A method and system for dynamic adjustment of access point energy saving window

Abstract

The application discloses a kind of dynamic adjustment method and system of access point energy-saving window, it is related to network communication technical field.The application includes: real-time acquisition of the multidimensional service characteristic data of access point;Business mode library is constructed, and business mode library includes multiple business modes;At preset periodic time point, based on the multidimensional service characteristic data collected in current period, each business mode in business mode library is matched with mode characteristic condition;When the matching result satisfies the preset switching condition from current business mode to target business mode, business mode switching is triggered;After triggering business mode switching, the energy-saving time window under target business mode is dynamically generated by calling the energy-saving window calculation mode associated with target business mode, and target energy-saving time window is issued to access point execution.The application enables energy-saving strategy to closely follow the real-time evolution of business mode, significantly improves the sensitive response capability to network service dynamic change.

Description

Technical Field

[0001] This invention relates to the field of network communication technology, and in particular to a method and system for dynamically adjusting the energy-saving window of an access point. Background Technology

[0002] The rapid development of Wireless Local Area Network (WLAN) technology, especially the widespread adoption of Wi-Fi 6 / 6E / 7, has made high-density, high-performance enterprise-grade WLAN deployments a critical infrastructure supporting modern digital parks, smart factories, smart healthcare, and smart education scenarios. In these scenarios, to achieve seamless coverage and high-capacity concurrency, the deployment density of Access Points (APs) has increased significantly (e.g., room-level and desktop-level deployments). This has led to a sharp increase in the overall energy consumption of AP devices, becoming a significant component of enterprise operating costs (OPEX) and a focus of attention in achieving "dual carbon" goals. Therefore, while ensuring a good user network experience, intelligently controlling APs to enter energy-saving states during off-peak hours (e.g., shutting down radio frequency, reducing transmission power, or entering deep sleep mode) has become a key technological direction for reducing network operating costs and improving energy efficiency.

[0003] Current mainstream energy-saving solutions rely on the analysis of historical traffic patterns (mainly the number of connected terminals) to generate a predefined, fixed energy-saving time window at specific times (such as initial deployment or manual triggering). The typical implementation process is as follows: the system only collects historical data for longer periods, such as 7 days or 30 days, for statistical fitting during deployment initialization or manual triggering, generates the energy-saving window, and fixes the configuration. After the energy-saving window takes effect, the system no longer continuously collects real-time business characteristics or establishes a dynamic monitoring mechanism. Regardless of subsequent business changes, it continues to use the initial strategy, and can only be re-collected, recalculated, and updated through manual intervention.

[0004] However, these energy-saving solutions are based on static learning or fixed configuration of historical data over a long period of time, lacking the ability to continuously monitor and adaptively adjust to the dynamic evolution of business models, and thus unable to respond sensitively to short-term or sudden changes in business models. Summary of the Invention

[0005] Therefore, it is necessary to provide a method and system for dynamically adjusting the energy-saving window of the access point to address the above-mentioned technical problems.

[0006] This invention provides a method for dynamically adjusting the energy-saving window of an access point, comprising: Real-time acquisition of multi-dimensional business characteristic data of access points; Construct a business model library, which includes multiple business models, and each business model includes model feature conditions consisting of multiple specific business features and their threshold ranges. At preset periodic time points, the multi-dimensional business feature data collected in the current period is matched with the pattern feature conditions of each business pattern in the business pattern library; when the matching result meets the preset switching conditions for switching from the current business pattern to the target business pattern, the business pattern switching is triggered. After triggering the business mode switch, the energy-saving window calculation method associated with the target business mode is invoked to dynamically generate the target energy-saving time window under the target business mode, and the target energy-saving time window is sent to the access point for execution.

[0007] Optionally, a business model library may be constructed, specifically including: The system receives the historical date range selected by the user through the user interface and defines the business mode name for the historical date range. Extract multi-dimensional business feature data corresponding to historical date ranges; A classification model is used to perform machine learning on the various business feature data to obtain pattern feature conditions for identifying the business pattern. Store the business model and its corresponding model feature conditions in the business model library.

[0008] Optionally, when a user defines a new business model, the system receives the energy-saving window determination method independently specified by the user for the new business model; the energy-saving window determination method includes: manually configuring a fixed time period or automatically learning and recommending a time period based on historical data of the business model.

[0009] Optionally, the method of calculating the energy-saving window associated with the target business model is invoked to dynamically generate the target energy-saving time window under the target business model, which also includes: When a target business model is configured to automatically learn and recommend time periods based on historical data of that business model, the tidal prediction algorithm is invoked to learn from all historical date data belonging to that target business model in the business model library, and obtain energy-saving time periods that conform to business patterns as the target energy-saving time window.

[0010] Optionally, the multidimensional service characteristic data includes at least a combination of the following three: the number of terminal accesses in different time periods, uplink and downlink traffic, average received signal strength indication of users, interference channel utilization, low confidence channel utilization time and user online duration data.

[0011] Optionally, after triggering a business mode switch, the energy-saving window calculation method associated with the target business mode is invoked to dynamically generate the target energy-saving time window under the target business mode, and the target energy-saving time window is sent to the access point for execution, specifically including: After the business mode switch, a corresponding target energy-saving time window is generated based on the energy-saving window calculation method pre-bound to the target business mode, including: If the energy-saving window of the target business model is determined by automatically learning and recommending time periods based on historical data of the business model, then the tidal prediction algorithm is called to learn and generate the optimal energy-saving time period based on the pure historical multidimensional business feature data of the target business model; if the energy-saving window of the target business model is determined by manually configuring a fixed time period, then the user-preset fixed time period in the model is read directly. The system performs security checks on the consistency of the execution mode, the legality of the time, and the status of the access point within the target energy-saving time window. Once the checks are passed, the Network Cloud Engine (NCE) controller sends a configuration command containing the start and end times of energy saving, the execution strategy, and the effective period to the corresponding access point. After receiving the instruction and responding with confirmation, the access point immediately caches it and automatically enters or exits the energy-saving state at the specified time.

[0012] This invention provides a dynamic adjustment system for the energy-saving window of an access point, comprising: The data acquisition module is used to acquire multi-dimensional business characteristic data of the access point in real time; The business pattern library construction module is used to build a business pattern library. The business pattern library includes a variety of business patterns, and each business pattern includes pattern feature conditions consisting of multiple specific business features and their threshold ranges. The business mode switching module is used to match the multi-dimensional business feature data collected in the current period with the mode feature conditions of each business mode in the business mode library at preset periodic time points; when the matching result meets the preset switching conditions for switching from the current business mode to the target business mode, the business mode switching is triggered. The window determination module is used to call the energy-saving window calculation method associated with the target business mode after triggering the business mode switch, dynamically generate the target energy-saving time window under the target business mode, and send the target energy-saving time window to the access point for execution.

[0013] Optionally, the business model library is configured as follows: The system receives the user's selection of historical dates through a graphical user interface and defines a new business model name for the date range associated with the selected date. By using a classification model to perform machine learning on multidimensional business feature data within a date range, pattern feature conditions consisting of multiple specific business features and their threshold ranges are automatically derived and extracted. These pattern feature conditions are used to identify new business patterns. Save the name and characteristic conditions of the new business model.

[0014] Optionally, the graphical user interface is also configured as follows: When a user defines a new business model, the system receives the energy-saving window determination method independently specified by the user for the new business model. The energy-saving window determination method includes: manually configuring a fixed time period or automatically learning and recommending a time period based on historical data of the business model.

[0015] Optionally, the multidimensional service characteristic data is uploaded from the access point to the Network Cloud Engine (NCE) analyzer, and includes at least a combination of the following three: the number of terminal accesses in different time periods, uplink and downlink traffic, user average received signal strength indication, interference channel utilization, low confidence channel utilization time and user online duration data.

[0016] The method and system for dynamically adjusting the energy-saving window of an access point provided in this invention have the following advantages compared with the prior art: This invention acquires multi-dimensional service characteristic data from access points in real time and matches it with a dynamically maintained service mode library based on periodic time points. When a match is detected between the current service characteristics and the preset conditions of the target service mode, a mode switch is proactively triggered. After the switch, the energy-saving window calculation method associated with the target service mode is invoked to dynamically generate an energy-saving time window adapted to the current service characteristics and execute it immediately. This mechanism enables energy-saving strategies to closely follow the real-time evolution of service modes, significantly improving the sensitivity to dynamic changes in network services. It avoids energy-saving timing mismatches or performance losses caused by static configuration or historical learning lags, achieving a dynamic balance between service quality and energy-saving effects. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a method for dynamically adjusting an access point energy-saving window in one embodiment. Figure 2 A basketball game service date diagram illustrating a dynamic adjustment method for an access point energy-saving window provided in one embodiment; Figure 3 This is a schematic diagram of a service mode library for a dynamic adjustment method of an access point energy-saving window provided in one embodiment; Figure 4 This is a service feature diagram of a basketball game, illustrating a dynamic adjustment method for an access point energy-saving window provided in one embodiment. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0019] This invention provides a method for dynamically adjusting the energy-saving window of an access point, the method comprising: Real-time acquisition of multi-dimensional business characteristic data of access points.

[0020] Construct a business model library, which includes multiple business models. Each business model includes model feature conditions consisting of multiple specific business characteristics and their threshold ranges.

[0021] At preset periodic time points, multi-dimensional business feature data collected in the current period is matched with the pattern feature conditions of each business pattern in the business pattern library. When the matching result meets the preset switching conditions for switching from the current business pattern to the target business pattern, the business pattern switch is triggered.

[0022] After triggering the business mode switch, the energy-saving window calculation method associated with the target business mode is invoked to dynamically generate the target energy-saving time window under the target business mode, and the target energy-saving time window is sent to the access point for execution.

[0023] Preferably, such as Figure 3 As shown, the construction of a business pattern library specifically includes: receiving historical date ranges selected by the user through a user interface and defining business pattern names for each historical date range; extracting multi-dimensional business feature data corresponding to the historical date ranges; using a classification model to perform machine learning on each business feature data to obtain pattern feature conditions used to identify the business pattern; and storing the business patterns and their corresponding pattern feature conditions in the business pattern library.

[0024] Preferably, when a user defines a new business model, the system receives the energy-saving window determination method independently specified by the user for the new business model. The energy-saving window determination method includes: manually configuring a fixed time period or automatically learning and recommending a time period based on historical data of the business model.

[0025] Preferably, the method of calling the energy-saving window calculation method associated with the target business mode to dynamically generate the target energy-saving time window under the target business mode also includes: when the target business mode is configured to automatically learn and recommend time periods based on the historical data of the business mode, the tidal prediction algorithm is called to learn the data of all historical dates belonging to the target business mode in the business mode library to obtain the energy-saving period that conforms to the business rules as the target energy-saving time window.

[0026] Preferably, the multidimensional service characteristic data is uploaded from the access point to the Network Cloud Engine (NCE) analyzer, and includes at least a combination of the following three: the number of terminal accesses in different time periods, uplink and downlink traffic, user average received signal strength indication, interference channel utilization, low confidence channel utilization time and user online duration data.

[0027] Preferably, after triggering a business mode switch, the energy-saving window calculation method associated with the target business mode is invoked to dynamically generate the target energy-saving time window under the target business mode, and the target energy-saving time window is sent to the access point for execution, specifically including: After a business mode switch, a corresponding target energy-saving time window is generated based on the energy-saving window calculation method pre-bound to the target business mode. This includes: if the energy-saving window determination method for the target business mode is to automatically learn and recommend time periods based on historical data of that business mode, then the tidal prediction algorithm is invoked to learn and generate the optimal energy-saving time period based on the pure historical multi-dimensional business feature data of that target business mode. If the energy-saving window determination method for the target business mode is to manually configure a fixed time period, then the user-preset fixed time period for that mode is directly read.

[0028] The system performs security checks on the consistency of the execution mode, the legality of the time, and the status of the access point within the target energy-saving time window. Once the checks are passed, the Network Cloud Engine (NCE) controller sends configuration instructions, including the start and end times of energy saving, the execution strategy, and the effective period, to the corresponding access point.

[0029] After receiving the instruction and responding with confirmation, the access point immediately caches it and automatically enters / exits the energy-saving state at the specified time.

[0030] like Figure 1 As shown, a specific embodiment of a method for dynamically adjusting the energy-saving window of an access point is provided: During the initial configuration, the AP continuously collects data from multiple dimensions mentioned in the data module and uploads it to the NCE platform in real time via telemetry. The NCE side has built-in business characteristics for weekdays and holidays. If users have scenarios with other business characteristics, they should click on the dates that match these business patterns in the "Add Business Pattern" section and edit the business pattern name to add it to the pattern library. The NCE platform will learn the feature threshold of this business pattern using the data from these clicked dates. After editing the business pattern, users also need to decide whether to manually set or automatically recommend the energy-saving period. If automatically recommended, the NCE algorithm will learn the energy-saving period based on the data from dates that match the business pattern; if manually recommended, the user-filled energy-saving period will always be provided.

[0031] During the energy-saving window change process, the NCE platform will initiate the window change process at the end of each day's data collection. Based on the feature data collected that day, it will determine whether the current situation meets a certain business mode (each mode has a threshold range for each feature data). If it does, it will determine that the current situation needs to be switched to that mode, and will issue the latest energy-saving time period according to the user's pre-set automatic / manual recommended energy-saving time period, thereby achieving daily adaptive business features for energy-saving time periods.

[0032] This invention introduces an automatic energy-saving mode switching mechanism, no longer relying on fixed configurations or simple calendars; it utilizes multi-dimensional data fusion to determine business patterns and calculate energy-saving windows, significantly improving the accuracy of judgments and scenario adaptability. Complex energy-saving strategies are abstracted into different business patterns, improving manageability and interpretability; and because business patterns are determined daily, the energy-saving time window adapts to changes in business patterns.

[0033] Based on the same inventive concept, embodiments of the present invention provide a dynamic adjustment system for the energy-saving window of an access point, the system comprising: The data acquisition module is used to acquire multi-dimensional business characteristic data of the access point in real time.

[0034] The business model library construction module is used to build a business model library. The business model library includes a variety of business models, and each business model includes model feature conditions consisting of multiple specific business features and their threshold ranges.

[0035] The business mode switching module is used to match the multi-dimensional business feature data collected in the current period with the mode feature conditions of each business mode in the business mode library at preset periodic time points; when the matching result meets the preset switching conditions for switching from the current business mode to the target business mode, the business mode switching is triggered.

[0036] The window determination module is used to call the energy-saving window calculation method associated with the target business mode after triggering the business mode switch, dynamically generate the target energy-saving time window under the target business mode, and send the target energy-saving time window to the access point for execution.

[0037] Preferably, the business model library is further configured as follows: The system receives user selections of historical dates via a graphical user interface and defines a new business pattern name for the date range associated with the selected date. A classification model is applied to perform machine learning on the multidimensional business feature data within the date range, automatically deriving and extracting pattern feature conditions—composed of multiple specific business features and their threshold ranges—used to identify the new business pattern. The name of the new business pattern and the pattern feature conditions are then saved.

[0038] Preferably, the graphical user interface is further configured to receive a method for determining the energy-saving window, which is independently specified by the user for the new business model. The energy-saving window determination method includes: manually configuring a fixed time period or automatically learning and recommending a time period based on historical data of the business model.

[0039] Preferably, the multidimensional service characteristic data is uploaded from the access point to the Network Cloud Engine (NCE) analyzer, and includes at least a combination of the following three: the number of terminal accesses in different time periods, uplink and downlink traffic, user average received signal strength indication, interference channel utilization, low confidence channel utilization time and user online duration data.

[0040] A specific embodiment of a dynamic adjustment system for access point energy-saving windows is provided, comprising the following modules: 1. Data Acquisition Module: Continuously collects various service characteristic data sources from APs and uploads them to the NCE (Network Cloud Engine) analyzer. The data needs to be retained for N days. Service characteristic data includes: number of terminal accesses in different time periods, uplink and downlink traffic, average received signal strength index (RSSI) of users, interference channel utilization (CU), low confidence channel utilization time, and user online duration data.

[0041] 2. Energy-saving window repository: Saves previously used energy-saving time windows.

[0042] 3. Business Mode Library: The NCE platform determines the current network's business mode based on daily data. The database has pre-defined business modes, such as the entry conditions for weekday and holiday modes: Weekday mode: Access count and traffic are both within a certain threshold range; Holiday energy-saving mode: Access count and traffic are less than the threshold, and the number of days entering the full-day energy-saving mode is greater than the threshold. Users can also customize special business modes (such as sports stadium event mode, campus exam week mode) by clicking on the nth day of collected data. The mode library will automatically learn (e.g., using decision trees) the business characteristics of the days clicked by the user. Business characteristics refer to the data collected by the AP meeting certain thresholds.

[0043] 4. Business Mode Switching: Determine if the current state matches the characteristics of other modes. If so, trigger a mode switch. After switching, immediately adopt the new mode's energy-saving window calculation method, dynamically generating and distributing the optimal future energy-saving time window based on the latest data and mode characteristics.

[0044] 5. Window Calculation: Each business mode is associated with an energy-saving window calculation method. The calculation method can be a fixed time period configured by the user or a time period learned from existing data on the number of days of that business mode by calling an automatic learning algorithm.

[0045] It can be seen that the core of this invention lies in: 1. Design and interaction process of the business model library module.

[0046] Users define new business models by selecting existing historical data dates (or date ranges) through a graphical interface. During this definition process, users can independently choose how to determine the energy-saving time window for each business model (manually configuring a fixed time period or automatically learning and recommending a time period based on historical data for that model).

[0047] After the user selects a date, the system (NCE analyzer) automatically applies a classification model (such as decision tree, SVM, clustering, Bayesian network, etc.) to perform machine learning on the multi-dimensional time-segmented network feature data corresponding to the user's selected date (which must explicitly include at least three of the following: number of terminal accesses, uplink and downlink traffic, average RSSI, interference CU, low confidence channel utilization CU, and user online duration, preferably all of them). It automatically derives logical judgment conditions for identifying the service pattern, consisting of multiple features and their threshold ranges. This is the interactive and processing flow of user click -> system automatically learning feature thresholds.

[0048] 2. Design of the business model switching module.

[0049] The system (NCE analyzer) performs matching calculations at preset periodic time points (such as the end of each day) based on the complete multidimensional business feature data collected that day, against the feature threshold conditions of all business patterns in the pattern library. When the matching calculation result meets the preset switching conditions (for example, the data significantly meets the threshold conditions of a non-current pattern and may meet the continuity requirements), the business pattern switching action is immediately triggered.

[0050] Once a business mode switch is triggered, the energy-saving window calculation process associated with the newly switched business mode is immediately initiated (whether by obtaining manual configuration values ​​or triggering automatic learning calculation), and the calculated / configured energy-saving time window is distributed to the AP side for execution via the controller within a short period of time (such as the next calendar day). This is a linkage mechanism of "daily data matching -> switching upon condition fulfillment -> immediate triggering of new mode window calculation and distribution".

[0051] In addition, such as Figure 2 As shown, taking a basketball court (where the AP energy-saving period differs between the basketball season and the off-season to adapt to actual business needs) as the implementation object, an application embodiment of the present invention is provided: 1. Front-end workflow.

[0052] Users can click on the days in the system calendar where AP collected data, such as clicking on days 19, 20, 22, and 23 of a past season and editing the business name as "Basketball Game". Figure 4As shown, the system automatically generates business characteristic thresholds for these four days as the basis for switching to the basketball game mode. After the user completes the editing, a basketball game entry appears in the business mode library, and the user can choose a recommended method for energy-saving time periods. If automatic is selected, the system will use tidal prediction algorithms to learn energy-saving time periods based on historical data of basketball game days; the recommended time period here is 12:00-08:00. If manual is selected, the energy-saving time periods can be configured manually. The user can choose whether to enable the mode; if disabled, the system will no longer determine whether a particular day belongs to that mode.

[0053] 2. Backend workflow.

[0054] On the NCE side, a supervised learning algorithm (such as a decision tree) is deployed to classify features based on the number of days the user clicks, generating a data threshold for entering the basketball game category. If the data on a subsequent day matches the threshold features, it is determined that the current business has entered the basketball game mode. Since the user has configured the automatic generation of energy-saving time periods, the tidal prediction algorithm will learn from the existing basketball game date data to determine the energy-saving time period of 12:00-08:00 that conforms to the business pattern, and finally issue this time period.

[0055] The embodiments described above are merely examples of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention.

Claims

1. A method for dynamically adjusting the energy-saving window of an access point, characterized in that, include: Real-time acquisition of multi-dimensional business characteristic data of access points; Construct a business model library, which includes multiple business models, each of which includes model feature conditions consisting of multiple specific business features and their threshold ranges; At preset periodic time points, the multidimensional business feature data collected in the current period is matched with the pattern feature conditions of each business pattern in the business pattern library. When the matching result meets the preset switching conditions for switching from the current business mode to the target business mode, the business mode switching is triggered; After triggering the business mode switch, the energy-saving window calculation method associated with the target business mode is invoked to dynamically generate the target energy-saving time window under the target business mode, and the target energy-saving time window is sent to the access point for execution.

2. The method according to claim 1, characterized in that, The construction of the business model library specifically includes: The system receives the historical date range selected by the user through the user interface and defines a business mode name for the historical date range. Extract the multi-dimensional business feature data corresponding to the historical date range; A classification model is used to perform machine learning on each of the aforementioned business feature data to obtain pattern feature conditions for identifying the business pattern. The business model and its corresponding model feature conditions are stored in the business model library.

3. The method for dynamically adjusting the energy-saving window of an access point as described in claim 2, characterized in that, When the user defines a new business model, the system receives the energy-saving window determination method independently specified by the user for the new business model; the energy-saving window determination method includes: manually configuring a fixed time period or automatically learning and recommending a time period based on historical data of the business model.

4. The method for dynamically adjusting the energy-saving window of an access point as described in claim 3, characterized in that, The step of calling the energy-saving window calculation method associated with the target business mode to dynamically generate the target energy-saving time window under the target business mode also includes: When the target business model is configured to automatically learn and recommend time periods based on historical data of the business model, the tidal prediction algorithm is called to learn all historical date data belonging to the target business model in the business model library, and obtain energy-saving time periods that conform to business rules as the target energy-saving time window.

5. The method for dynamically adjusting the energy-saving window of an access point as described in claim 1, characterized in that, The multidimensional service characteristic data includes at least a combination of the following three: the number of terminal accesses in different time periods, uplink and downlink traffic, average received signal strength indication of users, interference channel utilization, low confidence channel utilization time and user online duration data.

6. The method for dynamically adjusting the energy-saving window of an access point as described in claim 1, characterized in that, After triggering a business mode switch, the step of invoking the energy-saving window calculation method associated with the target business mode to dynamically generate a target energy-saving time window under the target business mode, and then sending the target energy-saving time window to the access point for execution, specifically includes: After the business mode switch, a corresponding target energy-saving time window is generated based on the energy-saving window calculation method pre-bound to the target business mode, including: If the energy-saving window of the target business model is determined by automatically learning and recommending time periods based on historical data of the business model, then the tidal prediction algorithm is called to learn and generate the optimal energy-saving time period based on the pure historical multidimensional business feature data of the target business model; if the energy-saving window of the target business model is determined by manually configuring a fixed time period, then the user-preset fixed time period in the model is read directly. The security verification of the target energy-saving time window execution mode consistency, time legality, and access point status is performed. After the verification is passed, the Network Cloud Engine (NCE) controller sends the configuration command containing the energy-saving start and end time, execution strategy, and effective period to the corresponding access point. After receiving the instruction and responding with confirmation, the access point immediately caches it and automatically enters or exits the energy-saving state at a specified time.

7. A dynamic adjustment system for the energy-saving window of an access point, characterized in that, include: The data acquisition module is used to acquire multi-dimensional business characteristic data of the access point in real time; A business pattern library construction module is used to build a business pattern library, which includes a variety of business patterns. Each business pattern includes pattern feature conditions consisting of multiple specific business features and their threshold ranges. The business mode switching module is used to match the multi-dimensional business feature data collected in the current period with the mode feature conditions of each business mode in the business mode library at preset periodic time points. When the matching result meets the preset switching conditions for switching from the current business mode to the target business mode, the business mode switching is triggered; The window determination module is used to call the energy-saving window calculation method associated with the target business mode after triggering the business mode switch, dynamically generate the target energy-saving time window under the target business mode, and send the target energy-saving time window to the access point for execution.

8. The dynamic adjustment system for the energy-saving window of an access point as described in claim 7, characterized in that, The business model library is configured as follows: The system receives the user's selection of historical dates through a graphical user interface and defines a new business model name for the date range associated with the selected date. The classification model is used to perform machine learning on the multidimensional business feature data within the date range, automatically deriving and extracting pattern feature conditions composed of multiple specific business features and their threshold ranges. The pattern feature conditions are used to identify the new business pattern. Save the name and feature conditions of the new business model.

9. The dynamic adjustment system for the energy-saving window of an access point as described in claim 8, characterized in that, The graphical user interface is also configured to: When the user defines a new business model, the system receives the energy-saving window determination method independently specified by the user for the new business model; the energy-saving window determination method includes: manually configuring a fixed time period or automatically learning and recommending a time period based on historical data of the business model.

10. The dynamic adjustment system for the energy-saving window of an access point as described in claim 7, characterized in that, The multidimensional service characteristic data is uploaded from the access point to the Network Cloud Engine (NCE) analyzer, and includes at least a combination of the following three: the number of terminal accesses in different time periods, uplink and downlink traffic, user average received signal strength indication, interference channel utilization, low confidence channel utilization time and user online duration data.

Images