time period is ES3 in Table 6 above, the network side device turns off the previously activated CC4 in the future time period. When specifically turning off CC4, no new UEs access the CC4 to be turned off, and data transmitted by CC4 can be uninstalled and transmitted to any one or more of CC1-CC3, and finally CC4 is turned off. The above situation applies to a situation where there is a lot of traffic transmitted by the CC, and this uninstallation method results in a certain signaling overhead and signaling interaction delay.

As another example, the power saving operation of the network side device may include flexibly operating one or two of the standby off method and the offload method based on the traffic transmitted by the CC waiting to be off and the number of UEs.

As another example, the power saving action taken by the network side device may include selecting the component carriers that need to be closed by combining one or more of the cell characteristics, the traffic and type of the connected UE, and the priority of the component carrier tasks.

Depending on the cell characteristics, CCs with small coverage (e.g., CCs with high frequency and narrow width, e.g., CC4 described with reference to FIG. 12A above or CC5 described with reference to FIG. 13A above) can be preferentially turned off, CCs with small amounts of transmitted data can be preferentially turned off, CCs with high transmission power can be preferentially turned off, or CCs with low unit power transmission capacity can be preferentially turned off. Depending on the actual application, CCs with large coverage, large amounts of transmitted data, low transmission power, or high unit power transmission capacity can be preferentially closed.

For the traffic and type of connected UEs, the CC with the least traffic transmitted by it can be preferentially turned off, the CC with the least number of UEs that set the CC as a primary cell (Cell, Pscell) or secondary cell (Cell, Scell) (in such a case, the signaling overhead caused is small), the CC with the least number of UEs that set the CC as a Pscell (in such a case, the signaling overhead caused is small), or the CC with the least number of low terminated UEs can be preferentially turned off. Alternatively, depending on the actual application, the CC with the largest amount of traffic transmitted by it and the largest number of UEs or the largest number of low terminated UEs can be preferentially turned off.

In relation to the priority of the component carrier task, a CC with a low priority, for example CC1 described with reference to FIG. 12A or FIG. 13A above, can be preferentially turned off. Or, a CC with a high priority can be preferentially turned off.

Above, another method provided by the present disclosure has been described in detail with reference to Figures 11 to 14. The above method provided by the present disclosure can accurately determine the power saving level of the network side device in a future time period based on the traffic of the future time period and a predetermined power saving level table, so that the network side device can accurately select the power saving level and accurately reserve only the minimum resources required to transmit traffic when operating based on the determined power saving level, thereby achieving an optimal power saving effect.

The method used for wireless communication in the network side device of the present disclosure has been described above with reference to Figures 1 to 14, and below, the network side device in the wireless communication network provided by the present disclosure will be described with reference to Figures 15 to 17. Network side device 1500 shown in Figure 15 corresponds to the method used for wireless communication in the network side device described with reference to Figures 1 to 4E, network side device 1600 shown in Figure 16 corresponds to the method used for wireless communication in the network side device described with reference to Figures 5 to 10B, and network side device 1700 shown in Figure 17 corresponds to the method used for wireless communication in the network side device described with reference to Figures 11 to 14. Therefore, in order to provide a simple description here, detailed description of the same contents will be omitted.

FIG. 15 is a block diagram of a network side device 1500 in a wireless communication network according to an embodiment of the present disclosure. FIG. 16 is a block diagram of a network side device 1600 in a wireless communication network according to an embodiment of the present disclosure. FIG. 17 is a block diagram of a network side device 1700 in a wireless communication network according to an embodiment of the present disclosure.

Referring to FIG. 15, the network side device 1500 may include a memory unit 1510 and a processing unit 1520. In this example, the network side device 1500 includes a memory unit 1510 and a processing unit 1520. Note that the network side device 1500 may further include other components, but these components are not related to the contents of the embodiments of the present disclosure, so illustration and description thereof will be omitted here.

As an example, the network side device 1500 may be a base station. Or, the network side device 1500 may be an OAM module or an SMO module. When the network side device is the OAM or SMO module, the OAM or SMO module needs to transmit traffic information for the predicted future time period to, for example, a base station by executing the method provided by the present disclosure, and the base station performs the corresponding operation.

As shown in FIG. 15, the memory unit 1510 may store information regarding traffic for a historical time period and information regarding traffic for an adjacent time period, where the adjacent time period is closer to a future time period than the historical time period.

For example, the information about the traffic of the historical time period and the information about the traffic of the adjacent time period can be stored in an internal memory of the network side device, such as a memory, such as a random access memory (RAM), a read-only memory (ROM), and an electrically erasable programmable ROM (EEPROM). Or, the information about the traffic of the historical time period and the information about the traffic of the adjacent time period can be stored in an external memory or cloud that interacts with the network side device.

As an example, a time period may be a time in hours or minutes. For example, a time period may be 1 hour or 2 hours. Also, a time period may be 15 minutes or 30 minutes.

In an embodiment of the present disclosure, the future time period may be a time period in the predicted future. For example, the future time period may be the time period immediately following the current time, e.g., if the current time is 12:00, the future time period may be 13:00-14:00. Also, for example, the future time period may be a time period separated from the current time by a predetermined time, e.g., if the current time is 12:00, the future time period may be 15:00-16:00, which is two hours or less apart.

The historical time band may be a time band prior to the current day on which the future time band exists. For example, the historical time band may be a time band of the day before the day on which the future time slot exists, or of the past few days. Accordingly, the traffic of the historical time band may be traffic generated within the time band on the day before the day on which the future time slot exists. For example, if the time band is one hour, the traffic of the historical time band may be traffic generated within each hour during the day before the day on which the future time slot exists.

The adjacent time period may also be a time period prior to a future time period within the current day. For example, the adjacent time period may be one or more time periods prior to a future time period within the current day. For example, if a future time period is from 4:00 pm to 5:00 pm on the current day, the adjacent time period may be from 3:00 pm to 4:00 pm on the current day, or from 10:00 am to 11:00 am on the current day, etc. Accordingly, the traffic of the adjacent time period may be traffic generated within a time period prior to the future time period within the current day. For example, if a time period is one hour, the traffic of the adjacent time period may be traffic generated within a time period prior to the future time period within the current day.

The processing unit 1520 may predict traffic for a future time period based on information about traffic for the historical time period and information about traffic for the adjacent time period, or based on stored information about traffic for multiple adjacent time periods.

For example, the traffic information for the historical time period and the traffic information for the adjacent time period can be acquired in real time by the network side device. Alternatively, the traffic information for the historical time period and the traffic information for the adjacent time period can be acquired by the network side device at a preset time according to actual needs.

According to one embodiment of the present disclosure, the processing unit 1520 may predict traffic for the future time period based on information about traffic for a time period in the history time period that corresponds to the future time period, information about traffic for a time period in the history time period that corresponds to a particular adjacent time period, and information about traffic for the particular adjacent time period.

According to another embodiment of the present disclosure, the historical time band and the adjacent time band include a related time band of the future time band and a reference time band corresponding to the future time band and the related time band, respectively, or a related time band including the future time band in the adjacent time band and a reference time band corresponding to the future time band and the related time band, respectively. In this case, the processing unit 1520 may further predict the traffic of the future time band based on a first coefficient indicating a relevance between the traffic of the future time band and the traffic of the related time band and the traffic of the reference time band.

For example, referring to FIG. 2A above, the historical time period may be the time period 10:00-11:00, 11:00-12:00, or 12:00-13:00 during the previous day. The adjacent time period may be the time period 10:00-11:00, or 11:00-12:00 during the current day.

As an example, the future time period may be the 12:00-13:00 time period of the current day, and the relevant time period of the future time period included in the historical time period and the adjacent time period may be the 12:00-13:00 time period of the previous day. The reference time periods corresponding to the future time period and the relevant time period included in the historical time period and the adjacent time period may be the 11:00-12:00 time period of the current day and the 11:00-12:00 time period of the previous day, respectively. The first coefficient may indicate the correlation between the traffic of the 12:00-13:00 time period of the current day and the 12:00-13:00 time period of the previous day, the traffic of the 11:00-12:00 time period of the current day and the traffic of the 11:00-12:00 time period of the previous day.

As another example, the future time period may be the 12:00-13:00 time period of the current day, and the related time period of the future time period included in the historical and adjacent time periods may be the 11:00-12:00 time period of the current day. Also, the 12:00-13:00 time period of the previous day and the 11:00-12:00 time period of the previous day may be the 12:00-13:00 time period of the previous day, respectively. The first coefficient may indicate a correlation between traffic in the 12:00-13:00 time period of the current day, traffic in the 11:00-12:00 time period of the current day, traffic in the 12:00-13:00 time period of the previous day, and traffic in the 11:00-12:00 time period of the previous day. For example, in this case, the first coefficient can indicate the correlation between the difference between the traffic during the 12:00-13:00 time slot of the current day and the traffic during the 11:00-12:00 time slot of the current day, and the difference between the traffic during the 12:00-13:00 time slot of the previous day and the traffic during the 11:00-12:00 time slot of the previous day. Alternatively, the first coefficient can indicate the correlation between the ratio between the traffic during the 12:00-13:00 time slot of the current day and the traffic during the 11:00-12:00 time slot of the current day, and the ratio between the traffic during the 12:00-13:00 time slot of the previous day and the traffic during the 11:00-12:00 time slot of the previous day.

As another example, the future time period may be a time period of 12:00-13:00 on the current day, and the related time period of the future time period included in the adjacent time period may be a time period of 11:00-12:00 on the current day. The reference time periods corresponding to the future time period and the related time period included in the adjacent time period may be a time period of 11:00 to 12:00 on the current day and a time period of 10:00 to 11:00 on the current day. The first coefficient may indicate a correlation between traffic during the time period of 12:00-13:00 on the current day, traffic during the time period of 11:00-12:00 on the current day, traffic during the time period of 11:00-12:00 on the current day, traffic during the time period of 10:00-11:00 on the current day.

Here, the formulas corresponding to α, β, γ, and χ are as shown in Table 1.

Note that in the example of the first coefficient described above with reference to FIG. 2A and Table 1, the example is the day before the current day, but the above method may also be applied to an example that uses a certain day or days in the past from the current day.

The first coefficient can be obtained by statistically analyzing historical traffic data. This has been described with reference to, for example, FIGS. 4A to 4E. Alternatively, the first coefficient can be preset as a fixed value according to actual needs. The larger the absolute value of the first coefficient, the higher the relevance, and the smaller the absolute value of the first coefficient, the lower the relevance. For example, a value of the first coefficient of +1 or -1 indicates high correlation, and a value of the first coefficient of 0 indicates low correlation. The value of the first coefficient may be different for different time periods in a day. In addition, the values of the first coefficients (e.g., the above α, β, γ, and δ) obtained based on different methods may be different for a certain time period. Therefore, according to another embodiment of the present disclosure, when actually predicting traffic for a future time period, any of the candidate coefficients obtained based on different methods according to the actual situation can be selected and predicted as the first coefficient. For example, the network side device may select the candidate coefficient having the highest correlation with the future time period from among a plurality of candidate coefficients as the first coefficient.

Figure 3 above shows the change in the values of coefficients α, β, and γ throughout the day, using coefficients α, β, and γ as examples. As shown in Figure 3 above, between 5:00 a.m. and 6:00 a.m., the absolute value of coefficient α is clearly smaller than the absolute values of coefficients β and γ, and is clearly larger than the absolute values of coefficients β and γ in other time period coefficients. In this case, processing unit 1520 may select the coefficient with the maximum correlation value as the first coefficient. For example, α, β, γ, and δ may be candidate coefficients for the first coefficient. Processing unit 1520 may select the maximum coefficient among candidate coefficients α, β, γ, and δ for the future time period for which traffic is to be predicted as the first coefficient.

Therefore, the network side device provided by the present disclosure can predict traffic for future time periods based on different candidate coefficients, making the network side device provided by the present disclosure flexible and versatile.

In addition, the network side device provided by the present disclosure can select the most relevant candidate coefficient to more accurately predict traffic changes in future time periods (e.g., increases or decreases in traffic in future time periods) and traffic in future time periods.

As can be seen from the detailed description of the network side device provided by the present disclosure, the network side device provided by the present disclosure can accurately predict traffic changes in future time periods by using the first coefficient, but there are situations in which the prediction of the magnitude of the change is accidentally inaccurate. For such situations, the present disclosure also provides further improvements to the network side device.

According to yet another embodiment of the present disclosure, the processing unit 1520 may further predict traffic for the future time period based on a second coefficient indicating a correlation between the magnitude of traffic for the relevant time period and the magnitude of traffic for a reference time period for the relevant time period.

For example, the relationship between the traffic magnitude of the relevant time band and the traffic magnitude of a reference time band for the relevant time band may be the ratio of the traffic magnitude of the relevant time band to the traffic magnitude of a reference time band for the relevant time band, or may be the ratio of the average traffic values or the ratio of the maximum traffic values therebetween.

For example, in the example shown in Table 2 above, the average value of traffic in a future time slot (e.g., 13:00-14:00) on the current day is denoted as x. The relevant time slot is the time slot of the previous day that corresponds to the future time slot, and its average value of traffic is denoted as a, with a magnitude of 10. The reference time slot of the relevant time slot is the time slot of the previous day that corresponds to the adjacent time slot, and its average value of traffic is denoted as d, with a magnitude of 3. Let the second coefficient be L. In such a case, the second coefficient L may be 10/3, i.e., the second coefficient L is equal to 3, in which case the disclosure of method 1 in Table 1 above can be transformed to x=a+L*α*(e-d). If α is 1, the traffic x of the future time slot predicted in this case is 30. Before transforming method 1, the unpredicted x=a+α*(e-d), i.e., x is 16. Then, the actual traffic in the future time slot is 30.

The relevant time slot may further be a time slot corresponding to a future time slot in the past two days, the traffic average value of which is 20, and the reference time slot of the relevant time slot may be a time slot corresponding to an adjacent time slot in the past two days, the traffic average value of which is 6. In such a case, the second coefficient L may be 20/6, i.e., the second coefficient L is equal to 3. In this case, the disclosure of method 1 in Table 1 above can be transformed to x=a+L*α*(e-d). If α is 1, the predicted traffic x in the future time slot in this case is 30. Before transforming method 1, the unpredicted x=a+α*(e-d), i.e., x is 16. Then, the actual traffic in the future time slot is 30.

Thus, by introducing the second coefficient, the network side device provided by the present disclosure can not only accurately predict traffic changes in future time periods, but also more accurately predict the magnitude of traffic changes in future time periods, thereby making the predicted traffic more accurate.

After accurately predicting the traffic of the future time band, the network side device only provides some resources that can handle the predicted traffic and turns off other resources, thereby saving power consumption in the network side device. For example, if the network side device can handle the traffic of the future time band with only three antennas, the network side device can turn off the remaining antennas and not operate the remaining antennas, thereby saving power consumption in the network side device. Furthermore, for example, if the network side device only needs two component carriers to handle the traffic of the future time band, the network side device does not transmit the remaining component carriers, thereby saving power consumption in the network side device.

The network side device in the wireless communication network provided by the present disclosure has been described above with reference to the attached drawings. The network side device provided by the present disclosure uses a first coefficient, which allows the network side device provided by the present disclosure to more accurately predict changes in traffic in future time periods. The network side device provided by the present disclosure also uses the second coefficient, which allows the network side device provided by the present disclosure to more accurately predict changes in the amount of traffic in future time periods. As a result, the network side device provided by the present disclosure can predict with high accuracy the traffic that the network side device needs to process in future time periods, and further, the network side device saves power consumption in the network side device by only providing some resources that can process the above traffic and turning off other resources.

Referring to FIG. 16, the network side device 1600 may include an acquisition unit 1610 and a processing unit 1620. In this example, the network side device 1600 includes an acquisition unit 1610 and a processing unit 1620. Note that the network side device 1600 may further include other components, but these components are not related to the contents of the embodiments of the present disclosure, and therefore will not be illustrated or described here.

As an example, the network side device 1600 may be a base station. Alternatively, the network side device 1600 may be an operation administration and maintenance (OAM) module or a service management and orchestration (SMO) module that interacts with the outside of the base station. When the network side device is the OAM or SMO module, the OAM or SMO module needs to transmit traffic information for a future time period predicted by executing the method provided by the present disclosure to, for example, the base station, and the base station performs a corresponding operation.

As shown in FIG. 16, the acquisition unit 1610 may be configured to acquire information regarding traffic in a past time period, time indicator information, and sector direction indication information, where the time indicator information may indicate whether the traffic in the past time period is a first time period or a second time period, and the sector direction indication information may indicate the sector direction of a cell corresponding to the network side device.

As an example, the information about the traffic of the past time period may be the information about the traffic of the historical time period described in FIG. 1 and the information about the traffic of the adjacent time period. The information about the traffic of the past time period may be stored in an internal memory of the network side device, for example a memory, such as a random access memory (RAM), a read-only memory (ROM) and an electrically erasable programmable ROM (EEPROM). Alternatively, the information about the traffic of the past time period may be stored in an external memory or cloud that interacts with the network side device.

For example, the first period may be a working day, e.g., one or all days from Monday to Friday. The second period may be a non-working day, e.g., one or all days from Saturday to Sunday.

For example, the sector direction may refer to the sector direction corresponding to each angle when combining uniform sectors as described in FIG. 12A, or the sector direction corresponding to each angle when combining non-uniform sectors as described in FIG. 13A. The traffic transmitted by the sector directions corresponding to different angles may be the same or different. Therefore, predicting future time zone traffic information based on the sector direction helps to make the predicted future time zone traffic information more accurate.

The processing unit 1620 may use a neural network model to predict traffic information for a future time period based on information about traffic for the past time period, the time sign information, and the sector direction indication information.

According to one embodiment of the present disclosure, the Neural Network (NN) model is a pre-trained neural network model. The Neural Network model may be any suitable conventional neural network model, such as a Convoluting Neural Network (CNN) model, a Recurrent Neural Network (RNN) model, etc.

The neural network model can be trained based on information about traffic over past time periods, time sign information, and sector direction indication information.

As an example, the information regarding traffic in the past time period may be traffic generated during each time of each day for the past 100 days. The time indicator information may be traffic indicating whether the traffic generated per day for the past 100 days is weekday traffic or holiday traffic. The sector direction indication information may indicate the sector direction of the cell corresponding to the network side device when generating the above traffic, for example, combining the sector direction corresponding to 60 degrees described in FIG. 12A.

As shown in Fig. 6A, suppose the training data is to predict traffic () within 5 hours on the 4th day (day 4). In this case, traffic 4 hours before the 4th day (i.e., _T1 to _T4 ) and traffic for each hour within three days before the 4th day (i.e., _T1 to _T24 on day 1, _T1 to _T24 on day 2, and _T1 to _T24 on day 3) may be selected as traffic for the past time period. Traffic for each hour within any one day or multiple days before the 4th day may also be selected.

Referring now to FIG. 6A, in this case, the time indicator information may be 0011, a total of four bits of information. A "0" in the first bit of "0011" may indicate that all traffic generated on day 1 is holiday traffic. A "0" in the second bit of "0011" may indicate that all traffic generated on day 2 is holiday traffic. A "1" in the third bit of "0011" may indicate that all traffic generated on day 3 is working day traffic. A "1" in the fourth bit of "0011" may indicate that all traffic generated on day 4 is working day traffic.

The sector direction indication information may be 6-bit information of 010000. This 6-bit information can indicate the directions corresponding to 30 degrees, 90 degrees, 150 degrees, 210 degrees, 270 degrees, and 330 degrees described in FIG. 12A from left to right, respectively. If the corresponding direction transmits traffic, the corresponding bit in the above 6-bit information can be set to 1. If traffic is not transmitted in the corresponding direction, the corresponding bit in the above 6-bit information can be set to 0. In this case, the above "010000" can indicate that traffic is transmitted only in the sector direction corresponding to 90 degrees.

Continuing with reference to FIG. 6A, after inputting the one training data into the neural network, the neural network obtains a predicted traffic T ^- ₅ . Based on the predicted traffic T ^- ₅ and the actual traffic _T5 at the time, the training of the neural network can be completed by adjusting the weight of the neural network based on the loss function.

After inputting the one training data into the neural network, the neural network obtains the predicted traffic T ^- ₅ , the predicted time indicator information and the predicted sector direction information, in which case, according to the predicted traffic T ^- ₅ and the actual traffic _T5 at the time, the predicted time indicator information and the actual time indicator information at the time, and the predicted sector direction information and the actual sector direction information at the time, the weight of the neural network is adjusted according to the loss function to complete the training of the neural network.

Furthermore, as shown in FIG. 6B, in order to improve the predictive effect of the trained neural network, in the actual training, a large amount of data similar to the one training data above from many days (e.g., as shown in Table 3 above) is first selected and input to the neural network for training.

In Table 3 above, the training data is used to train the neural network. The neural network is trained multiple times, and the validation data is used to validate the trained neural network each time, and only the neural network with the highest validation data is reserved during the training process. For example, if the current neural network has been trained 50 times, the highest validation data is the neural network trained 25th time. After the 51st training, if the probability of the neural network trained 51st time in the validation data is lower than the probability of the neural network trained 25th time, the neural network trained 51st time is not reserved and training continues. If the probability of the neural network trained 51st time in the validation data is higher than the probability of the neural network trained 25th time, the neural network trained 51st time is reserved, that is, the neural network trained 25th time is deleted and only the neural network trained 51st time is left, and training continues until a neural network that meets the specified requirements is trained. The test data is used to test the final trained neural network.

Based on the traffic predicted based on each training data and the actual traffic at the time, the neural network training is completed by adjusting the weights of the neural network based on a loss function. The loss function may be any suitable loss function, such as a mean square error (MSE) loss function, a cross entropy loss function (e.g., softmax), etc.

Finally, after the above training is completed, each weight of the neural network is fixed. That is, after the above process, the neural network is trained and can predict traffic information for future time periods.

As an example, the predicted future time slot traffic information may be one or more of the following: traffic for the future time slot, time marker for the future time slot, and sector direction for the future time slot.

The network side device provided by the present disclosure has been described in detail above. Note that the network side device provided by the present disclosure uses a neural network trained through a large amount of data, so that the method provided by the present disclosure can accurately predict traffic information for future time periods. Compared with the network side device described with reference to FIG. 15 above, the network side device described with reference to FIG. 16 is more convenient and faster.

Furthermore, only by further determining traffic information for a future time period predicted by the network side device described with reference to FIG. 15 above can the power saving level to be used by the network side device for that future time period be determined, and further a corresponding power saving operation can be performed by the network side device. In some cases, the manner of determining the power saving level may be inappropriate, for example, not very accurate in terms of the vicinity of the boundary value of the power saving level. Based on this, the present disclosure provides a further improvement of the network side device described with reference to FIG. 16.

According to an embodiment of the present disclosure, the acquisition unit 1610 is further configured to acquire power saving level information for the network side device corresponding to the traffic in the past time period, and the processing unit 1620 may further predict traffic information for a future time period using the neural network model based on information regarding the traffic in the past time period, the time sign information, the sector direction indication information, and the power saving level information.

As an example, the energy saving level information may be predetermined energy saving (ES) level information, for example, as shown in Table 4 above, where n and N are positive integers.

In one embodiment of the present disclosure, the power saving level information can also be used as input data to train the neural network. When the trained neural network is actually used, the traffic information for the predicted future time period includes the power saving level information corresponding to the future time period, thereby causing the network side device to perform a corresponding power saving operation.

As an example, power saving level information corresponding to each traffic can be added to the data used in training the neural network. As shown in FIG. 7A, traffic _T0 has an ES level of _ES0 in Table 4. Traffic _T1 has an ES level of _ES1 corresponding to the ES level in Table 4. Traffic _Tn has an ES level of _ESn corresponding to the ES level in Table 4. The _ES0 to _ESn are also used to train the neural network as data used in training the neural network. Thus, when the trained neural network is actually used, the traffic information for the predicted future time period may include power saving level information corresponding to the future time period.

As another example, in some cases, it may be necessary to know only whether the current power saving level has decreased or increased compared to the power saving level corresponding to a future time period. In a scene where only the change in the power saving level is required, as shown in FIG. 7B, the neural network can be trained with the power saving level change information used in training the neural network. Specifically, assume that the ES level corresponding to traffic T ₀ changes to 0. The ES level corresponding to traffic T ₁ is (ES ₁ -ES ₀ ). The ES level corresponding to traffic T _n is (ES _n -ES _n-1 ). The 0, (ES ₁ -ES ₀ ) ... (ES _n -ES _n-1 ) are also used to train the neural network as data used in training the neural network. Thus, when the trained neural network is actually used, the traffic information for the predicted future time period may include the power saving level change information corresponding to the future time period.

As another example, in some cases, it may be desired to know only the boundary value of the power saving level corresponding to the traffic in the future time period. In a scene where only the boundary value of the power saving level is desired to be known, as shown in FIG. 7C, the neural network can be trained using the boundary value of the power saving level corresponding to each traffic as data used in training the neural network. Specifically, the boundary value of the power saving level corresponding to the traffic T ₀ is [Boundary 0, Boundary 1] (e.g., [0, 100] in Table 4 above). The boundary value of the power saving level corresponding to the traffic T ₁ is [Boundary 1, Boundary 2] (e.g., [100, 200] in Table 4 above). The boundary value of the power saving level corresponding to the traffic T _n is [Boundary n-1, Boundary n] (e.g., [N-1, N] in Table 4 above). The boundary value corresponding to each traffic is also used as data used in training the neural network to train the neural network. Thus, when the trained neural network is actually used, the traffic information for the predicted future time period may include the boundary value of the power saving level corresponding to the traffic in the future time period.

In the above three embodiments, when training the neural network model, the neural network model is pre-trained by adjusting the parameters of the neural network model based on the model prediction loss (e.g., the above loss function) of the neural network model calculated using the power saving level information, and the traffic information of the future time period predicted by the trained neural network during actual use includes power saving level information corresponding to the future time period, so that the network side device performs the corresponding power saving operation.

In all three of the above examples, the neural network model is pre-trained by increasing the amount of data used to train the neural network, and adjusting the parameters of the neural network model based on the model prediction loss of the neural network model calculated using the power saving level information.

By improving the model prediction loss without increasing the data used to train the neural network, traffic information for future time periods predicted when the trained neural network is actually used includes power saving level information corresponding to the future time periods, thereby causing the network side device to perform corresponding power saving operations.

In yet another embodiment of the present disclosure, the model prediction loss is calculated by calculating the model prediction loss of the neural network model based on traffic range information corresponding to the power saving level information and traffic information predicted by the neural network model.

As an example, we will explain that the model prediction loss is an MSE loss function, as shown in Figures 8A and 8B.

Referring to FIG. 8A above, when training the neural network before it is improved, the MSELOSS of the neural network performs a comparison operation between the predicted traffic T'n (n=0, 1...m, n and m are all integers) and the actual traffic Tn (n=0, 1...m, n and m are all integers), and adjusts the weights of the neural network based on the loss function to complete the training of the neural network.

Referring to FIG. 8B, when training the neural network after being improved, first determine the corresponding power saving level of the actual traffic corresponding to each predicted traffic, for example, the power saving level corresponding to the actual traffic Tn (n=0, 1...m, n and m are all integers) is _ESn (n=0, 1...m, n and m are all integers).

Then, the traffic range corresponding to the power saving level is determined; for example, the traffic range corresponding to the determined ES2 is 100 to 200 (as shown in Table 4 above).

A representative value within the range is then selected as the traffic value corresponding to the power saving level. For example, a median value mid _n (n=0, 1...m, where n and m are both integers) within the range is selected as the traffic value corresponding to the power saving level (as shown in FIG. 8B). Alternatively, an extreme value within the range may be selected as the traffic value corresponding to the power saving level. For example, if the range is 0 to 100, the minimum value 0 may be selected as the traffic value corresponding to the power saving level. If the range is 100 to 200, the minimum value 100 may be selected as the traffic value corresponding to the power saving level. Furthermore, for example, if the range is 0 to 100, the maximum value 100 may be selected as the traffic value corresponding to the power saving level. If the range is 100 to 200, the maximum value 200 may be selected as the traffic value corresponding to the power saving level.

Finally, the neural network MSELOSS compares the predicted traffic (n=0, 1...m, where n and m are both integers) with the traffic value corresponding to the selected power saving level (e.g., mid _n (n=0, 1...m, where n and m are both integers)) and adjusts the weights of the neural network based on the loss function to complete the training of the neural network.

With the above improvements, the trained neural network model can include power saving level information corresponding to future time periods in the traffic information predicted for future time periods when actually used, thereby enabling the network side device to perform corresponding power saving operations.

Note that when training a neural network, a large amount of data is used for training, and the above-mentioned training neural network embodiment can well mine rules based on the data, but cannot combine human experiential knowledge, thereby further improving the trained neural network to obtain more favorable results. For the above case, the present disclosure provides a further improvement of the method described with reference to FIG. 5.

Specifically, when training the neural network, the neural network can be trained based on the information on traffic in the historical time period and the information on traffic in the adjacent time period used in the method described with reference to FIG. 1 above, and at least one of the first coefficients (e.g., the above-mentioned α, β, γ, δ), and a small amount of traffic information in the past time period used in the manner described with reference to FIG. 5 above, predicted time marker information, and predicted sector direction information, for example, by employing a neural network and training them in combination.

When training the neural network, the small amount of past traffic information, the predicted time sign information, and the predicted sector direction information are input to the first neural network for training, and the weight for training the first neural network is α _m (m=0, 1, 2...n, where m and n are both integers).

At least one of the information regarding traffic in the historical time period and the information regarding traffic in the adjacent time period used in the method described with reference to Figure 1 above, and the first coefficients (e.g., the above-mentioned α, β, γ, δ) is given a predetermined weight, for example β _m (m = 0, 1, 2...n, where m and n are both integers).

The two weights α _m and β _m are input to a second neural network for training to obtain weights α _m and β _m (denoted as p). Based on the weights α _m , β _m and p, a fusion weight of the whole neural network including the first neural network and the second neural network can be obtained. The second neural network can be any suitable conventional neural network, such as a Convoluted Neural Network (CNN), a Recurrent Neural Network (RNN), etc.

Finally, the loss function is minimized by adjusting the fusion weights of the entire neural network including the first neural network and the second neural network, thereby completing the training of the neural network.

According to an embodiment of the present disclosure, fusion weights can be obtained according to different needs.

For example, as shown in Fig. 9A, the weights α _m and β _m are input into the second neural network, and then the weight p is obtained. Then, the fusion weight is obtained based on α _m , β _m , and p.

As another example, as shown in Fig. 9B, after inputting the weights _αm and _βm into the second neural network, we obtain a set of corresponding weights _pn (m = 0, 1, 2...n, where m and n are both integers), and then obtain the fusion weights based on _pn .

As another example, as shown in FIG. 9C, after inputting the weights α _m and β _m into the second neural network, the fusion weight w _n (m=0, 1, 2...n, where m and n are both integers) is directly obtained.

According to an embodiment of the present disclosure, the information regarding traffic in the past time periods may include multiple sets of traffic data for the past time periods, and the processing unit 1620 may further normalize the traffic data for each past time period in the multiple sets by dividing the traffic data for each past time period in the multiple sets by the maximum value in the multiple sets, and predict traffic information for a future time period using the neural network model based on the normalized traffic data. Alternatively, the processing unit 1620 may further normalize each traffic data in each set by dividing each traffic data in the multiple sets by the maximum value in the multiple sets, and predict traffic information for a future time period using the neural network model based on the normalized traffic data.

In the above description, when training a neural network, the neural network performs normalization processing on the data in batches, and each batch of data is composed of several random data. For example, the data in each batch may be traffic for n (n is an integer) time periods as shown in FIG. 6A. In other words, the information on traffic for past time periods in the training of each batch may include multiple sets of traffic data for past time periods.

Referring to Fig. 10A, the batch of data input to the neural network for training includes n sets (i.e., 1-n, n is an integer). For the i-th set, the data of the set may include traffic in the past 72 time slots, i.e., _x0,0 to _x0,71 . Before improving the aspect of the normalization process, the neural network divides all the traffic data in each time slot in the batch of data by the sum of all the traffic data in the batch of data to achieve the purpose of the normalization process, as shown in Fig. 10A.

The value of each traffic data is relatively large (e.g., reaching hundreds of thousands), so the sum of the traffic data of the batch is large. The traffic data in each time period in the batch data is divided by the sum of all the traffic data in the batch data, and the resulting data is very small. Therefore, although the error generated in the subsequent training process is small, the predicted traffic error due to the small error caused by multiplying the sum is large.

In contrast, the improved method of the present disclosure achieves the purpose of normalization by dividing all traffic data in each time period in the batch of data by the maximum value (max) among all traffic data in the batch of data, as shown in FIG. 10A. Because the difference between the traffic data value in each time period in the batch of data and the maximum value is small, the trained neural network is faster.

In addition, the information on traffic in the past time period of each batch may include multiple sets of traffic data in the past time period. For example, the first batch of data is composed of three sets of data, 1, 3, and 5. The second batch of data is composed of three sets of data, 1, 2, and 4. When performing normalization processing based on the above improvement method, it is necessary to divide the first set of data in the first batch of data by the maximum value of the data in that batch, max 1, and to divide the first set of data in the second batch of data by the maximum value of the data in that batch, max 2. Since max 1 and max 2 are always different, when a neural network is trained, a similar set of data is divided by different values when normalized, which may reduce the prediction accuracy of the trained neural network.

For the above case, another improvement method of the present disclosure is to divide all the traffic data in each time period in each set of data in the batch of data by the maximum value (max) of all the traffic data in the set of data, thereby achieving the purpose of normalization processing, so that the trained neural network is faster and more accurate, as shown in FIG. 10B.

The different improvements to the network side device described with reference to the drawings described above may be implemented simultaneously or selectively, so that the improved network side device includes at least one of the improvements described above.

The network side device in the wireless communication network provided by the present disclosure has been described above with reference to the accompanying drawings. The network side device provided by the present disclosure uses a neural network to enable the network side device provided by the present disclosure to quickly and accurately predict traffic information for a future time period. The network side device provided by the present disclosure further introduces power saving level information when training the neural network, so that the network side device provided by the present disclosure can accurately predict power saving level information corresponding to traffic for a future time period, and thus the network side device can more conveniently perform a corresponding power saving operation. The network side device provided by the present disclosure further uses a fusion weight to further improve the accuracy of the predicted traffic information for a future time period. The network side device provided by the present disclosure further improves the normalization aspect to enable the network side device provided by the present disclosure to more quickly and accurately predict traffic information for a future time period. Furthermore, the network side device only provides some resources that can process the above traffic and turns off other resources, thereby saving power consumption in the network side device.

The network side device and the corresponding method described with reference to the above drawings can accurately predict traffic in a future time period, so that the network side device only provides some resources that can handle the traffic and turns off other resources, thereby achieving power consumption savings in the network side device for the entire time period.

Referring to FIG. 17, the network side device 1700 may include a processing unit 1710 and a control unit 1720. In this example, the network side device 1700 includes a processing unit 1710 and a control unit 1720. Note that the network side device 1700 may further include other components, but these components are not relevant to the contents of the embodiments of the present disclosure, and therefore will not be illustrated or described here.

As an example, the network side device 1700 may be a base station. Or, the network side device 1700 may be the above-mentioned OAM module or SMO module that interacts with the base station and with the outside of the base station. When the network side device is the above-mentioned OAM or SMO module, the OAM or SMO module needs to transmit information on the power saving level obtained by executing the method provided by the present disclosure, traffic information for the predicted future time period, and resource information corresponding to the power saving level, for example, to the base station, and the base station performs the corresponding power saving operation.

The processing unit 1710 is configured to determine a power saving level to be used by the network side device in a future time period based on traffic in the future time period and a predetermined power saving level table, where the power saving level table may include a plurality of power saving levels corresponding to the wireless transmission capabilities of the network side device and the resources used when the network side device transmits wirelessly.

As an example, the network side device can acquire traffic for the above future time period in real time or at a preset time.

In the embodiment of the present disclosure, the resources used by the network side device for wireless transmission may be at least one of time domain resources, frequency domain resources, spatial domain resources, and power domain resources, or a combination thereof. The time domain resources may include resources such as subframes, slots, microslots, and symbols. The frequency domain resources may include resources such as component carriers (CCs), a portion of bandwidth (BWP), subbands, reference signal groups, and subcarrier groups. The spatial domain resources may include resources such as antenna elements, transceiver units (TxRUs), and virtual antenna ports. The power domain resource may include resources required to adjust at least one of the maximum transmit power, power spectral density (PSD), etc. of one or more channels of transmitted data (Data), control (control), reference signal (RS), synchronization signal block (SSB), etc.

As an example, the predetermined power saving level table may be determined in advance based on the wireless transmission capability of the network side device and the resources used when the network side device performs wireless transmission. For example, an energy saving (ES) level table such as that shown in Table 5 above may be determined in advance.

In the above Table 5, the ES level is divided into five levels, namely ES1-ES4. The ES level may be divided into any multiple levels according to actual needs.

In Table 5 above, the power saving levels ES1-ES4 are in the order of ES1, ES2, ES3, and ES4 from highest to lowest. Power saving level ES1 may indicate the lowest energy consumption in the network side device. Power saving level ES4 may indicate the highest energy consumption in the network side device. Power saving levels ES2 and ES3 indicate that the power consumption in the network side device is between ES1 and ES4. Other manners may be designed to indicate high or low energy consumption in the network side device.

The resources in Table 5 above can indicate resources that the network side device uses during wireless transmission. For example, at least one of time domain resources, frequency domain resources, spatial domain resources, and power domain resources, or a combination thereof.

The radio transmission capability in Table 5 above can indicate the radio transmission capability of the network side device when using the corresponding resource. As an example, the transmission capability of the network side device may be the maximum value, average value, weighted average value, etc. of the transmission data amount of the network side device within a certain time period in the past. For example, the transmission capability of the network side device may be the maximum value or weighted average value of the data amount within a unit time of the component carrier within a past time period. Also, for example, the transmission capability of the network side device may be the product of the average value of all UE transmission rates on a specific component carrier within a certain time period and the bandwidth of the specific component carrier. Also, for example, the transmission capability of the network side device may be the maximum rate of the component carrier for a specific modulation and coding scheme (MCS) or (Rank) number. Also, for example, the transmission capability of the network side device may be the ratio between the actual transmission data amount and the resource usage rate.

When the resources include multiple component carriers available to the network side device, according to one embodiment of the present disclosure, each component carrier can have a priority corresponding to the component carrier, where in the power saving level table, the higher the energy consumption, the more low priority component carriers the power saving level turns on.

The resources may include four component carriers available to the network side device. As shown in FIG. 12A, the four component carriers are denoted as CC1, CC2, CC3, and CC4. CC1 has a center frequency of 800 M, a beamwidth of 55 M, and a beam coverage radius of 902 meters (m). CC2 has a center frequency of 1.5 G, a beamwidth of 1.5 G, and a beam coverage radius of 654 m. CC3 has a center frequency of 1.7 G, a beamwidth of 45 M, and a beam coverage radius of 613 m. CC4 has a center frequency of 2.0 G, a beamwidth of 45 M, and a beam coverage radius of 564 m. The lower the center frequency of a component carrier, the lower the priority. That is, the priority of the four component carriers is CC4, CC3, CC2, and CC1, in order of decreasing priority. In addition, the priority of a component carrier can be determined based on factors such as the coverage range of the component carrier, the traffic that is permitted to be transmitted, and the power consumed in traffic transmission.

In this case, if the contents of the resource column in Table 5 above are specifically the above component carriers CC1-CC4, the content shown in Table 6 above is obtained.

In the power saving level table 6 above, the power saving level ES4 with the highest energy consumption turns on all four component carriers. The power saving level ES1 with the lowest energy consumption activates only component carrier CC1. The number of component carriers turned on by ES2 and ES3, which have energy consumption between ES4 and ES1, is between ES4 and ES1. In other words, a power saving level with high energy consumption activates more low priority component carriers than a power saving level with low energy consumption.

In addition, the wireless communication network may include one or more sectors, and the one or more sectors are also resources available to the network side device. The multiple sectors may be uniform, in which case the size of the coverage area of each sector is the same. The multiple sectors may be non-uniform, in which case the size of the coverage area of each sector is different. For the situation of uniform sectors, the determination of the wireless transmission capabilities of the base station in Tables 5 and 6 above was described above with reference to Figures 12A and 12B. For the situation of non-uniform sectors, the determination of the wireless transmission capabilities of the base station in Tables 5 and 6 above was previously described with reference to Figures 13A to 13C. The description will be omitted here.

In yet another embodiment of the present disclosure, the processing unit 1720 may adjust the traffic for the future time period according to one or more of a prediction error, a target power saving amount, or a target data throughput, and determine a power saving level to be used for the network side device in the future time period based on the adjusted traffic for the future time period and a predetermined power saving level table.

In one embodiment of the present disclosure, there may be a certain error between the traffic in the future time slot and the actual traffic. Although the error is small, the traffic in the future time slot can be adjusted to more accurately determine the power saving level of the network side device in the future time slot. For example, it is determined based on empirical or statistical values that the relative error between the traffic in the future time slot and the actual traffic is x%, in which case the traffic in the future time slot can be adjusted according to the above formula (4).

In another embodiment of the present disclosure, the higher the power saving level of the network side device in the determined future time period, the fewer resources the network side device uses, and therefore the smaller the data throughput, so that the scaling factor can be determined based on at least one of the target power saving amount and the target data throughput when adjusting the traffic in the future time period. For example, statistical data indicates that when the scale factor x=1.1, the decrease in data throughput in the network side device can be controlled within 2%, thereby causing the service impact on the UE to be within a controllable range when adjusting the power saving level. In this case, the traffic in the future time period can be adjusted according to the above formula (5).

In yet another embodiment of the present disclosure, the scaling coefficient can be further obtained based on a target scene. For example, the target scene may include scenes such as urban macro base station (Umi, Urban Microcell), urban macro base station (Uma, Urban Macrocell), rural (Rma, Rural Macrocell), indoor hotspot (InH, Indoor Hotxt), etc. A scaling coefficient mode or a scaling coefficient table may be designed in advance based on the different scenes and experiences, and then a specific scaling coefficient may be selected based on the target data throughput and target scene requirements that are achieved according to actual commercial needs. By considering one or more of the prediction error, the target power saving amount, the target data throughput, or the target scene, the minimum power saving level required for the network side device when transmitting all traffic may be selected as much as possible based on the traffic of the future time period.

As shown in FIG. 14, as the scaling factor increases, traffic in the future time period is constantly amplified, which causes a lower power saving level to be selected. That is, as the scaling factor increases, the network side device uses more resources to transmit data, which continuously reduces the data throughput reduction rate in the network side device, thereby increasing the power consumption in the video unit (AAU) in the network side device from time to time. For example, when the scale factor is 1.1, the data throughput in the network side device decreases by 1.7%, and the power consumption in the video unit (AAU) in the network side device increases to 53.9%.

Returning to FIG. 17, the control unit 1720 may be configured to perform power saving operations based on the determined power saving level.

As an example, the power saving operation performed by the network side device may include closing resources that do not need to continue to be used by waiting for shutdown. For example, if the power saving level of the network side device in the future time period is determined to be ES3 in Table 6 above, the network side device turns off the previously activated CC4 in the future time period. When specifically turning off CC4, no new UE accesses the CC4 to be turned off, and after transmitting all the data transmitted by CC4, for example, when the amount of data transmitted by CC4 is small (there is a certain waiting delay in such a situation), CC4 is turned off.

As another example, the power saving operation performed by the network side device may include turning off resources that do not need to continue to be used by an uninstallation method. For example, if it is determined that the power saving level of the network side device in a future time period is ES3 in Table 6 above, the network side device turns off the previously activated CC4 in the future time period. When specifically turning off CC4, no new UEs access the CC4 to be turned off, and data transmitted by CC4 can be uninstalled and transmitted to any one or more of CC1-CC3, and finally CC4 is turned off. The above situation applies to a situation where there is a lot of traffic transmitted by the CC, and this uninstallation method results in a certain signaling overhead and signaling interaction delay.

As another example, the power saving operation of the network side device may include flexibly operating one or two of the standby off method and the offload method based on the traffic transmitted by the CC waiting to be turned off and the number of UEs.

As another example, the power saving action taken by the network side device may include selecting the component carriers that need to be closed by combining one or more of the cell characteristics, the traffic and type of the connected UE, and the priority of the component carrier tasks.

Depending on the cell characteristics, CCs with small coverage (e.g., CCs with high frequency and narrow width, e.g., CC4 described with reference to FIG. 12A above or CC5 described with reference to FIG. 13A above) can be preferentially turned off, CCs with small amounts of transmitted data can be preferentially turned off, CCs with high transmission power can be preferentially turned off, or CCs with low unit power transmission capacity can be preferentially turned off. Depending on the actual application, CCs with large coverage, large amounts of transmitted data, low transmission power, or high unit power transmission capacity can be preferentially closed.

For the traffic and type of connected UEs, the CC with the least traffic transmitted by it can be preferentially turned off, the CC with the least number of UEs that set the CC as a primary cell (Cell, Pscell) or secondary cell (Cell, Scell) (in such a case, the signaling overhead caused is small), the CC with the least number of UEs that set the CC as a Pscell (in such a case, the signaling overhead caused is small), or the CC with the least number of low terminated UEs can be preferentially turned off. Alternatively, depending on the actual application, the CC with the largest amount of traffic transmitted by it and the largest number of UEs or the largest number of low terminated UEs can be preferentially turned off.

In relation to the priority of the component carrier task, a CC with a low priority, for example CC1 described with reference to FIG. 12A or FIG. 13A above, can be preferentially turned off. Or, a CC with a high priority can be preferentially turned off.

The network side device provided by the present disclosure has been described in detail above with reference to the accompanying drawings. The network side device provided by the present disclosure can accurately determine the power saving level of the network side device in a future time period based on the traffic of the future time period and a predetermined power saving level table, so that the network side device can accurately select the power saving level and accurately reserve only the minimum resources required to transmit traffic when operating based on the determined power saving level, thereby achieving an optimal power saving effect.

<Hardware Configuration>
Moreover, the block diagrams used in the description of the above embodiments show functional blocks. These functional blocks (components) are realized by any combination of at least one of hardware and software. Furthermore, the method of realizing each functional block is not particularly limited. That is, each functional block may be realized using one device that is physically or logically coupled, or may be realized using two or more devices that are physically or logically separated and directly or indirectly (for example, using wires, wirelessly, etc.) connected to these multiple devices. The functional blocks may be realized by combining the one device or the multiple devices with software.

In terms of functions, there are judgment, decision, determination, calculation, computation, processing, derivation, investigation, search, confirmation, reception, transmission, output, access, resolution, selection, selection, establishment, comparison, assumption, expectation, estimation, broadcasting, notifying, communicating, forwarding, configuring, reconfiguring, allocating, mapping, and assigning, but are not limited to these. For example, a functional block (component) that performs a function is called a transmission unit or a transmitter. As mentioned above, there are no particular limitations on the method of realization for each.

For example, a base station, a user terminal, etc. in an embodiment of the present disclosure may function as a computer that performs processing of the wireless communication method of the present disclosure. FIG. 18 is a diagram showing an example of a hardware configuration of a network side device in an embodiment of the present disclosure. The above-mentioned network side device (for example, the network side device 15, the network side device 16, or the network side device 17) may be physically configured as a computer device including a processor 1801, a memory 1802, a storage 1803, a communication device 1804, an input device 1805, an output device 1806, a bus 1807, etc.

In addition, in the following description, the term "device" can be replaced with circuit, device, section, etc. The hardware configuration of the network side device may be configured to include one or more of the devices shown in the figure, or may be configured to not include some of the devices.

For each function in the network side device, specific software (programs) are loaded into hardware such as processor 1801 and memory 1802, and the functions are realized by the processor 1801 performing calculations and controlling at least one of communication by communication device 1804 and reading and writing data in memory 1802 and storage 1803.

The processor 1801, for example, runs an operating system to control the entire computer. The processor 1801 may be configured as a central processing unit (CPU) including an interface with peripheral devices, a control device, an arithmetic unit, a register, etc. For example, the control unit, processing unit, etc. of the network side device may be realized by the processor 1801.

The processor 1801 also reads out programs (program codes), software modules, data, etc. from at least one of the storage 1803 and the communication device 1804 into the memory 1802, and executes various processes accordingly. The programs used are those that cause a computer to execute at least a part of the operations described in the above-mentioned embodiments. For example, the processing unit or control unit of the network side device may be realized by a control program stored in the memory 1802 and operated by the processor 1801, and may be similarly realized for other functional blocks. Although the above various processes have been described as being executed by one processor 1801, they may be executed simultaneously or sequentially by two or more processors 1801. The processor 1801 may be realized by one or more chips. The programs may also be transmitted from the network via telecommunications lines.

Memory 1802 may be a computer-readable recording medium, and may be, for example, at least one of a ROM (Read Only Memory), an EPROM (Erasable Programmable ROM), an EEPROM (Electrically Erasable Programmable ROM), a RAM (Random Access Memory), etc. Memory 1802 may also be called a register, a cache, a main memory (primary storage device), etc. Memory 1802 may store a program (program code), software module, etc. that is executable to implement a wireless communication method according to an embodiment of the present disclosure.

Storage 1803 may be a computer-readable recording medium, and may be, for example, at least one of an optical disk such as a CD-ROM (Compact Disc ROM), a hard disk drive, a flexible disk, a magneto-optical disk (for example, a compressed disk, a digital multifunction disk, a Blu-ray (registered trademark) disk), a smart card, a flash memory (for example, a card, a stick, a key drive, a floppy (registered trademark) disk, a magnetic stripe (PE), etc. Storage 1803 may be called an auxiliary storage device. The above recording medium may be, for example, a database, a server, or other suitable medium including at least one of memory 1802 and storage 1803.

The communication device 1804 is hardware (transmitting/receiving device) for communicating between computers via at least one of a wired network and a wireless network, and is also called, for example, a network device, a network controller, a network card, a communication module, etc. The communication device 1804 may be configured to include a high-frequency switch, a duplexer, a filter, a frequency synthesizer, etc., to realize, for example, at least one of Frequency Division Duplex (FDD) and Time Division Duplex (TDD).

The input device 1805 is an input device (e.g., a keyboard, a mouse, a microphone, a switch, a button, a sensor, etc.) that accepts input from the outside. The output device is an output device (e.g., a display, a speaker, an LED lamp, etc.) that outputs to the outside. The input device 1805 and the output device may also be integrated into one structure (e.g., a touch panel).

In addition, each device, such as the processor 1801 and memory 1802, is connected by a bus for communicating information. The bus may be configured using a single bus, or different buses may be used between each device.

The network side device may also be configured to include hardware such as a microprocessor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a programmable logic device (PLD), or a field programmable gate array (FPGA), and the hardware implements some or all of the functional blocks. For example, the processor 1801 may be implemented using at least one of these pieces of hardware.

<Modification>
In addition, in the present disclosure, the notification of information is not limited to the aspects/embodiments described in the present disclosure, and may be performed using other methods. For example, the notification of information may be performed by physical layer signaling (e.g., Downlink Control Information (DCI)), Uplink Control Information (UCI)), higher layer signaling (e.g., Radio Resource Control (RRC) signaling, Medium Access Control (MAC) signaling, broadcast information (Master Information Block (MIB)), System Information Block (SIB)), other signals, or combinations thereof. In addition, the RRC signaling may be referred to as an RRC message, such as an RRC Connection Setup message, an RRC Connection Reconfiguration message, or the like.

Each aspect/embodiment described in this disclosure is applicable to LTE (Long Term Evolution), LTE-A (LTE-Advanced), SUPER 3G, IMT-Advanced, 4G (fourth generation mobile communication system), 5G (fifth generation mobile communication system), sixth generation mobile communication system (6th generation mouse ication system (6G)), xth generation mobile communication system (xth generation mouse ication system (xG)) (xG (x is, for example, an integer or a decimal number)), FRA (F radio Access), NR (new Radio), New radio access (NX), next-generation wireless access (FX), W-CDMA (registered trademark), GSM (registered trademark), CDMA 2000, UMB (Ultra Mobile Broadband), IEEE 802.11 (Wi-Fi (registered trademark)), IEEE 802.16 (Wi MAX (registered trademark)), IEEE 802.20, UWB (Ultra-Wide Band), Bluetooth (registered trademark), and other suitable systems, and may be applied to at least one of next-generation systems extended, modified, created, or defined based on these. In addition, multiple systems may be applied in combination (for example, a combination of at least one of LTE and LTE-A and 5G, etc.).

The processing procedures, sequences, flow charts, etc. of each aspect/embodiment described in this disclosure may be reordered unless inconsistent. For example, the methods described in this disclosure present elements of various steps in an illustrated order and are not limited to the particular order presented.

In the present disclosure, certain operations performed by a base station may also be performed by its upper node depending on the circumstances. In a network consisting of one or more network nodes having a base station, various operations performed for communication with a terminal are obviously performed by at least one of the base station and other network nodes other than the base station (e.g., MME or S-GW, etc., but are not limited to these). Although the above example shows a case where there is one other network node other than the base station, it may also be a combination of multiple other network nodes (e.g., MME and S-GW).

Information, etc. (see "Information, Signals" section) can be output from a higher layer (or a lower layer) to a lower layer (or a higher layer). It may be input and output via multiple network nodes.

The information, etc. that is input and output may be stored in a specific location (e.g., memory) or may be managed using a management table. The information, etc. that is input and output may be overwritten, updated, or added. The information, etc. that is output may be deleted. The information, etc. that is input may be transmitted to another device.

The determination may be made based on a value represented by one bit (0 or 1), a boolean value (true or false), or a comparison of numerical values (e.g., with a specific value).

Each aspect/embodiment described in this disclosure may be used alone or in combination, or may be switched between depending on the implementation. In addition, notification of specific information (e.g., notification of "X") is not limited to being done explicitly, but may be done in an implicit manner (e.g., not notifying the specific information).

Although the present disclosure has been described in detail above, it is clear to those skilled in the art that the present disclosure is not limited to the embodiments described herein. The present disclosure can be implemented in modified and altered forms without departing from the spirit and scope of the present invention as defined by the claims. Therefore, the description of the present disclosure is intended as an illustrative example and does not have any limiting meaning on the present disclosure.

In this disclosure, software may be referred to as software, firmware, middleware, microcode, hardware description language, or by other names, and shall be construed broadly to mean instructions, calls, code, code segments, program code, programs, subprograms, software modules, applications, software applications, software packages, routines, subroutines, objects, executable files, threads of execution, processes, functions, etc.

Software, instructions, information, etc. may also be transmitted and received via a transmission medium. For example, if the software is transmitted from a website, server, or other remote source using at least one of wired technologies (such as coaxial cable, optical cable, twisted pair, Digital Subscriber Line (DSL)), and/or wireless technologies (such as infrared, microwave, etc.), then at least one of these wired and wireless technologies is included within the definition of a transmission medium.

The information, signals, etc. described in this disclosure may be represented using any of a variety of different technologies. For example, data, commands, information, signals, bits, symbols, chips, etc. that may be referred to throughout the above description may be represented as voltages, currents, electromagnetic waves, magnetic fields or magnetic particles, light fields or photons, or any combination thereof.

In addition, the terms described in this disclosure and the terms necessary for understanding this disclosure may be replaced with terms having the same or similar meanings. For example, at least one of the channel and the symbol may be a signal (signaling). Also, the signal may be a message. Also, the component carrier (CC) may be called a carrier frequency, a cell, a frequency carrier, etc.

As used in this disclosure, the terms "system" and "network" are used interchangeably.

In addition, the information, parameters, etc. described in this disclosure may be expressed using absolute values, may be expressed as relative values to a specific value, or may be expressed using other corresponding information. For example, radio resources may be indicated by an index.

The names used for the above parameters are not intended to be limiting in any respect. Furthermore, the formulas etc. using these parameters may differ from those expressly disclosed in this disclosure. The various channels (e.g., PUCCH, PDCCH, etc.) and information elements may be identified by any suitable names, and thus the various names assigned to these various channels and information elements are not intended to be limiting in any respect.

In this disclosure, terms such as "base station (BS)", "radio base station", "fixed station", "NodeB", "eNodeB (eNB)", "gNodeB (gNB)", "access point", "transmission point", "reception point", "transmission/reception point", "cell", "sector", "cell group", "carrier", and "component carrier" are used interchangeably. A base station may also be referred to by terms such as macrocell, small cell, femtocell, and picocell.

A base station can accommodate one or more (e.g., three) cells. When a base station accommodates multiple cells, the entire coverage area of the base station can be divided into smaller areas, and each smaller area can also provide communication services by a base station subsystem (e.g., a small base station for indoor use (Remote Radio Head (RRH)). The term "cell" or "sector" refers to a part or the entire coverage area of at least one of the base station and base station subsystem that provides communication services within that coverage.

In this disclosure, the terms "Mobile Station (MS)," "user terminal," "User Equipment (UE)," "terminal," etc. may be used interchangeably.

A mobile station may also be referred to by those skilled in the art as a subscriber station, mobile part, subscriber part, radio part, remote part, mobile equipment, radio equipment, radio communication equipment, remote device, mobile subscriber station, access terminal, mobile terminal, radio terminal, remote terminal, handset, user agent, mobile client, client, or some other suitable terminology.

At least one of the base station and the mobile station may be called a transmitting device, a receiving device, a communication device, etc. At least one of the base station and the mobile station may be a device mounted on a moving body, a moving body, etc. The moving body is a movable object, and the moving speed is arbitrary. Of course, it also includes a case where the moving body is stopped. The moving body includes, for example, a vehicle, a transport vehicle, an automobile, a motorcycle, a bicycle, a networking automobile (connected car), a forklift, a bulldozer, a wheel loader, a dump truck, a forklift, a train, a bus, a rear trailer, a rickshaw, a ship (ship and other water transport tool (and other wraft)), an aircraft, a rocket, an artificial satellite, an unmanned aerial vehicle (drone) (registered trademark), a multicopter, a four-rotor (vertical take-off and landing) helicopter (quadcopter), a balloon, and objects mounted thereon, or is not limited thereto. The moving object may be an autonomous moving object based on an operation command. It may be a moving object (e.g., a car, an airplane, etc.), an unmanned moving object (e.g., an unmanned aerial vehicle, an autonomous vehicle, etc.), or a robot (manned or unmanned). At least one of the base station and the mobile station may further include a device that does not necessarily move during communication operations. For example, at least one of the base station and the mobile station may be an IoT (Internet of Through) device such as a sensor.

In addition, the base station in the present disclosure may be decoded by a user terminal. For example, each aspect/embodiment of the present disclosure may be applied to a structure in which communication between a base station and a user terminal is replaced with communication between multiple user terminals (e.g., called D 2D (Device-to-Device), V 2X (Vehicle-to-Everything), etc.). In this case, the user terminal may be configured to have the functions of the base station described above. Furthermore, terms such as "uplink" and "downlink" may be replaced with terms corresponding to terminal-to-terminal communication (e.g., "side"). For example, the uplink channel, downlink channel, etc. may be replaced with a side channel.

Similarly, the user terminal in this disclosure may be decrypted by the base station. In this case, the base station may be configured to have the functions of the user terminal described above.

The term "determining" as used in this disclosure may include a variety of actions. For example, "determining" may be considered to include judging, calculating, computing, processing, deriving, investigating, looking up, swiping, inquiry (e.g., searching in a table, database, or other data structure), ascertaining, and the like. Also, "determining" may be considered to include receiving (e.g., received information), transmitting (e.g., transmitting information), input, output, accessing (e.g., data accessed from memory), "judging," "deciding," and the like. In addition, "judgment" and "decision" can be considered to be resolving, selecting, choosing, establishing, comparing, etc. In other words, "judgment" and "decision" can be considered to be a "judgment" or "decision" for some actions. In addition, "judgment" may be replaced with "assuming," "expecting," "considering," etc.

In this disclosure, the terms "connected" and "coupled" or any variation thereof refer to any connection or coupling, direct or indirect, between two or more elements, and may include the presence of one or more intermediate elements between the two elements that are "connected" or "coupled" to each other. The coupling or connection between the elements may be physical, logical, or a combination thereof. For example, "connected" may be replaced with "access." As used in this disclosure, it may be considered to use at least one of one or more wires, cables, and printed electrical connections, and may use wavelengths having the radio frequency range, microwave range, and light (both visible and non-visible) range, as some non-limiting and non-limiting examples.

As used in this disclosure, the phrase "based on" does not mean "based only on," unless expressly stated otherwise. In other words, the phrase "based on" means both "based on" and "based at least on."

Any reference to elements using designations such as "first," "second," etc., as used in this disclosure is not intended to inclusively limit the amount or order of those elements. These designations may be used in this disclosure as a convenient way to distinguish between two or more elements. Thus, reference to a first and second element does not imply that only two elements may be employed or that the first element must take precedence over the second element in some form.

In this disclosure, the "parts" in the configuration of each of the above-mentioned devices may be replaced with "circuits," "equipment," etc.

When used in this disclosure, the terms "include," "including," and variations thereof are meant to be inclusive, similar to the term "comprising." Additionally, the term "or" as used in this disclosure is not meant to be an exclusive or.

Although the present disclosure has been described in detail above, it is clear to those skilled in the art that the present disclosure is not limited to the embodiments described in the present specification. The present disclosure can be implemented as modified and altered forms without departing from the spirit and scope of the present disclosure as defined by the claims. Therefore, the description in the present specification is intended to be illustrative and does not have any limiting meaning on the present disclosure.

Claims (6)

A network side device in a wireless communication network,
an acquisition unit configured to acquire information on traffic in a past time period, time indicator information indicating whether the traffic in the past time period is traffic in a first time period or traffic in a second time period, and sector direction indication information indicating a sector direction of a cell corresponding to the network side device;
and a processing unit configured to predict traffic information for a future time period using a neural network model based on information about traffic for the past time period, the time sign information, and the sector direction indication information indicating a sector direction transmitting the traffic . The acquisition unit is further configured to acquire power saving level information for the network side device corresponding to the traffic in the past time period,
The network side device of claim 1, wherein the processing unit is further configured to predict traffic information for a future time period using the neural network model based on the information on traffic for the past time period, the time sign information, the sector direction indication information, and the power saving level information. The network side device according to claim 2, wherein the neural network model is pre-trained by adjusting parameters of the neural network model based on a model prediction loss of the neural network model calculated using the power saving level information. The model prediction loss is
The network side device according to claim 3 , further comprising: calculating a model prediction loss of the neural network model based on traffic range information corresponding to the power saving level information and traffic information predicted by the neural network model. The information about traffic over a historical time period includes a plurality of sets of traffic data over a historical time period;
2. The network side device according to claim 1, wherein the processing unit is configured to: normalize traffic data for each past time period in a plurality of sets by dividing the traffic data for each past time period by a maximum value in the plurality of sets, and predict traffic information for a future time period using the neural network model based on the normalized traffic data; or the processing unit is configured to normalize traffic data for each past time period in each set by dividing the traffic data for each past time period by a maximum value in the plurality of sets, and predict traffic information for a future time period using the neural network model based on the normalized traffic data. A wireless communication method for a network side device, comprising:
Obtaining information on traffic in a past time period, time indicator information indicating whether the traffic in the past time period is traffic in a first time period or traffic in a second time period, and sector direction indication information indicating a sector direction of a cell corresponding to the network side device;
and predicting traffic information for a future time period using a neural network model based on information about traffic for the past time period, the time sign information, and the sector direction indication information indicating a sector direction for transmitting traffic .