Automatic monitoring method for radio frequency single board, network management device, and storage medium

Abstract

The present application discloses an automatic monitoring method for a radio frequency single board, a network management device, and a storage medium. The method comprises: acquiring a thermal imaging graph of an indoor radio frequency single board, and current operation state data of multiple elements of the radio frequency single board; according to the thermal imaging graph and the operation state data, determining a current operation condition of the indoor radio frequency single board, the current operation condition of the radio frequency single board comprising normal operation or a fault; according to the current operation condition of the indoor radio frequency single board, controlling a total coverage range of an indoor radio frequency signal to at least reach a total coverage range corresponding to a preset minimum coverage threshold.

Description

Automatic monitoring method, network management equipment and storage medium for radio frequency single board

[0001] Cross-references

[0002] This application claims priority to a Chinese patent application filed with the Patent Office of China on June 13, 2023, with application number 202310698969.X and titled “Automatic Monitoring Method, Network Management Equipment and Storage Medium for Radio Frequency Boards”. The entire contents of the application are incorporated by reference into this application. Technical Field

[0003] The present application relates to the field of intelligent monitoring technology, and in particular to an automatic monitoring method, network management equipment and storage medium for a radio frequency board. Background Art

[0004] In the field of mobile communication networks, with the increasing popularity of 5G (5th Generation Mobile Communication Technology) services, technologies such as the Internet of Things (IoE), the Internet of Things (IoT), AI (Artificial Intelligence), and holographic projection will affect all aspects of life. Therefore, a good indoor signal is crucial. In the 5G and 4G (fourth Generation Mobile Communication Technology) eras, solutions such as RRU (Remote Radio Unit) board diagnosis were provided to address indoor signal coverage. However, the data sources used to diagnose RRU boards in these solutions were relatively single and lacked diversity, resulting in inaccurate RRU board diagnosis. Furthermore, after diagnosis, only prompts and alarms were issued regarding current RRU board faults. The resolution of RF signal issues caused by RRU board faults was significantly delayed. In actual user experience, the signal impact caused by RRU board faults was very direct and serious.

[0005] Summary of the Invention

[0006] Based on this, the embodiments of the present application provide an automatic monitoring method, network management device and storage medium for a radio frequency board, which can improve the accuracy of radio frequency board diagnosis and accurately ensure that the total coverage range of indoor radio frequency signals always meets user requirements.

[0007] In a first aspect, the present application provides an automatic monitoring method for a radio frequency board, which is applied to a network management device, and the method includes: obtaining a thermal imaging image of the radio frequency board indoors and the current operating status data of multiple components of the radio frequency board; determining the current operating status of the radio frequency board indoors based on the thermal imaging image and the operating status data, the current operating status of the radio frequency board including normal operation or a fault; and controlling the total coverage range of the radio frequency signal indoors to at least reach the total coverage range corresponding to a preset minimum coverage threshold based on the current operating status of the radio frequency board indoors.

[0008] In a second aspect, the present application provides a network management device, which includes: a communication module, a memory, and a processor, wherein the communication module is used for communication, the memory is used for storing a computer program; the processor is used to execute the computer program and implement the automatic monitoring method of the radio frequency board as described above when executing the computer program.

[0009] In a third aspect, the present application provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor implements the automatic monitoring method of the radio frequency board as described above. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] FIG1 is a schematic structural diagram of an embodiment of an RRU board in the related art;

[0011] FIG2 is a flow chart of an embodiment of an automatic monitoring method for a radio frequency board according to the present invention;

[0012] FIG3 is a flow chart of another embodiment of the automatic monitoring method of the radio frequency board of the present application;

[0013] FIG4 is a flow chart of another embodiment of the automatic monitoring method of the radio frequency board of the present application;

[0014] FIG5 is a flow chart of another embodiment of the automatic monitoring method of the radio frequency board of the present application;

[0015] FIG6 is a flow chart of another embodiment of the automatic monitoring method of the radio frequency board of the present application;

[0016] FIG7 is a flow chart of another embodiment of the automatic monitoring method of the radio frequency board of the present application;

[0017] FIG8 is a flow chart of another embodiment of the automatic monitoring method of the radio frequency board of the present application;

[0018] FIG9 is a schematic structural diagram of an embodiment of a network management device of the present application;

[0019] FIG10 is a thermal image of an embodiment of an automatic monitoring method for a radio frequency board of the present application;

[0020] 11 is a schematic diagram of three ROC curves formed through training in an embodiment of an automatic monitoring method for a radio frequency board of the present application;

[0021] 12 is a structural diagram of an embodiment of an RF board and a network management device in an automatic monitoring method of the RF board of the present application;

[0022] 13 is a schematic diagram of a method for automatically monitoring a radio frequency board according to an embodiment of the present invention, wherein a portion of the indoor signal coverage in a blind spot is retained;

[0023] FIG14 is a schematic diagram showing that the signal coverage range approximately conforms to the disk Boolean model in an embodiment of the automatic monitoring method of the radio frequency board of the present application. DETAILED DESCRIPTION

[0024] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0025] The flowcharts shown in the accompanying drawings are for illustrative purposes only and do not necessarily include all contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may be decomposed, combined, or partially merged, so the actual execution order may vary depending on the actual situation.

[0026] In the subsequent description, suffixes such as "module," "component," or "unit" used to represent elements are used only to facilitate the description of the present application and have no specific meaning. Therefore, "module," "component," or "unit" may be used interchangeably.

[0027] Before introducing the embodiments of the present application in detail, the relevant technical content is first introduced.

[0028] In the mobile communications network sector, 40% of complaints in the 4G era were related to network coverage, 70% of which stemmed from indoor use. In the 4G era, 70%-80% of all services took place indoors. With the increasing popularity of 5G services, technologies like the Internet of Everything, the Internet of Things, AI, and holographic projection will impact every aspect of life. Therefore, good indoor signal quality will be crucial. Precisely because of the need for good indoor signal coverage, it's crucial to be able to quickly detect changes in indoor signal quality and dynamically adjust the power of RF-related equipment to ensure a usable indoor signal. This ability to adjust the signal based on the current environment can also reduce user complaints related to indoor use.

[0029] The commonly used indoor distribution solution in the 5G era is digital indoor distribution, and the commonly used digital indoor distribution system is Qcell, which consists of a baseband unit (BBU), a remote convergence unit (P-Bridge), and a remote radio frequency unit (pRRU). The commonly used indoor distribution solution in the 4G era is a base station with an RRU radio frequency unit to solve the problem of indoor signal coverage. However, in these commonly used indoor distribution solutions, the RRU board, as the last link and the most critical equipment in wireless communication, is like a bridge in the sky, providing a stable and reliable channel for user information exchange and ensuring accurate and real-time delivery of information. Referring to Figure 1, in some embodiments, the RRU board 100 may include a transceiver, a power amplifier module (including a power amplifier), a duplex filter, and a power supply module; the transceiver includes an AISG\MON interface and a CPRI (Common Public Radio Interface). The AISG\MON interface is an electrical adjustment interface that provides an RRU board commissioning port; the input of the power supply module is a DC (Direct Current) input; the duplex filter is grounded and includes four ANTs (antenna terminals). The RRU converts baseband digital signals into radio waves through complex and sophisticated circuitry, transmitting them through the antenna. Simultaneously, it receives information from user terminals and transmits it to the core network for information exchange. To ensure wide signal coverage, the RRU board is associated with a power amplifier (PA). The PA is often one of the most power-hungry and inefficient components on an RRU board. Therefore, given the same bandwidth and frequency band, indoor RF signal coverage is influenced by the RRU board, the power amplifier, and other related factors.

[0030] A common RF signal problem on the RF board is the RF signal coverage problem caused by RRU-related faults. Current network management also provides relevant RRU diagnosis and other solutions for this problem. However, the data sources used to diagnose RRU boards in these solutions are relatively single and lack diversity. The diagnosis of RRU boards is inaccurate, and after diagnosis, prompts and alarms are often only given for the current RRU board fault. The solution to the RF signal problem caused by the RRU board fault is relatively delayed. In actual experience, users are very intuitive and serious about the signal impact caused by the RRU board fault.

[0031] In the embodiment of the present application, a thermal image of the indoor RF board and the current operating status data of multiple components of the RF board are obtained; based on the thermal image and the operating status data, the current operating status of the indoor RF board is determined, and the current operating status of the RF board includes normal operation or a fault; based on the current operating status of the indoor RF board, the total coverage range of the indoor RF signal is controlled to at least reach the total coverage range corresponding to the preset minimum coverage threshold. Compared with the single data source when diagnosing whether the RF board is faulty in the related art, in the embodiment of the present application, the basis for determining the current operating status of the RF board includes two data sources: the thermal image of the RF board and the current operating status data of multiple components of the RF board. Therefore, the current operating status of the RF board can be determined more accurately, and the total coverage range of the indoor RF signal can be more accurately ensured to always meet the user's requirements when the current operating status of the RF board is more accurate.

[0032] The embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0033] 2 , which is a flow chart of an embodiment of a method for automatically monitoring a radio frequency board of the present application, the method is applied to a network management device, and the method includes: step S101 , step S102 , and step S103 .

[0034] Step S101: obtaining a thermal image of the radio frequency board indoors and current operating status data of multiple components of the radio frequency board.

[0035] A radio frequency board may refer to a board that can transmit and receive radio frequency signals, including but not limited to an RRU board. A radio frequency board may include one or more radio frequency units. For example, an RRU board may include one RRU unit or multiple RRU units. An infrared imager with a communication function may be provided inside the radio frequency board. The infrared imager may perform real-time infrared monitoring of the inside of the radio frequency board and form a thermal image at the corresponding moment (a thermal image of the entire radio frequency board). The infrared imager may send the thermal image to a network management device. A data acquisition module with a communication function may also be provided inside the radio frequency board. The data acquisition module may collect operating status data of multiple components (also called devices, belonging to hardware) inside the radio frequency board and form operating status data at the corresponding moment. The data acquisition module may send the operating status data to a network management device. Typically, the data acquisition module collects operating status data of multiple components inside the radio frequency board through software.

[0036] Different components within a RF board have different operating status data. The current operating status data for the multiple components on the RF board includes at least one of electrical data (e.g., current, voltage, power, etc.), temperature data (e.g., temperature, etc.), and RF data (e.g., signal strength, frequency, bandwidth, link connectivity, noise level, etc.). The current operating status data for the multiple components on the RF board may also include data such as rotational speed, for example, fan speed.

[0037] The radio frequency signal related data includes radio frequency related data and power in electrical related data.

[0038] It should be noted that the infrared imager and data acquisition module inside the RF board perform real-time monitoring, which can be performed at predetermined time intervals or triggered when a fault occurs.

[0039] Step S102: determining the current operating status of the indoor radio frequency board according to the thermal image and the operating status data, where the current operating status of the radio frequency board includes normal operation or failure.

[0040] The thermal image of the RF board can directly reflect the temperature conditions of multiple components on the RF board. Based on the thermal image of the RF board, abnormal components with abnormal temperatures among the multiple components on the RF board can be analyzed. Combined with the operating status data of the multiple components on the RF board, the current operating status of the RF board can be determined relatively accurately, that is, whether the RF board is operating normally or has a fault.

[0041] It should be noted that if any component on the RF board fails, the RF board is considered faulty. For example, if a RF board contains a single RF unit and any component on that unit fails, the RF board is considered faulty. Alternatively, if a RF board contains multiple RF units and any one of them fails, the RF board is considered faulty.

[0042] Step S103: According to the current operation status of the indoor radio frequency board, the total coverage range of the indoor radio frequency signal is controlled to at least reach the total coverage range corresponding to the preset minimum coverage threshold.

[0043] After determining the current operating status of the indoor RF board, the total coverage of the indoor RF signal can be controlled accordingly to at least reach the total coverage corresponding to a preset minimum coverage threshold. The preset minimum coverage threshold can be the minimum coverage percentage at which indoor signal coverage meets user needs, i.e., the percentage between the minimum total coverage at which indoor signal coverage meets user needs and the total coverage when the room is fully covered. Because the current operating status of the indoor RF board is determined relatively accurately, the total coverage of the indoor RF signal can be controlled more accurately accordingly.

[0044] In the embodiment of the present application, a thermal image of the indoor RF board and the current operating status data of multiple components of the RF board are obtained; based on the thermal image and the operating status data, the current operating status of the indoor RF board is determined, and the current operating status of the RF board includes normal operation or a fault; based on the current operating status of the indoor RF board, the total coverage range of the indoor RF signal is controlled to at least reach the total coverage range corresponding to the preset minimum coverage threshold. Compared with the single data source when diagnosing whether the RF board is faulty in the related art, in the embodiment of the present application, the basis for determining the current operating status of the RF board includes two data sources: the thermal image of the RF board and the current operating status data of multiple components of the RF board. Therefore, the current operating status of the RF board can be determined more accurately, and the total coverage range of the indoor RF signal can be more accurately ensured to always meet the user's requirements when the current operating status of the RF board is more accurate.

[0045] In some embodiments, step S103, controlling the total coverage range of the indoor radio frequency signal to at least reach the total coverage range corresponding to the preset minimum coverage threshold based on the current operating status of the indoor radio frequency board, may include: when there is a fault in the indoor radio frequency board, controlling the total coverage range of the indoor radio frequency signal to at least reach the total coverage range corresponding to the preset minimum coverage threshold based on the current operating status of the indoor radio frequency board.

[0046] According to the current operation status of the indoor RF board, the total coverage range of the indoor RF signal is controlled to at least reach the total coverage range corresponding to the preset minimum coverage threshold. The indoor RF signal can be monitored in real time, and the indoor RF signal can also be monitored when there is a fault in the indoor RF board.

[0047] Generally speaking, if the indoor RF board operates normally and has no faults, the total coverage range of the indoor RF signal is usually within the normal coverage range, that is, it usually at least reaches the total coverage range corresponding to the preset minimum coverage threshold. Only when the indoor RF board has a fault will the total coverage range of the indoor RF signal be affected. Real-time monitoring of indoor RF signals also generates a lot of consumption (such as power consumption, resource consumption, etc.). Therefore, in order to reduce consumption, when the indoor RF board has a fault, the total coverage range of the indoor RF signal can be controlled to at least reach the total coverage range corresponding to the preset minimum coverage threshold based on the current operating status of the indoor RF board.

[0048] In some embodiments, step S103, controlling the total coverage range of the indoor radio frequency signal to at least reach the total coverage range corresponding to the preset minimum coverage threshold based on the current operation status of the indoor radio frequency board, may also include: sub-step S1031 and sub-step S1032, as shown in Figure 3.

[0049] Sub-step S1031: determining the total coverage of the indoor radio frequency signal according to radio frequency signal related data of the indoor radio frequency board, wherein the radio frequency signal related data belongs to the data in the operating status data.

[0050] Sub-step S1032: According to the total coverage range of the indoor radio frequency signal, control the total coverage range of the indoor radio frequency signal to at least reach the total coverage range corresponding to the preset minimum coverage threshold.

[0051] In the embodiments of the present application, the RF signal-related data may refer to data related to the RF signal. The RF signal-related data is included in the operating status data. In this way, the RF signal-related data in the operating status data can be directly used to avoid recollecting other data. Most of the RF signal-related data is RF-related data (for example, signal strength, bandwidth, frequency, etc.). Since the power, especially the power of the power amplifier module (including the power amplifier), affects the RF signal, the RF signal-related data may also include the power in the electrical-related data.

[0052] The failure of the RF board can be divided into unavailable failure and available failure; the unavailable failure makes the RF unit (such as RRU unit) of the RF board unable to send RF signals normally, and the coverage range of the RF signal of the RF board with the unavailable failure is basically zero, and the power amplifier effect therein is actually ineffective. The power amplifier will generate unnecessary power consumption if it is always running; the available failure often makes the RF unit (such as RRU unit) of the RF board unable to reach normal power, and the coverage range of the RF signal of the RF board with the available failure is basically smaller than the coverage range of the RF signal of the normally operating RF board, thereby affecting the coverage range of the RF signal and reducing the strength of the RF signal.

[0053] Indoor RF signals are typically transmitted by normally functioning RF boards, with a small portion also transmitted by available faulty RF boards. When determining the total indoor RF signal coverage, if there are no faulty RF boards, the total indoor RF signal coverage can be determined based on the RF signal data from normally functioning RF boards. If there are faulty RF boards, a coarse approach can be used to determine the total indoor RF signal coverage based on the RF signal data from normally functioning RF boards. A more detailed approach can be used to determine the total indoor RF signal coverage based on the RF signal data from both normally functioning RF boards and the available faulty RF boards.

[0054] Since the coverage range of the RF signal obtained from the RF signal-related data of the unavailable faulty RF board is basically zero, determining the total coverage range of the indoor RF signal based on the RF signal-related data of the indoor RF board is actually determining the total coverage range of the indoor RF signal based on the RF signal-related data of the indoor normally operating RF board, or determining the total coverage range of the indoor RF signal based on the RF signal-related data of the indoor normally operating RF board and the RF signal-related data of the available faulty RF board.

[0055] After obtaining the total coverage range of the indoor RF signal, combined with the preset minimum coverage threshold, the total coverage range of the indoor RF signal can be controlled to at least reach the total coverage range corresponding to the preset minimum coverage threshold. That is, if the total coverage range of the indoor RF signal at least reaches the total coverage range corresponding to the preset minimum coverage threshold, the state of the current indoor RF board can be directly maintained. If the total coverage range of the indoor RF signal is less than the total coverage range corresponding to the preset minimum coverage threshold, it is necessary to adjust the state of the normally operating indoor RF board, increase the power of the power amplifier module, etc., so that the adjusted total coverage range of the indoor RF signal at least reaches the total coverage range corresponding to the preset minimum coverage threshold.

[0056] In some embodiments, sub-step S1032, controlling the total coverage range of the indoor radio frequency signal to at least reach the total coverage range corresponding to a preset minimum coverage threshold based on the total coverage range of the indoor radio frequency signal, may include: when the current total coverage range of the indoor radio frequency signal is less than the total coverage range corresponding to the preset minimum coverage threshold, controlling the normally operating radio frequency board indoors to perform power adjustment so that the adjusted total coverage range of the indoor radio frequency signal is equal to the total coverage range corresponding to the preset minimum coverage threshold.

[0057] Since available faults often make it impossible for the RF unit (such as RRU unit) of the RF board to reach normal power, when adjusting the total coverage range of the indoor RF signal, it is usually necessary to adjust the normally operating RF board and perform power adjustment on the normally operating RF board so that the adjusted total coverage range of the indoor RF signal is equal to the total coverage range corresponding to the preset minimum coverage threshold. For example, the coverage range of the RF signal of certain normally operating RF boards can be increased, thereby increasing the current total coverage range of the indoor RF signal. Of course, the coverage range of the RF signal of certain normally operating RF boards can also be reduced. For example, when the overlapping coverage range of the RF signals of multiple RF boards is relatively large, the coverage range of the RF signal of certain normally operating RF boards can be reduced to reduce the overlapping coverage range, thereby reducing consumption.

[0058] Compared with the related art, after diagnosis, only prompts and alarms are given for the current fault of the RF board, and the RF signal problem can only be solved after the faulty RF board is repaired, resulting in a large lag in the solution of the RF signal problem. In the embodiment of the present application, when the current total coverage range of the indoor RF signal is less than the total coverage range corresponding to the preset minimum coverage threshold, it is also possible to adjust the power of the normally operating RF board to always keep the adjusted total coverage range of the indoor RF signal equal to the total coverage range corresponding to the preset minimum coverage threshold, that is, the current indoor RF signal is automatically repaired (i.e., self-healed), thereby solving the problem of a large lag in the solution of the RF signal problem, and avoiding the user experience being affected by the current total coverage range of the indoor RF signal being less than the total coverage range corresponding to the preset minimum coverage threshold; at the same time, on the basis of ensuring that the user's indoor video signal experience is not affected before the fault is handled, sufficient repair time is provided for the operation and maintenance personnel to handle the fault, which can also reduce the operation and maintenance pressure of the operation and maintenance personnel, improve the overall network maintenance environment, reduce indoor distribution-related complaints, and minimize consumption.

[0059] In some embodiments, when the coverage range of the adjusted indoor RF signal is equal to the total coverage range corresponding to the preset minimum coverage threshold, the total power consumed by the adjusted RF board among the RF boards operating normally indoors is the lowest.

[0060] To consistently maintain the adjusted total coverage range of the indoor RF signal equal to the total coverage range corresponding to the preset minimum coverage threshold, various adjustment schemes are possible. To minimize energy consumption while meeting user needs, the total power consumed by the adjusted RF board among the normally operating indoor RF boards can be minimized when the adjusted indoor RF signal coverage range equals the total coverage range corresponding to the preset minimum coverage threshold.

[0061] In some embodiments, sub-step S1032, controlling the power regulation of the indoor normally operating RF board, may further include: sub-step S10321 and sub-step S10322, as shown in FIG4 .

[0062] Sub-step S10321: using a resource allocation algorithm to determine the RF board to be adjusted and the target power value to be adjusted among the RF boards operating normally indoors.

[0063] Sub-step S10322: dynamically controlling the power of the power amplifier module of the RF board to be adjusted among the RF boards operating normally indoors based on the target power value.

[0064] A resource allocation algorithm refers to an algorithm that divides all available resources to each network user or resource consumer. It mainly solves the problem of allocating limited resources fairly and accurately to each user. This algorithm can fairly allocate available resources to each user or organization based on the user's needs. According to the different characteristics of different resource allocation algorithms, they can be divided into three categories: spatial allocation algorithm, time allocation algorithm, and algorithm allocation algorithm. In an embodiment of the present application, a resource allocation algorithm is used to determine the RF board to be adjusted and the target power value to be adjusted in the RF board operating normally indoors. In this way, when the coverage range of the adjusted indoor RF signal is equal to the total coverage range corresponding to the preset minimum coverage threshold, the power of the power amplifier module of the RF board to be adjusted in the RF board operating normally indoors is dynamically controlled based on the target power value, which can more fairly increase the power of the power amplifier module of the RF board to be adjusted.

[0065] In some embodiments, when the current total coverage range of the indoor RF signal is less than the total coverage range corresponding to a preset minimum coverage threshold, a greedy algorithm is used to determine the RF board to be adjusted and the target power value to be adjusted among the RF boards operating normally indoors.

[0066] A greedy algorithm (also known as a greedy algorithm) is a simpler and faster technique for finding the optimal solution to certain problems. It often makes the best choice based on a certain optimization measure based on the current situation, without considering various possible overall situations, thus saving a lot of time that must be spent on exhausting all possibilities to find the optimal solution.

[0067] In some embodiments, the method further includes: when it is determined that the faulty RF board has resumed normal operation and the current total coverage range of the indoor RF signal is greater than the total coverage range corresponding to the preset minimum coverage threshold, controlling the adjusted RF board among the indoor normally operating RF boards to restore to the power before adjustment.

[0068] In an embodiment of the present application, if the maintenance personnel have completed the repair of the faulty RF board, determined that the faulty RF board has resumed normal operation, and confirmed that the current total coverage range of the indoor RF signal is greater than the total coverage range corresponding to the preset minimum coverage threshold, then the adjusted RF board among the indoor normally operating RF boards can be controlled to restore to its original state, that is, to restore to the power before adjustment, to avoid long-term overload operation of the adjusted RF board affecting its service life.

[0069] In some embodiments, sub-step S1031, determining the total coverage range of the indoor RF signal based on the RF signal related data of the indoor RF board, may include: sub-step S10311 and sub-step S10312, as shown in FIG5 .

[0070] Sub-step S10311: determining the coverage range of the radio frequency signal of each radio frequency board in the room according to the radio frequency signal related data of the radio frequency board in the room.

[0071] Sub-step S10312: determining a current total coverage range of the indoor radio frequency signal according to the coverage ranges of the radio frequency signals of the multiple radio frequency boards currently in the room and the indoor location.

[0072] In an embodiment of the present application, the coverage range of the RF signal of each current RF board in the room is first determined based on the RF signal-related data of the current RF board in the room; since the RF signals between the RF boards may have overlapping coverage problems, the current total coverage range of the indoor RF signal can be determined more accurately in combination with the indoor positions of multiple RF boards.

[0073] In one embodiment, sub-step S10311, determining the coverage range of the RF signal of each current RF board in the room based on the RF signal-related data of the current RF board in the room, includes: based on a recurrent neural network, determining the coverage range of the RF signal of each current RF board in the room based on the RF signal-related data of the current RF board in the room.

[0074] A recurrent neural network (RNN) is a type of recursive neural network that takes sequence data as input and recursively evolves in the direction of the sequence, with all nodes (recurrent units) connected in a chain-like fashion. Bidirectional RNNs (Bi-RNNs) and long short-term memory networks (LSTMs) are common RNNs. RNNs possess memory, parameter sharing, and Turing completeness, making them particularly advantageous for learning nonlinear features of sequences.

[0075] In some embodiments, based on a long short-term memory network, the coverage range of the radio frequency signal of each current radio frequency board in the room is determined according to radio frequency signal related data of the current radio frequency board in the room.

[0076] In some embodiments, sub-step S10311, determining the current total coverage of the indoor RF signal based on the coverage of the RF signals of the current multiple RF boards in the room and the indoor location, may also include: determining the current total coverage area of ​​the indoor RF signal based on the coverage of the RF signals of the current multiple RF boards in the room and the indoor location based on a grid-based approximation algorithm.

[0077] A grid-based approximation algorithm divides the entire indoor area into a large number of grids. It calculates the coverage of any circle based on the distance from each grid to the center of the circle (i.e., the RF board). Finally, it summarizes all covered grids to derive the total coverage area. In some embodiments, a diamond grid solution can be used to achieve complete and seamless coverage of the area by placing circles at each vertex of the diamond grid. This can provide an analytical expression for the minimum number of circles required for complete and seamless coverage.

[0078] In some embodiments, the method further includes: when determining the current total coverage range of the indoor radio frequency signal, constructing a current signal coverage map of the indoor radio frequency signal.

[0079] In an embodiment of the present application, when determining the current total coverage range of the indoor radio frequency signal, a current signal coverage map of the indoor radio frequency signal is constructed. The signal coverage map can more intuitively and vividly display the current total coverage range of the indoor radio frequency signal to the user.

[0080] In some embodiments, the method further includes: when it is determined that the RF board has an unavailable fault, controlling the power of the power amplifier module of the RF board with the unavailable fault to be set to zero, and the fault includes the unavailable fault and the available fault.

[0081] When an RF board is unavailable, it cannot send RF signals normally. The power amplifier module in the RF board is actually ineffective. Continuously running the power amplifier module consumes unnecessary power. Therefore, setting the power of the power amplifier module of the unavailable RF board to zero can avoid unnecessary power consumption.

[0082] In some embodiments, the method further includes: when it is determined that the radio frequency board is faulty, issuing a prompt message, wherein the prompt message is used to notify a user to repair the faulty radio frequency board.

[0083] In the embodiment of the present application, if it is determined that the RF board is faulty, a prompt message is issued to prompt the user to repair the faulty RF board, so that the faulty RF board can resume normal operation as soon as possible.

[0084] In some embodiments, step S102, determining the current operating status of the indoor RF board based on the thermal imaging image and the operating status data, may include: sub-step S1021 and sub-step S1022, as shown in FIG6 .

[0085] Sub-step S1021: Analyze the thermal image using a convolutional neural network to obtain analysis data of the thermal image.

[0086] Sub-step S1022: determining the current operating status of each of the radio frequency boards in the room according to the analysis data of the thermal imaging image and the operating status data.

[0087] The embodiment of the present application uses a convolutional neural network to process thermal imaging images, which can obtain corresponding analysis data with a smaller amount of calculation, thereby improving the efficiency of diagnosing whether the radio frequency board is faulty.

[0088] In some embodiments, sub-step S1021, analyzing the thermal image using a convolutional neural network to obtain analysis data of the thermal image, may include: sub-step S10211, sub-step S10212, sub-step S10213, and sub-step S10214, as shown in FIG7 .

[0089] Sub-step S10211: Based on the convolution kernel of the first convolution layer, obtain first data of the thermal image according to the thermal image, and the first data of the thermal image is used to determine the positions of multiple components of the RF board on the thermal image.

[0090] Sub-step S10212: determining fault weights of the plurality of components of the RF board according to the first data of the thermal imaging image.

[0091] Sub-step S10213: obtaining second data of the thermal imaging image according to the first data of the thermal imaging image and the fault weights of the plurality of components of the RF board.

[0092] Sub-step S10214: Based on the convolution kernel of the second convolution layer, obtain the fault feature vector of the RF board according to the second data of the thermal imaging image.

[0093] The embodiment of the present application uses two convolution kernels to obtain a fault feature vector, which has high computational efficiency. In addition, since different types of RF boards may have different positions of components thereon, the degree of impact of the fault on the RF board may be different. Therefore, the fault weights of the multiple components of the RF board are determined according to their positions on the thermal imaging image, which enables the processing method to adapt to various different RF boards.

[0094] In some embodiments, sub-step S1022, determining the current operating status of the indoor RF board based on the analysis data of the thermal imaging image and the operating status data, may include: sub-step S10221 and sub-step S10222, as shown in Figure 8.

[0095] Sub-step S10221: Based on the fault binary classification model, according to the analysis data of the thermal image and the operating status data, determine whether the radio frequency board is operating normally or has a fault.

[0096] Sub-step S10222: When it is determined that the RF board has a fault, based on a binary classification model of the fault type, according to the analysis data of the thermal imaging image of the faulty RF board and the operating status data, determine whether the RF board has an unavailable fault or an available fault.

[0097] The two binary classification models described above can be used to determine whether the RF board is operating normally or has a fault. If a fault exists, it can be further distinguished as an available fault or an unavailable fault.

[0098] It should be noted that the above method of the embodiment of the present application can be applied to the daily operation and maintenance of indoor signals in the communications industry and to alarm processing application scenarios of other key facilities.

[0099] Refer to Figure 9, which is a structural diagram of an embodiment of the network management device of the present application. It should be noted that the network management device of the embodiment of the present application can implement the above method. For detailed description of the relevant content, please refer to the above method part, which will not be repeated here.

[0100] The network management device 100 includes a communication module 3, a memory 1, and a processor 2. The communication module 3 is used for communication, and the memory 1 is used to store computer programs. The processor 2 is used to execute the computer programs and implement any of the above-described methods for automatically monitoring a radio frequency board when executing the computer programs. The communication module 3 and the memory 1 are connected to the processor 2 via a bus.

[0101] The processor 2 may be a microcontroller unit, a central processing unit or a digital signal processor, etc. The memory 1 may be a Flash chip, a read-only memory, a magnetic disk, an optical disk, a USB flash drive or a mobile hard disk, etc.

[0102] The present application also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the processor implements any of the above-described methods for automatically monitoring a radio frequency board.

[0103] The computer-readable storage medium may be an internal storage unit of the network management device, such as a hard disk or memory. The computer-readable storage medium may also be an external storage device of the network management device, such as a plug-in hard disk, smart memory card, secure digital card, flash memory card, etc.

[0104] The following describes in detail the above methods of the embodiments of the present application by taking the RRU board as an example and combining them together as a specific embodiment.

[0105] The core point of the embodiment of the present application is simply that: a convolutional neural network analyzes the thermal imaging image taken by the infrared imager and obtains the abnormal points (i.e., abnormal components) of the RRU board-related hardware (i.e., multiple components), and combines the operating status data of multiple components of the RRU board to determine whether the RRU board has a fault, as well as the impact range of the current fault on the indoor radio frequency signal when the RRU board fails, and constructs a visualization page of the signal coverage map, and finally triggers related operations to automatically repair the current indoor radio frequency signal.

[0106] The method of the embodiment of the present application can improve the efficiency of locating single-board hardware faults, reduce the impact of faults on users, improve the service quality of indoor coverage, reduce the fault handling pressure of operation and maintenance personnel, and can be applied to other radio frequency boards.

[0107] The method of the embodiment of the present application generally includes the following process:

[0108] 1. Data Acquisition: The infrared imager built into the RRU board is used to perform real-time infrared monitoring of the RRU board and generate a thermal image at the corresponding moment, as shown in Figure 10 (Note: The original image is a color image. Components with normal temperatures generally appear red in the color image, and appear darker in the gray image in Figure 10, such as the components on the left, right, and top of Figure 10. Components with abnormally high temperatures generally appear bright orange in the color image, and appear brighter in the gray image in Figure 10, such as the component on the lower right of Figure 10). Real-time thermal images of the current RRU board can be acquired through real-time monitoring and monitoring triggered by faults. The data acquisition module built into the RRU board is used to acquire the current operating status data of multiple components on the RRU board and send the operating status data and thermal images to the network management (i.e., network management equipment).

[0109] 2. Hardware Analysis: Use a convolutional neural network to analyze thermal images within the current time period and mark the locations of abnormally high temperature areas on the RRU board. Based on this, fault weights are set and, combined with the special areas marked by the model (such as sensors, chips, capacitors, resistors, and other related areas), the main areas of abnormally high temperature are calculated to obtain the fault feature vector of the RRU board.

[0110] 3. Fault Diagnosis: First, train two binary classification models. One is a fault classification model, used to diagnose whether the RRU board is operating normally or has a fault. The other is a fault type classification model, used to diagnose whether the RRU board fault is an available fault or an unavailable fault. Using the KNN (K-Nearest Neighbor) or SVM (Support Vector Machine) classification algorithm, calculate the value by combining the original operating status data with the fault feature vector. Then, use the sigmoid function to convert the value into a categorical value between 0 and 1.

[0111] For the binary fault classification model, multiple receiver operating characteristic (ROC) curves are formed based on the classification value and the probability of a fault occurring in the sample at that classification value. This means that multiple binary fault classification models are trained. The model with the area under the ROC curve (AUC) closest to 1 is selected as the actual binary fault classification model. If the classification value is greater than x1, the probability of a fault is greater than y1%, where y1 is the required probability. In this case, x1 is used as the threshold for determining a fault. The resulting data is then fed into the binary fault classification model, and the relative magnitude of the classification value and x1 is used to determine whether the RRU board is faulty.

[0112] For the fault type binary classification model, the classification value and the probability of unavailability in faulty samples at that classification value form multiple ROC curves. This means that multiple fault type binary classification models are trained, and the model with the AUC closest to 1 is selected as the actual fault type binary classification model. If the classification value is greater than x2, the unavailability probability is greater than y2%, and y2 is the required probability. In this case, x2 is used as the threshold for determining an unavailability fault. The resulting data is fed into the fault type binary classification model, and the magnitude of the classification value and x2 is used to determine whether the RRU board has an unavailability fault.

[0113] For example, as shown in Figure 11, the horizontal axis represents specificity (also known as the true negative rate), and the vertical axis represents sensitivity (also known as the true positive rate). When the binary classification model is a fault classification model, the sensitivity is the proportion of faulty boards in all samples. When the binary classification model is a fault type classification model, the sensitivity is the proportion of boards with unavailable faults in all faulty samples. After training, three ROC curves are formed: Curve A, Curve B, and Curve C. Curve A has the closest AUC to 1, so Curve A is selected as the binary classification model.

[0114] 4. Real-time detection of indoor RF signals: Utilizing a long short-term memory network trained with operating status data and fault feature vectors, the current operating status data and fault feature vectors are input to calculate the RF signal coverage of the current RRU board. Based on the location of the RRU board and the coverage of the RRU board's RF signal, the total coverage of the indoor RF signal and the corresponding signal coverage map are formed. The total coverage area of ​​the indoor RF signal is calculated using a grid-based approximation algorithm, and a minimum coverage area threshold is set. When the total coverage area of ​​the indoor RF signal falls below the minimum coverage area threshold, a low signal alarm is issued, and signal self-healing, i.e., automatic adjustment, is triggered.

[0115] 5. Signal self-healing: Using the long short-term memory network from the previous step, the model calculates the current RRU board's RF signal coverage and power adjustment range. The model then calculates the adjustable signal range of the RRU board. A greedy algorithm is combined with a grid-based approximation algorithm to calculate different solutions that can restore the current signal coverage area. The solution with the smallest ratio of increased power consumption to increased area is selected as the self-healing solution.

[0116] 6. Fault recovery: When the RRU board-related fault is resolved, the power amplifier module with increased power is automatically adjusted to restore to its original power to ensure its service life, and the power changes are dynamically adjusted during daily use.

[0117] The application environment of an embodiment of the present application is: it involves an indoor radio frequency signal coverage alarm management, a system that monitors the current indoor 5G-related signal coverage and the operating status of related equipment. After the operation and maintenance personnel monitor the alarm, they process the alarm and can detect the alarm in real time.

[0118] The application of the embodiments of the present application is to perform self-healing repair of indoor RF signals. Through relevant resource configuration, real-time hardware detection of RRU boards is performed, relevant hardware information is automatically reported and alarms are issued, faulty RRUs are marked, and changes in relevant indoor signals are dynamically adjusted to achieve the purpose of automatic repair of indoor scene signals. By adjusting relevant parameters of the deep learning model, fault adjustment for most RF boards and self-healing of RF signals are also achieved.

[0119] The structure diagram of the radio frequency board and the network management device according to an embodiment of the present application is shown in FIG12 :

[0120] The components described in the embodiment of the present application are divided into four modules, namely, a single board acquisition module 121, a hardware diagnosis module 122, a real-time signal detection module 123, and a dynamic adjustment module 124. The components of each component are described as follows:

[0121] Single board acquisition module 121: It integrates a small infrared imager for real-time acquisition of single board infrared thermal imaging images and a data acquisition module for acquiring operating status data of multiple components of the RRU single board (also known as hardware operating parameters, such as power, signal strength, frequency, etc.).

[0122] Hardware diagnosis module 122: uses a convolutional neural network using the thermal image and the identifier of the area where the relevant components (i.e., electronic devices) are located as feature values ​​to calculate the abnormal numerical features appearing in the key positions of the corresponding components, and outputs the fault feature vector of the single board. Then, based on the operating status data of multiple components such as the power, bandwidth, frequency, and link connection status of the current single board, it further accurately confirms whether the current RRU single board has related hardware faults, constructs a sigmoid function to determine the extent of the impact of this fault, and implements the ROC curve to determine whether the fitting degree of the model meets the required level, providing a basis for subsequent real-time signal detection and dynamic adjustment.

[0123] Real-time signal detection module 123: Utilizes a long short-term memory network to calculate the range of the current RRU board's RF signal coverage based on RF signal-related data (e.g., bandwidth, signal strength, power amplifier power, noise level, etc.), and forms a related signal color block diagram (i.e., the corresponding signal coverage diagram). Based on the specific conditions of the indoor environment, a threshold that can reserve some space gaps can be used to detect the current RF signal in real time.

[0124] Dynamic Adjustment Module 124: Utilizing a model formed by a long-short-term memory network, combined with a resource allocation algorithm, dynamically calculates the power increase required for each RRU board's power amplifier module to ensure that the current indoor RF signal reaches the total coverage area corresponding to the preset minimum coverage threshold. The calculated result is then sent to the corresponding RRU board's power amplifier module for dynamic adjustment. For RRU boards that are unavailable due to a fault, the power of their power amplifier module is set to 0 to reduce power consumption. If the fault is subsequently detected and recovered, the power of the previously adjusted RRU board's power amplifier module is reset to its original level. Automatic adjustment can also be performed during daily use based on actual conditions.

[0125] The method of the embodiment of the present application may include the following steps:

[0126] Step 1: Data acquisition, that is, acquiring hardware fault related information (ie, thermal imaging image and operation status data) from the board acquisition module 121 of the RRU board.

[0127] Step 2: Hardware diagnosis, that is, diagnosing whether there is a related fault from the information related to the hardware fault.

[0128] Step 3: Real-time signal detection, that is, building the current indoor RF signal coverage situation in real time according to the fault situation of the RRU board managed by the current indoor network management equipment, and when a fault is detected, the alarm will be displayed in real time on the page of the network management equipment.

[0129] Step 4: Dynamic adjustment, that is, dynamically adjusting the relevant configurations of the normally operating RRU board according to the current indoor RF signal coverage.

[0130] Step 5: Self-healing adjustment, that is, performing signal self-healing on the current indoor RRU board based on the dynamic adjustment feedback, dynamically adjusting the relevant RRU board until the RF signal meets the minimum requirement. If the minimum requirement cannot be met, the relevant user is notified.

[0131] Step 6: Alarm recovery, that is, after the subsequent fault is restored, the alarm at the network management device is restored, and the power of the power amplifier module of each adjusted RRU board is adjusted back to the initial level.

[0132] Step 1 may correspond to the following specific embodiments:

[0133] To determine whether an RRU board is faulty, it is necessary to collect fault information about the RRU board hardware.

[0134] First, the board acquisition module 121 inside the RRU board is used to obtain the current RRU board related data, and its power interface, optical cable interface, module overall temperature, fan speed, current room temperature and other key related data are called for extraction, and the location and specific related data are marked. It can detect in real time and can also actively call related data in the command line of the network management device page.

[0135] Next, we need to obtain the fault feature vector of the hardware dimension. A miniature infrared imaging device is built into the RRU board, which regularly performs infrared imaging of the RRU board in real time. The thermal images of the corresponding time are filtered and cropped and sent to the network management device for use in data processing in step 2.

[0136] In practice, as long as there is a board that can obtain operating status data and a thermal image can be obtained through infrared imaging, this method can be used to obtain the operating status information of the RRU board for subsequent system processing.

[0137] Step 2 may correspond to the following specific embodiments:

[0138] Before implementing hardware fault analysis, it is necessary to read the features of the acquired thermal images and convert the image features into fault feature vectors for subsequent data processing. The specific method is to use convolutional neural networks to extract features from thermal images.

[0139] Convolutional neural networks use convolution kernels to extract features. The overall model's fit can be controlled by varying the convolution and pooling layers, as well as the size of the final output feature vector. This can reduce the dimension of the fault feature vector in the case of overfitting, and increase the output dimension of the convolution layer in the case of underfitting. Compared to other feature extraction methods, this method offers greater flexibility.

[0140] By convolving the image's pixel matrix with the convolution kernel, we can obtain eigenvalues. The convolution kernel, when placed in a neural network, represents the corresponding weight. The dot product of the convolution kernel and the image represents the weight of the convolution kernel at the corresponding position. First, a convolutional kernel with the original component location is set in the convolutional layer to perform convolution feature extraction on the current board image. This is used to determine the location of the corresponding component in the current thermal image. After processing in the pooling layer, a fault weight is assigned to the corresponding component based on its importance. A convolution kernel dynamically generated based on the fault weight and the corresponding component location is then used to extract features from the current image. An error cost is set in the form of cross-entropy. A corresponding number of images are extracted for training each time. The sig value generated by each iteration is used. If underfitting is detected, the number of iterations is increased. Finally, an X-dimensional feature vector is generated for the fully connected layer. The value of X is affected by the image pixels and the number of kernels in the first convolution kernel. For example, if the initial image has 256x256 pixels and the first convolution kernel (component feature extraction kernel) is a 5x5 matrix, the pooling layer will output a 32-dimensional component feature vector. The resulting fault convolution feature vector is used as the eigenvalue for subsequent machine learning.

[0141] The two key points in this algorithm are the convolution kernel and fault weights of the first-layer convolutional neural network. The convolution kernel determines its matrix value based on the board type, which is used to extract the positions of relevant components on the current board. Therefore, it can be applied to component position extraction on various boards. Secondly, the fault weight is assigned to the relevant fault weight based on the current component position. The importance of components on different boards varies. Therefore, when applied to boards other than RRUs, the fit of the entire model can be adjusted by adjusting the convolution kernel and fault weights, ultimately resulting in a fault feature vector that is also applicable to all boards.

[0142] Using a classification algorithm such as KNN or SVM, we determine whether the current board is faulty. If so, we determine whether the fault is unavailable. This includes two binary classification models: one for fault classification and one for fault type classification.

[0143] The data injected into the model consists of two components: the operating status data of multiple components on the RRU board, with related variables grouped together for calculation; and the fault feature vector extracted from the thermal image analysis above. This classification algorithm generates a corresponding classification value.

[0144] The calculated classification values ​​are often quite broad. Since classification results often exhibit a clear normal distribution, a special function is needed to compress the range of the classification values ​​to between 0 and 1. This requires the use of the sigmoid function. The sigmoid function compresses all classification values ​​to between 0 and 1. This function requires setting a threshold for the relevant positive value (for example, what value between 0 and 1 is considered an unavailable fault).

[0145] Taking the fault type binary classification model as an example, this threshold is used to determine whether a fault is unavailable. To determine the threshold, we use the 0-1 range of values ​​formed by combining all current variables to determine the probability that the current value is an unavailable fault. This probability is then plotted as an ROC curve. Based on the curve, we determine the threshold so that the accuracy rate of positive judgments meets the expected threshold. The AUC value of the ROC curve is used to determine whether the current classification algorithm meets the requirements. The closer the AUC value is to 1, the more accurate the prediction algorithm.

[0146] The detected data is placed in the current model to calculate the actual operating status of the current RRU board, and the operating status is passed to step 3 for subsequent calculation of the actual signal coverage range.

[0147] Step 3 may correspond to the following specific embodiments:

[0148] Observing RF signal changes at each stage is generally not done from the beginning. Therefore, current signal coverage needs to be assessed based on the actual conditions at each stage. Recurrent neural networks can address this problem. These loops allow for persistent information. Using long- and short-term neural networks, relevant cell states are constructed and selectively incorporated into these cell states by incorporating operational status data and fault feature vectors extracted from images.

[0149] When the signal strength of the RF signal of a single board exceeds a certain distance, the signal can hardly realize normal business functions. Therefore, the range beyond this distance is regarded as a no-signal state. Therefore, the actual signal coverage range of the RF signal of the RRU single board can be approximately regarded as conforming to the disk Boolean model, which facilitates subsequent signal coverage-related calculations.

[0150] Based on the current t value and input data, the system determines the hidden state of the current RRU board's RF signal. Based on this state, it outputs the signal coverage range to the next state. Using the recurrent neural network, it can predict signal changes over time by acquiring operational status data and image fault feature vectors. For RRU boards identified as unavailable, signal statistics are not collected and the system reports these failures to the network management device. A signal coverage map is constructed based on the signal location over this period, retaining coverage in some blind spots (as shown in Figure 13). This allows for a visual prediction and display of indoor signal coverage and strength over time.

[0151] In the actual case of wireless signal propagation, wireless signals are very unstable and irregular. Signal irregularity is a common problem of all signal coverage and cannot be ignored. In particular, the radio frequency signals transmitted by the RRU unit are often affected by the current environment, the current radio environment interference, etc. Therefore, the range of signal propagation often changes irregularly and packets are lost in different directions. To address this situation, it is necessary to set an interference feature among the incoming related features. This feature is the average impact of the current environment due to signal irregularity. In this way, the ideal signal range can be reduced by the comprehensive impact. In this way, the relevant predicted range value can be obtained more accurately, and a more accurate signal coverage map under relevant interference conditions can be more intuitively displayed.

[0152] There is a time lag between when a fault occurs and when the user perceives a signal change. Therefore, the current signal situation can be used to predict future signal changes. This time lag can be used to implement self-healing measures when a signal change occurs, minimizing the impact of the current hardware fault on users.

[0153] Step 4 and step 5 may correspond to the following specific embodiments:

[0154] Because the signal coverage of the RRU board approximately conforms to the disk Boolean model, as shown in Figure 14, a grid-based approximation algorithm is used. This algorithm divides the entire area into a large number of grids. The distance from each grid to the center of the circle is calculated, and the total coverage area is finally derived by summing up all the covered grids. Currently, the main solution used is a diamond grid. By deploying circles at each vertex of the diamond grid, the area is completely seamlessly covered. This analytical expression for the minimum number of circles required for complete seamless coverage can be realized.

[0155] Then, using the correlation model obtained in step 3, the RRU boards that are still able to operate are calculated. The power of the input power amplifier module is within a certain range of thresholds. By combining the greedy algorithm with the grid-based approximation algorithm, the correlation power of the nearest RRU board is first increased, and the relevant adjustment input value is calculated as the actual range value of each RRU board. When the coverage area is increased, the ratio of the increased power to the increase in coverage area is calculated. By adjusting the order of increasing the power, different improvement schemes are obtained. Finally, the scheme with the lowest ratio of increased power to increased area is output when the coverage range reaches the lowest range. The actual signal range value of each RRU board is calculated, and the power size that needs to be adjusted for the corresponding RRU board is obtained.

[0156] The relevant numerical adjustments are sent to the corresponding RRU board for real-time adjustment. The coverage of the relevant signals is promptly updated based on the feedback data. The signal change adjustment sets the relevant time frequency to ensure that relatively accurate signal change results can be obtained within the signal adjustment cycle, and also reduce the risk of failure caused by repeated adjustment of the power amplifier module of the RRU board. By achieving the lowest ratio of output power increase to area increase, the current indoor signal problem is restored with the lowest energy consumption, achieving energy conservation and consumption reduction goals while self-healing the fault.

[0157] Step 6 may correspond to the following specific embodiments:

[0158] When the fault has been handled by the operation and maintenance personnel and has been determined to have disappeared, the alarm at the network management equipment is restored based on the currently restored signal coverage, and the power of the power amplifier module of each adjusted single board is adjusted back to the initial level. After the processing is completed, the relevant data of this fault handling is uploaded to the sample library, and sufficient data is collected to further train the model and improve the accuracy and versatility of the overall model prediction.

[0159] In the absence of any faults, the system monitors the current signal changes in real time and adjusts the power of the power amplifier module of the current RRU board in real time to achieve the relevant coverage range. This can not only ensure the user's indoor signal experience in the absence of any faults, but also save the operation and maintenance costs of the operation and maintenance personnel and reduce the manpower consumption caused by signal weakening in the absence of any faults.

[0160] In summary, the embodiment of the present application relies on the operating status data and thermal imaging images of the actual RRU board. The sample source is sufficient and sufficient. As time goes by and the amount of data continues to increase, combined with the infrared imaging of the actual board and the location of the faulty component, the convolutional neural network model can be continuously trained, so that the fault weight of the component failure within the corresponding area can be determined, and the fault weight, operating status data and actual fault cause of the corresponding component are continuously trained on the deep learning model to adjust the proportion of each characteristic value to the RRU board hardware failure, the operating status data (voltage, link status, bit error, power, board temperature, etc.), and the hardware failure (component fault location, component fault weight) on the current alarm impact are comprehensively evaluated to judge the current working status of the RRU board, and use deep learning to construct the current signal coverage map, and dynamically adjust the uncovered area to achieve signal self-healing adjustment.

[0161] By adopting the technical solution of the embodiment of the present application, the following technical effects can be achieved:

[0162] 1. In response to the problem that single-board diagnostic data in related technologies cannot provide comprehensive diagnosis, the embodiments of the present application provide diagnostic data in the hardware dimension (i.e., thermal imaging images), and provide operating status data of multiple components of a single board, and judge the degree of impact of the fault, so as to accurately judge the impact of the current hardware fault on the existing basis, and use the results as a basis for subsequent fault handling.

[0163] 2. Build a signal coverage map to visualize the current indoor signal status and dynamically adjust the power of the power amplifier module in areas with insufficient signal until the signal reaches the required level. This further provides O&M personnel with hardware fault support and emergency self-healing when faults occur. Automated processing also facilitates application in larger indoor environments. Furthermore, this system can also be applied to all RF boards with relevant acquisition modules (collecting operating status data and infrared imaging). By adjusting the relevant parameters of the deep learning model, fault adjustment and RF signal self-healing can be achieved for RF boards capable of collecting operating status data and infrared imaging.

[0164] Those skilled in the art will appreciate that all or some of the steps in the methods, systems, and functional modules / units in the devices disclosed above may be implemented as software, firmware, hardware, or appropriate combinations thereof.

[0165] In a hardware implementation, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or temporary medium). As known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer. Furthermore, as is well known to those skilled in the art, communication media typically embodies computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0166] The preferred embodiments of the present application are described above with reference to the accompanying drawings, but are not intended to limit the scope of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and essence of the present application shall fall within the scope of the present application.

Claims

1. A method for automatically monitoring a radio frequency board, the method being applied to a network management device, the method comprising: Obtaining a thermal image of the indoor radio frequency board and current operating status data of multiple components of the radio frequency board; Determine the current operation status of the indoor radio frequency board according to the thermal imaging image and the operation status data, where the current operation status of the radio frequency board includes normal operation or failure; According to the current operation status of the indoor radio frequency board, the total coverage range of the indoor radio frequency signal is controlled to at least reach the total coverage range corresponding to the preset minimum coverage threshold.

2. The method according to claim 1, wherein: According to the current operation status of the indoor radio frequency board, controlling the total coverage range of the indoor radio frequency signal to at least reach the total coverage range corresponding to the preset minimum coverage threshold includes: determining a total coverage range of an indoor radio frequency signal according to radio frequency signal related data of an indoor radio frequency board, wherein the radio frequency signal related data belongs to the data in the operation status data; According to the total coverage range of the indoor radio frequency signal, the total coverage range of the indoor radio frequency signal is controlled to at least reach the total coverage range corresponding to the preset minimum coverage threshold.

3. The method according to claim 2, wherein: According to the total coverage range of the indoor radio frequency signal, controlling the total coverage range of the indoor radio frequency signal to at least reach the total coverage range corresponding to the preset minimum coverage threshold includes: When the current total coverage range of the indoor radio frequency signal is less than the total coverage range corresponding to the preset minimum coverage threshold, the radio frequency board operating normally indoors is controlled to perform power adjustment so that the adjusted total coverage range of the indoor radio frequency signal is equal to the total coverage range corresponding to the preset minimum coverage threshold.

4. The method according to claim 3, wherein: The controlling the normal operation of the indoor radio frequency board to perform power regulation includes: The resource allocation algorithm is used to determine the radio frequency to be adjusted in the indoor normally operating radio frequency board. frequency board and the target power value to be adjusted; The power of the power amplifier module of the radio frequency board to be adjusted in the radio frequency board normally operating indoors is dynamically controlled based on the target power value.

5. The method according to claim 3, wherein: The method further comprises: When it is determined that the faulty RF board has resumed normal operation and the total coverage range of the indoor RF signal is greater than the total coverage range corresponding to the preset minimum coverage threshold, the adjusted RF board among the indoor normally operating RF boards is controlled to resume the power to the power before adjustment.

6. The method according to claim 2, wherein: The determining of the total coverage of the indoor radio frequency signal according to the radio frequency signal related data of the indoor radio frequency board includes: Determine the coverage of the radio frequency signal of each radio frequency board in the room according to the radio frequency signal related data of the radio frequency board in the room; According to the coverage of the radio frequency signals of multiple radio frequency boards in the room and the indoor position, the current total coverage of the indoor radio frequency signal is determined.

7. The method according to claim 6, wherein: The method of determining the coverage range of the radio frequency signal of each current radio frequency board in the room according to the radio frequency signal related data of the current radio frequency board in the room includes: Based on the recurrent neural network, the coverage range of the RF signal of each current RF board in the room is determined according to the RF signal related data of the current RF board in the room.

8. The method according to claim 6, wherein: The determining of the current total coverage of the indoor radio frequency signal according to the coverage of the radio frequency signals of the multiple radio frequency boards in the room and the indoor position includes: Based on the grid approximation algorithm, according to the coverage range of the radio frequency signals of multiple radio frequency boards in the room and the indoor position, the current total coverage area of ​​the indoor radio frequency signal is determined, and the current signal coverage map of the indoor radio frequency signal is constructed.

9. The method according to claim 1, wherein: The method further comprises: When it is determined that the radio frequency board has an unavailable fault, the radio frequency board with the unavailable fault is controlled. The power of the power amplifier module of the single board is set to zero, and the fault includes the unavailable fault and the available fault; and / or, When it is determined that the radio frequency board has a fault, a prompt message is issued, wherein the prompt message is used to notify the user to repair the radio frequency board with the fault.

10. The method according to claim 1, wherein: Determining the current operation status of the indoor radio frequency board according to the thermal imaging image and the operation status data includes: Analyzing the thermal imaging image using a convolutional neural network to obtain analysis data of the thermal imaging image; The current operating status of each of the radio frequency boards in the room is determined according to the analysis data of the thermal imaging image and the operating status data.

11. The method according to claim 10, wherein: The method of analyzing the thermal imaging image using a convolutional neural network to obtain analysis data of the thermal imaging image includes: Based on the convolution kernel of the first convolution layer, first data of the thermal imaging image is obtained according to the thermal imaging image, and the first data of the thermal imaging image is used to determine the positions of the multiple components of the radio frequency board on the thermal imaging image; Determine the fault weights of the plurality of components of the RF board according to the first data of the thermal imaging image; Obtaining second data of the thermal imaging image according to the first data of the thermal imaging image and the fault weights of the plurality of components of the radio frequency board; Based on the convolution kernel of the second convolution layer, a fault feature vector of the radio frequency board is obtained according to the second data of the thermal imaging image.

12. A network management device, comprising: A communication module, a memory and a processor, wherein the communication module is used for communication and the memory is used for storing computer programs; The processor is used to execute the computer program and implement the automatic monitoring method of the radio frequency board as described in any one of claims 1-11 when executing the computer program.

13. A computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor implements the automatic monitoring method of a radio frequency board according to any one of claims 1 to 11.

Images