Indoor distribution system monitoring method and device and electronic equipment

Abstract

The invention discloses a monitoring method and device of an indoor distribution system and electronic equipment. The method comprises the following steps: acquiring standing-wave ratio data of an indoor distribution system; first indoor distribution systems are determined according to the standing-wave ratio data, and the first indoor distribution systems are part of the indoor distribution systems in the abnormal state; user feature data of the first indoor distribution system and a second indoor distribution system are determined, the residual network is trained according to the user feature data, an indoor distribution monitoring model is obtained, and the second indoor distribution system is the indoor distribution system in a normal state in the indoor distribution systems; and monitoring the indoor distribution system through the indoor distribution monitoring model to obtain a monitoring result of the indoor distribution system. According to the invention, the technical problem that the network maintenance efficiency and the user experience are affected due to the fact that the indoor distribution monitoring technology in the related technology is difficult to effectively identify the signal quality reduction caused by the small change of the passive device or the loose connection, and the false alarm or the missing alarm is easily generated especially in the environment of obvious user tidal effect and large index fluctuation is solved.

Description

Technical Field

[0001] The present application relates to the field of communication technologies, and in particular, to a method, device, and electronic device for monitoring an in-building distribution system. Background Art

[0002] In modern wireless communication networks, in-building distribution systems play a crucial role. Especially in scenarios with high pedestrian flow and complex building structures such as hospitals, large shopping malls, and office buildings, in-building distribution systems are the main means for operators to solve the problem of insufficient indoor deep coverage. However, a large number of passive devices in in-building distribution systems, such as couplers, power dividers, and feeders, once the connections become loose, the devices are physically damaged, or their performance deteriorates, it will lead to a significant deterioration in signal transmission quality, thereby affecting the quality of indoor mobile communication services. Such problems are often only discovered after users complain about the degradation of network quality, which not only delays the time for fault location and repair but also directly affects the user experience and the network service reputation of the operator.

[0003] In related technologies, in-building monitoring technologies usually rely on monitoring changes in user traffic, quantity, and user performance indicators to identify potential system anomalies. However, in practical applications, these methods have significant limitations. First, the large fluctuations in indicators caused by the user tidal effect make it difficult to accurately determine whether there is a real fault in the in-building distribution system based on changes in user performance indicators. Second, the existence of blind spots in the coverage of in-building distribution systems can lead to misjudgments in the telecommunication network. Even if the system is normal, false reports of network quality degradation may be generated due to ineffective coverage in some areas. In addition, faults in passive devices usually do not trigger any direct alarm signals, which further increases the difficulty of monitoring in-building distribution systems.

[0004] To address the above problems, no effective solutions have been proposed yet. Summary of the Invention

[0005] Embodiments of the present application provide a method, device, and electronic device for monitoring an in-building distribution system, so as to at least solve the technical problem that in-building monitoring technologies in related technologies are difficult to effectively identify signal quality degradation caused by minor changes or loose connections of passive devices, especially prone to false alarms or missed alarms in an environment with a significant user tidal effect and large fluctuations in indicators, affecting network maintenance efficiency and user experience.

[0006] According to one aspect of the embodiments of the present application, a method for monitoring a distributed antenna system is provided, including: obtaining the standing wave ratio data of the distributed antenna system; determining a first distributed antenna system based on the standing wave ratio data, where the first distributed antenna system is a part of the distributed antenna system in an abnormal state; determining the user characteristic data of the first distributed antenna system and a second distributed antenna system, and training a residual network based on the user characteristic data to obtain a distributed antenna monitoring model, where the second distributed antenna system is a distributed antenna system in a normal state; monitoring the distributed antenna system through the distributed antenna monitoring model to obtain a monitoring result of the distributed antenna system, where the monitoring result is used to reflect the abnormal state of the distributed antenna system.

[0007] Optionally, determining the first distributed antenna system based on the standing wave ratio data includes: determining the true fluctuation amplitude, trend value, trend index, and trend strength of the standing wave ratio of the distributed antenna system based on the standing wave ratio data; determining the average trend index of the distributed antenna system based on the true fluctuation amplitude, trend value, trend index, and trend strength of the standing wave ratio; determining the first distributed antenna system based on the average trend index and a first preset threshold.

[0008] Optionally, determining the user characteristic data of the first distributed antenna system and the second distributed antenna system includes: determining target users in the first distributed antenna system and the second distributed antenna system, where the target users are used to represent users who can reflect the overall performance of the distributed antenna system within a preset time period; obtaining the behavioral characteristic data of the target users, where the behavioral characteristic data includes at least one of the following: the reference signal received power, signal-to-noise ratio, and power headroom report of the target users; converting the behavioral characteristic data into a feature matrix form to obtain the user characteristic data.

[0009] Optionally, determining the target users in the first distributed antenna system and the second distributed antenna system includes: determining the abnormal data in the first distributed antenna system and the second distributed antenna system; removing the abnormal data and determining the single-day call duration of all users in the first distributed antenna system and the second distributed antenna system, where the single-day call duration is used to represent the time length of a user's effective communication with the network within a day; determining the users with a single-day call duration lower than a second preset threshold as the target users.

[0010] Optionally, training the residual network based on the user characteristic data includes: preliminarily processing the user characteristic data through the convolutional layer in the residual network to obtain a first feature map; performing deep feature learning on the first feature map through multiple residual blocks in the residual network to obtain a second feature map; performing fusion processing on the second feature map through the fully connected layer in the residual network to obtain a prediction result corresponding to the user characteristic data, where the prediction result includes a first classification label for judging the abnormal state of the distributed antenna system.

[0011] Optionally, the method further includes: determining a difference value between a first classification label and a second classification label through a target loss function, where the second classification label is a true label manually marked to represent the state of the in-building distribution system; adjusting model parameters of the residual network according to the difference value to obtain an in-building distribution monitoring model.

[0012] Optionally, the method further includes: determining true positives, false positives, true negatives, and false negatives of the in-building distribution system, where true positives are used to represent the number of in-building distribution systems with abnormal states correctly identified by the in-building distribution monitoring model; false positives are used to represent the number of in-building distribution systems with normal states misidentified as abnormal by the in-building distribution monitoring model; true negatives are used to represent the number of in-building distribution systems with normal states correctly identified by the in-building distribution monitoring model; false negatives are used to represent the number of in-building distribution systems with abnormal states misidentified as normal by the in-building distribution monitoring model; determining the block accuracy of the user feature data according to the true positives, false positives, true negatives, and false negatives, where the block accuracy is used to represent the classification accuracy of each feature matrix in the user feature data.

[0013] Optionally, the method further includes: determining the station accuracy of the in-building distribution system according to the block accuracy, where the station accuracy is used to represent the overall classification accuracy of the in-building distribution system.

[0014] According to another aspect of the embodiments of the present application, there is also provided a monitoring device for an in-building distribution system, including: an acquisition module, configured to acquire standing wave ratio data of the in-building distribution system; a determination module, configured to determine a first in-building distribution system according to the standing wave ratio data, where the first in-building distribution system is a part of the in-building distribution system in an abnormal state; a training module, configured to determine user feature data of the first in-building distribution system and a second in-building distribution system, and train a residual network according to the user feature data to obtain an in-building distribution monitoring model, where the second in-building distribution system is an in-building distribution system in a normal state; a monitoring module, configured to monitor the in-building distribution system through the in-building distribution monitoring model to obtain a monitoring result of the in-building distribution system, where the monitoring result is used to reflect the abnormal state of the in-building distribution system.

[0015] According to still another aspect of the embodiments of the present application, there is also provided an electronic device, including: a memory and a processor, where the memory is configured to store program instructions; the processor is connected to the memory and configured to execute to implement the above-mentioned monitoring method for the in-building distribution system.

[0016] According to yet another aspect of the embodiments of the present application, there is also provided a non-volatile storage medium, where the non-volatile storage medium includes a stored computer program, and the device where the non-volatile storage medium is located executes the above-mentioned monitoring method for the in-building distribution system by running the computer program.

[0017] According to another aspect of the embodiments of the present application, there is also provided a computer program product, including computer instructions, which implement the above-mentioned monitoring method of the in-building distribution system when executed by a processor.

[0018] In the embodiments of the present application, by obtaining the standing wave ratio data of the in-building distribution system; determining the first in-building distribution system according to the standing wave ratio data, where the first in-building distribution system is a part of the in-building distribution system in an abnormal state; determining the user characteristic data of the first in-building distribution system and the second in-building distribution system, and training the residual network according to the user characteristic data to obtain an in-building distribution monitoring model, where the second in-building distribution system is the in-building distribution system in a normal state; monitoring the in-building distribution system through the in-building distribution monitoring model to obtain the monitoring result of the in-building distribution system, where the monitoring result is used to reflect the abnormal state of the in-building distribution system, achieving the purpose of accurately identifying and locating the obvious and hidden obstacles in the in-building distribution system, thereby realizing early warning and rapid response, and effectively improving the technical effect of the indoor mobile network service quality. Furthermore, it solves the technical problem that the in-building distribution monitoring technology in the related art is difficult to effectively identify the signal quality degradation caused by small changes or loose connections of passive devices, especially prone to false alarms or missed alarms in an environment with significant user tidal effects and large index fluctuations, affecting the network maintenance efficiency and user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings described herein are used to provide a further understanding of the present application, and constitute a part of the present application. The schematic embodiments of the present application and their descriptions are used to explain the present application, and do not constitute an improper limitation to the present application. In the drawings:

[0020] Figure 1 is a hardware structure diagram of a computer terminal for implementing the monitoring method of the in-building distribution system according to the embodiments of the present application;

[0021] Figure 2 is a flowchart of a monitoring method of the in-building distribution system according to the embodiments of the present application;

[0022] Figure 3 is a schematic structural diagram of a residual block according to the embodiments of the present application;

[0023] Figure 4 is a schematic diagram of a loss value change curve according to the embodiments of the present application;

[0024] Figure 5 is a schematic diagram of a block accuracy change curve according to the embodiments of the present application;

[0025] Figure 6 is a schematic diagram of a station accuracy according to the embodiments of the present application;

[0026] Figure 7is a confusion matrix diagram of a station accuracy according to an embodiment of the present application;

[0027] Figure 8 It is a structural diagram of a monitoring device for a room distribution system according to an embodiment of the present application. DETAILED DESCRIPTION

[0028] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0029] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0030] First, some nouns or terms that appear in the process of explaining the embodiments of the present application are subject to the following explanations:

[0031] Indoor distribution system: A system used to improve the coverage and quality of wireless signals inside buildings. It distributes the signal source to every corner of the room through a distributed antenna network. It is especially used to solve signal coverage problems in large buildings or underground spaces.

[0032] Passive components: In indoor distributed systems, passive components refer to components that do not require external power to work, such as couplers, power dividers, combiners, feeders, connectors, etc. They are used for signal distribution and transmission.

[0033] Residual Network (ResNet): A neural network structure in deep learning. By introducing residual blocks, it solves the problems of gradient vanishing and gradient explosion in deep networks, enabling the network to train deeper layers without significantly degrading performance. Residual blocks allow the network to learn the residual function, i.e., the difference between the input and output, rather than directly learning the mapping from input to output, which makes the training of deep residual networks more effective.

[0034] VSWR (Voltage Standing Wave Ratio): A key parameter for measuring the impedance matching degree between an antenna and a transmission line (feeder) in a wireless communication system. In the field of wireless communication, especially in mobile communication networks, VSWR is often used to evaluate the health status and signal transmission efficiency of in-building distribution systems.

[0035] RSRP (Reference Signal Received Power): A key technical indicator in LTE (Long-Term Evolution) and NR (New Radio, 5G New Radio) networks, used to measure the power level of the reference signal received by a user equipment. In the monitoring of in-building distribution systems, a continuous decrease in the RSRP value may indicate problems with the connection between the signal source and the receiver, or changes in the indoor environment that affect signal propagation. Therefore, it is an important parameter for detecting the health status of in-building distribution systems.

[0036] SINR (Signal-to-Interference plus Noise Ratio): A commonly used indicator for measuring signal quality in wireless communication, representing the ratio of signal power to the sum of interference and noise power. The higher the SINR, the better the signal quality, and the higher the stability and reliability of communication. When monitoring in-building distribution systems, changes in SINR can reveal the presence of external interference or a decline in antenna performance, helping to diagnose potential faults in the system.

[0037] PHR (Power Headroom Report): Part of the uplink in LTE and NR networks, mainly used to report the power headroom of a user equipment in uplink transmission. PHR provides the difference between the maximum power that a user equipment can use under current conditions and the actual power used. In in-building distribution systems, monitoring PHR can help evaluate whether the power of the UE is limited and whether the in-building distribution system can effectively support the uplink transmission of the user equipment, indirectly reflecting the performance and stability of the in-building distribution system.

[0038] To solve the problem of poor efficiency in in-building distribution monitoring in related technologies, an embodiment of the present application provides a method for monitoring an in-building distribution system, and this method can run on Figure 1 the computer terminal shown below. The following describes this computer terminal.

[0039] The method embodiment for monitoring the in-building distribution system provided by the embodiment of the present application can be executed on a mobile terminal, a computer terminal or a similar computing device. Figure 1 The following shows a hardware structure block diagram of a computer terminal for implementing the method for monitoring the in-building distribution system. As Figure 1 shown, the computer terminal 10 may include one or more processors (the processors may include, but are not limited to, processing devices such as a microprocessor MCU or a programmable logic device FPGA, shown as 102a, 102b,..., 102n in the figure), a memory 104 for storing data, and a transmission module 106 for communication functions connected by a wired and / or wireless network. In addition, it may further include: a display, a keyboard, a cursor control device, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the I / O interface), a network interface, and a BUS bus. Those of ordinary skill in the art can understand that Figure 1 the structure shown is only schematic and does not limit the structure of the above-mentioned electronic device. For example, the computer terminal 10 may further include more or fewer components than Figure 1 shown, or have a different configuration from Figure 1 shown.

[0040] It should be noted that the above one or more processors and / or other data processing circuits are generally referred to as "data processing circuits" in this article. The data processing circuit may be embodied in software, hardware, firmware or any combination thereof, in whole or in part. In addition, the data processing circuit may be a single independent processing module, or be incorporated in whole or in part into any one of the other elements in the computer terminal 10. As involved in the embodiment of the present application, the data processing circuit is a kind of processor control (such as the selection of a variable resistor terminal path connected to an interface).

[0041] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the monitoring method of the in-building distribution system in the embodiments of the present application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, the monitoring method of the in-building distribution system described above is implemented. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memories. In some instances, the memory 104 may further include a memory remotely disposed relative to the processor, and these remote memories may be connected to the computer terminal 10 through a network. Examples of the above network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0042] The transmission module 106 is used to receive or send data via a network. Specific examples of the above network may include a wireless network provided by a communication provider of the computer terminal 10. In one instance, the transmission module 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station and thus can communicate with the Internet. In one instance, the transmission module 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0043] The display can be, for example, a touch-screen liquid crystal display (LCD), and the liquid crystal display enables a user to interact with the user interface of the computer terminal 10.

[0044] It should be noted here that in some alternative embodiments, the above Figure 1 shown computer terminal may include hardware elements (including circuits), software elements (including computer code stored on a computer-readable medium), or a combination of both hardware elements and software elements. It should be pointed out that Figure 1 is only an example of a specific specific instance, and is intended to show the types of components that may exist in the above computer terminal.

[0045] Under the above operating environment, an embodiment of a monitoring method for an in-building distribution system is provided in the embodiments of the present application. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0046] Figure 2 is a flowchart of a monitoring method for an in-building distribution system according to an embodiment of the present application, as Figure 2As shown in the figure, the method includes the following steps:

[0047] Step S202: Obtain the standing wave ratio data of the in-building distribution system.

[0048] In the above step S202, regular data collection of the in-building distribution system is required. For example, measure the standing wave ratio (VSWR) of the in-building distribution system twice a day. Among them, the standing wave ratio is an important indicator to measure the matching degree of the antenna system. An ideal antenna and feeder should be perfectly matched, so that the VSWR is 1, indicating that all the transmitted energy is absorbed by the antenna without reflection. In the real environment, the VSWR value is usually greater than 1, indicating that part of the energy is reflected back to the feeder, reflecting the efficiency of energy transmission. By obtaining the standing wave ratio data, we can further analyze the physical link signal transmission quality of the in-building distribution system.

[0049] Step S204: Determine the first in-building distribution system according to the standing wave ratio data, where the first in-building distribution system is part of the in-building distribution system in an abnormal state.

[0050] In the above step S204, the improved statistical characteristics of quantitative financial data can be used to evaluate indicators such as the true fluctuation range, trend value, trend index, and trend strength of the standing wave ratio, and finally obtain the average directional index (exponentially decreasing weighted moving average) of the in-building distribution system. Subsequently, by analyzing the average directional index of the in-building distribution system, identify the first in-building distribution system with relatively obvious changes in the wireless link transmission quality, that is, the system with hidden or obvious faults. Through in-depth analysis of the VSWR data, potential problems of the in-building distribution system can be discovered at an early stage without waiting for user complaints or obvious network performance degradation.

[0051] Step S206: Determine the user characteristic data of the first in-building distribution system and the second in-building distribution system, and train the residual network according to the user characteristic data to obtain the in-building distribution monitoring model, where the second in-building distribution system is the in-building distribution system in a normal state.

[0052] In the above step S206, the core task is to construct and optimize the deep learning model for in-building distribution system status monitoring. Specifically, first, select target users (high-quality users) from the identified first in-building distribution system (abnormal system) and the second in-building distribution system (normal system) and process the user characteristic data, including but not limited to indicators such as RSRP, SINR, and PHR, to form a high-quality feature dataset. Subsequently, use the residual network (ResNet) to train these datasets, and extract and learn the user behavior characteristics in normal and abnormal states through deep learning.

[0053] Different from the traditional monitoring approach, this application abandons the consideration of the time continuity of a single user, but focuses on the pattern recognition of different user characteristics, processes the data into a feature matrix, and effectively improves the training efficiency and judgment accuracy of the model. Finally, after multiple rounds of training and optimization, a in-building distribution monitoring model is obtained, which can accurately judge whether the in-building distribution system is in an abnormal state based on the user characteristic data, providing a solid foundation for subsequent intelligent monitoring and fault location.

[0054] Step S208: Monitor the in-building distribution system through the in-building distribution monitoring model to obtain the monitoring result of the in-building distribution system, where the monitoring result is used to reflect the abnormal state of the in-building distribution system.

[0055] In the above step S208, after the model training is completed, the in-building distribution monitoring model is applied to the full-scale in-building distribution system to perform intelligent monitoring on the in-building distribution system in real time or at regular intervals. Among them, the in-building distribution monitoring model can judge the current state of the in-building distribution system according to the pattern of the learned user characteristic data and output the monitoring result, and this detection result reflects whether the system is in an abnormal state.

[0056] Through the above steps S202 to S208, the purpose of accurately identifying and locating the obvious and hidden obstacles in the in-building distribution system is achieved, thereby realizing early warning and rapid response, and effectively improving the technical effect of the indoor mobile network service quality. Furthermore, it solves the technical problem that it is difficult for the in-building distribution monitoring technology in the related art to effectively identify the signal quality degradation caused by small changes or loose connections of passive devices, especially in an environment with a significant user tidal effect and large index fluctuations, which is prone to false alarms or missed alarms, affecting the network maintenance efficiency and user experience. The following is a detailed description.

[0057] In the above step S204, determining the first in-building distribution system according to the standing wave ratio data includes: determining the true fluctuation amplitude, trend value, trend index, and trend strength of the standing wave ratio of the in-building distribution system according to the standing wave ratio data; determining the average trend index of the in-building distribution system according to the true fluctuation amplitude, trend value, trend index, and trend strength of the standing wave ratio; and determining the first in-building distribution system according to the average trend index and the first preset threshold.

[0058] In the embodiment of this application, the non-linear fitting ability of deep learning technology is fully utilized, and the data analysis method in the financial field is introduced into the monitoring of the in-building distribution system, solving problems such as wireless environment differences, user tidal effects, and lack of passive device alarms that are difficult to handle by traditional monitoring methods. The specific process can be as follows:

[0059] First, by analyzing the standing wave ratio data, the real fluctuation amplitude of the standing wave ratio of the indoor system (reflecting the degree of change of the standing wave ratio), the trend value (indicating the rising or falling trend of the standing wave ratio), the trend index (quantifying the strength of the trend) and the trend strength (measurement of the overall trend) are calculated. The calculation of these indicators is based on the improved quantitative financial data statistical method. By comparing the standing wave ratio values ​​at different time points, its short-term and long-term fluctuation patterns are identified. The following is an explanation of the above indicator data:

[0060] (1) The true fluctuation amplitude TR of the standing wave ratio:

[0061] The true fluctuation range of the standing wave ratio TR is an indicator for calculating the degree of change of the standing wave ratio. It measures the fluctuation range of the standing wave ratio by comparing the maximum and minimum values ​​of the standing wave ratio on the day with the average value of the previous day. The larger the TR value, the more drastic the fluctuation of the standing wave ratio, which may mean that the system state is unstable or there is a fault.

[0062] The specific expression is as follows:

[0063] TR=max(V H -V L , V H -V PF , V PF -V L )

[0064] Where V H is the maximum value of the standing wave ratio on that day, V L is the minimum value of the standing wave ratio on that day, V PF is the average value of the standing wave ratio of the previous day.

[0065] (2) Positive and negative trend values ​​(+DM and -DM):

[0066] The trend value is used to describe the rising (+DM) or falling (-DM) trend of the standing wave ratio over a period of time. By comparing the maximum and minimum values ​​of the previous day, it is possible to calculate whether the standing wave ratio of the current day has risen or fallen compared to the previous day, as well as the magnitude of the change. A positive trend value indicates an upward trend, while a negative trend value indicates a downward trend.

[0067] The specific expression is as follows:

[0068]

[0069] Where V PH is the maximum value of the standing wave ratio of the previous day, V PL It is the minimum value of the standing wave ratio of the previous day.

[0070] In order to reduce the impact of fluctuations, it can be processed by a decreasing weighted moving average in the form of an M-day exponential. The specific expression is as follows:

[0071] EX M = [X × 2 + EX M-1 × (M - 1)] / (M + 1)

[0072] Wherein, EX M is the exponential average value on the Mth day, and EX M-1 is the exponential average value on the (M - 1)th day, and X is the index to be processed, such as TR or ±DM.

[0073] (3) Positive and negative directional indicators (+DI and -DI):

[0074] The directional indicator reduces the impact of short-term fluctuations on trend judgment by weighting the trend values. It uses an exponentially decreasing weighted moving average process to calculate the average value of the trend values over multiple consecutive days (such as M days), so as to more accurately reflect the long-term trend of the standing wave ratio. +DI and -DI respectively represent the intensities of the positive and negative directions.

[0075] The specific expressions are as follows:

[0076] +DI = EX +DM / EX TR × 100

[0077] -DI = EX -DM / EX TR × 100

[0078] Wherein, EX +DM represents the exponential average value of +DM, EX -DM represents the exponential average value of -DM, and EX TR represents the exponential average value of TR.

[0079] (4) Trend intensity (DX):

[0080] The trend intensity is measured by the absolute value of the difference between the positive and negative directional indicators. It represents the relative intensity between the rising trend and the falling trend of the standing wave ratio. The larger the DX value, the more obvious the change trend of the standing wave ratio, which may be caused by system anomalies.

[0081] The specific expression is as follows:

[0082] DX = |+DI - (-DI)| / |+DI + (-DI)| × 100

[0083] Secondly, after obtaining the above statistical indicators of the standing wave ratio, the average directional index (ADX) of the in-building distribution system is further calculated. This is a comprehensive index optimized for wireless environment differences. ADX can more accurately reflect the overall state of the in-building distribution system over a period of time, including whether it tends to be abnormal.

[0084] The specific expression is as follows:

[0085] ADX = Exponential Moving Average(DX, M)

[0086] Finally, by comparing ADX with the first preset threshold, those first distributed antenna systems with anomalies can be identified. For example, for ADX N :

[0087] When the average value of ADX 1,…,N-1 ≤ 0.5, if ADX N ≥ 10, then the distributed antenna system is abnormal;

[0088] When the average value of ADX 1,…,N-1 > 0.5, if ADX N ≥ 25, then there is a problem with the distributed antenna system.

[0089] So far, some distributed antenna systems with hidden obstacles can already be monitored. After on-site verification, the accuracy rate of the above-mentioned standing wave ratio monitoring scheme is 99.85%.

[0090] Meanwhile, through the accumulation of data monitored and verified over a certain period of time, combined with the daily inspection and testing of the distributed antenna scenarios, a list of distributed antenna sectors with sufficient positive and negative sample quantities can be obtained.

[0091] In the above step S206, determining the user characteristic data of the first distributed antenna system and the second distributed antenna system includes: determining the target users in the first distributed antenna system and the second distributed antenna system, where the target users are used to represent the users who can reflect the overall performance of the distributed antenna system within a preset time period; obtaining the behavior characteristic data of the target users, where the behavior characteristic data includes at least one of the following: the reference signal receiving power, signal-to-noise ratio, and power margin report of the target users; converting the behavior characteristic data into the form of a feature matrix to obtain the user characteristic data.

[0092] Among them, determining the target users in the first distributed antenna system and the second distributed antenna system includes: determining the abnormal data in the first distributed antenna system and the second distributed antenna system; removing the abnormal data, and determining the single-day call duration of all users in the first distributed antenna system and the second distributed antenna system, where the single-day call duration is used to represent the time length of the effective communication between the user and the network within one day; determining the users with a single-day call duration lower than the second preset threshold as the target users.

[0093] In the embodiment of the present application, the process of determining the target users and their behavior characteristic data in the first distributed antenna system and the second distributed antenna system is an important link for the refined management of the monitoring data of the distributed antenna system. The specific steps can be as follows:

[0094] S1: Determine the target users.

[0095] Target users refer to those users who can reflect the overall performance of the indoor distribution system within a preset time period. Their behavioral characteristic data is more representative, so it is suitable for training and validating deep learning models.

[0096] Specifically, first, to avoid the interference of Internet of Things lag and abnormal users, it is necessary to detect and remove abnormal data from the first indoor distribution system and the second indoor distribution system to ensure the quality and integrity of the dataset. Among them, abnormal data includes but is not limited to incorrect measurement values, extreme values, or invalid values caused by network failures. Then, count the single-day call duration of all users, that is, the length of time that users effectively interact with the network within a day. Finally, set a second preset threshold (such as P hours) to screen target users. Users with a call duration lower than this threshold will be determined as target users. The purpose of this step is to exclude the data of users with extremely short call times or very little usage, because this data may not be sufficient to reflect the comprehensive performance of the indoor distribution system.

[0097] S2: Obtain behavioral characteristic data.

[0098] Behavioral characteristic data mainly includes the reference signal received power (RSRP), signal-to-noise ratio (SINR), and power headroom report (PHR) of target users. These metrics are directly related to the network quality perceived by users and the operating status of the indoor distribution system.

[0099] S3: Data conversion and processing.

[0100] Clean the behavioral characteristic data and convert it into the form of a feature matrix to provide structured data input for the deep learning model. Among them, the feature matrix can effectively organize the behavioral characteristic data of multiple users, facilitating batch processing and feature learning of the neural network. The hourly feature matrix data format of a certain indoor distribution system sector can be as follows:

[0101]

[0102] Among them, DataHour KY represents the performance data of the Kth target user in the Yth hour.

[0103] Through the above process, not only can high-quality user behavioral characteristic data be obtained, but also it can be ensured that each piece of data in the dataset comes from target users who can truly reflect the state of the indoor distribution system. In this way, whether the data is collected in the first indoor distribution system (abnormal state) or the second indoor distribution system (normal state), it can provide valuable input for the training of the neural network model, thereby improving the accuracy and robustness of the model in monitoring the state of the indoor distribution system.

[0104] Further, training the residual network based on user feature data includes: preliminarily processing the user feature data through a convolutional layer in the residual network to obtain a first feature map; performing deep feature learning on the first feature map through multiple residual blocks in the residual network to obtain a second feature map; and performing fusion processing on the second feature map through a fully connected layer in the residual network to obtain a prediction result corresponding to the user feature data, where the prediction result includes a first classification label for determining the abnormal state of the in-building distribution system.

[0105] In the embodiment of the present application, ResNet performs shortcut connections by introducing residual blocks as shown in Figure 3 and does not directly learn complex mapping relationships. Instead, it learns the residuals between the input and output, making the gradient transfer more efficient during backpropagation and alleviating the vanishing gradient problem. Specifically, the input x of the residual block undergoes operations such as a weight layer (a series of convolutional operations) and an activation function to obtain a residual function F(x). F(x) is added element-wise to x to obtain the output H(x) of the residual block. Subsequently, for the residual network, it directly learns F(x) = H(x) - x. By learning the residuals, the network can also distinguish effective and redundant features in different layers, avoiding network degradation.

[0106] Among them, ResNet has five different network structures with 18, 34, 50, 101, and 152 layers. The present application uses ResNet-18, which consists of a convolutional layer, residual blocks, and a fully connected layer. Specifically, first, the user feature data is preliminarily feature-extracted through the convolutional layer of ResNet (a convolutional layer with a dimension of 7×7, a stride of 2, and a padding of 3), and the generated first feature map captures the spatial correlation in the data. Subsequently, the first feature map is deeply learned through multiple residual blocks to enhance the learning ability and feature expression of the network. Among them, each residual block contains two convolutional layers with a dimension of 3×3, a stride of 1, and a padding of 1 and a residual connection. After each convolutional layer, there are batch normalization and ReLU functions. Between multiple residual blocks, the size of the feature map of the convolutional data map is halved, and the number of channels is doubled. Finally, the second feature map obtained through deep learning is fused through a fully connected layer to output the first classification label, that is, the prediction result of whether there is an abnormal state in the in-building distribution system.

[0107] This series of operations makes full use of the advantages of ResNet, effectively handles the complexity of the wireless environment and the diversity of the dataset, realizes the accurate detection of hidden obstacles in the in-building distribution system, greatly improves the accuracy and efficiency of monitoring, and provides strong support for the maintenance and optimization of the in-building distribution system.

[0108] Optionally, the above indoor distribution monitoring model is determined as follows: Determine the difference value between the first classification label and the second classification label through the target loss function, where the second classification label is the true label manually marked to represent the state of the indoor distribution system; Adjust the model parameters of the residual network according to the difference value to obtain the indoor distribution monitoring model.

[0109] In the embodiments of the present application, the target loss function is used to quantify the difference between the first classification label generated by the residual network and the second classification label manually marked, that is, the gap between the model prediction result and the actual state of the indoor distribution system. By calculating this difference value, the prediction accuracy of the model during the training process can be clearly evaluated, and the model parameters of the network can be adjusted accordingly to optimize the learning process of the network.

[0110] The specific expression is as follows:

[0111]

[0112] In the formula, J is the difference value, P is the number of sample points (user feature data), K is the number of label categories, y ic is the digital encoding of the sample target value, h θ (x i ) c is the predicted probability value that the observed sample x i belongs to category c.

[0113] Specifically, after each training iteration, the loss function calculates the gap between the classification label predicted by the model and the actual label. This difference value guides the update direction of the model parameters to ensure that the learning direction of the model is consistent with the target. Subsequently, the learning rate is optimized using an adaptive moment estimation algorithm (such as Adam), which can dynamically adjust the step size of parameter updates to help the model converge to the optimal state faster. In addition, techniques such as adding Dropout to prevent overfitting are also used to further improve the generalization ability of the model, enabling it to accurately judge the state of the indoor distribution system on unseen data.

[0114] Figure 4 is the curve of the loss value change during the training of ResNet-18, where green and red are the loss value curves during training and testing respectively. As Figure 4 can be seen, as the number of iterations increases, the loss value gradually stabilizes, and the difference between the model prediction result and the actual result gradually decreases.

[0115] Optionally, the above method further includes: determining the true positives, false positives, true negatives, and false negatives of the in-building distribution system, where the true positives are used to represent the number of in-building distribution systems for which the in-building distribution monitoring model correctly identifies the abnormal state; the false positives are used to represent the number of in-building distribution systems in the normal state that are incorrectly identified as abnormal by the in-building distribution monitoring model; the true negatives are used to represent the number of in-building distribution systems in the normal state that are correctly identified by the in-building distribution monitoring model; the false negatives are used to represent the number of in-building distribution systems in the abnormal state that are incorrectly identified as normal by the in-building distribution monitoring model; determining the block accuracy of the user feature data based on the true positives, false positives, true negatives, and false negatives, where the block accuracy is used to represent the classification accuracy of each feature matrix in the user feature data.

[0116] Furthermore, the site accuracy of the in-building distribution system can be determined based on the block accuracy, where the site accuracy is used to represent the overall classification accuracy of the in-building distribution system.

[0117] In the embodiments of the present application, the hourly data of a single sector within Y days is divided into multiple feature matrices (a total of N), and then the in-building distribution monitoring model independently predicts each feature matrix and outputs the judgment result of the entire in-building distribution system, that is:

[0118] S = [s 1 , s 2 , …, s N

[0119] where N is the number of user feature matrices; s i is the block accuracy, that is, the accuracy of each feature matrix, taking values of 0 or 1, representing normal or abnormal states respectively, i = 1, …, N; S is the site accuracy, that is, the accuracy of the entire in-building distribution system.

[0120] It should be noted that accuracy is a comprehensive indicator used to measure the correct classification ratio of a classification model on test data, that is, the sum of the correctly predicted samples of all categories (including normal and abnormal states) by the model accounts for the proportion of all classification decisions.

[0121] Specifically, the block accuracy is calculated for the accuracy of each feature matrix s i , which is applicable to evaluating the classification ability of the model for a single user feature matrix and can directly reflect the accuracy of the model in processing each data block.

[0122] The specific calculation formula is as follows:

[0123] Acc = (TP + TN) / (TP + TN + FP + FN)

[0124] In the formula, TP, FP, TN, and FN are the true positives, false positives, true negatives, and false negatives of the in-building distribution system respectively. The specific analysis is as follows: ​

[0125] True Positive (TP): The number of times the model correctly identifies a distributed antenna system with actual anomalies as an abnormal state, reflecting the effectiveness of the model in detecting abnormal situations.

[0126] False Positive (FP): The number of times the model incorrectly identifies a normal distributed antenna system as an abnormal state, usually reflecting the false alarm rate of the model under specific conditions.

[0127] True Negative (TN): The number of times the model correctly identifies a distributed antenna system in an actual normal state as a normal state, demonstrating the accuracy of the model in excluding abnormal situations.

[0128] False Negative (FN): The number of times the model incorrectly identifies a distributed antenna system in an abnormal state as a normal state, meaning that the model may miss real problems in some cases.

[0129] The station accuracy rate is the accuracy rate for judging the state of the entire distributed antenna system. Considering that the prediction of a single data block may be affected by local interference or data noise, the station accuracy rate adopts a decision-making mechanism, that is, if the prediction results of the majority of data blocks s i tend to be abnormal (≥3 / 5N), then the entire distributed antenna system is determined to be abnormal; otherwise, it is normal.

[0130] The specific calculation formula is as follows:

[0131]

[0132] Figure 5 is the change curve of the block accuracy rate during the training of ResNet-18, where the green and red lines are the accuracy rate curves during training and testing respectively. As Figure 5 can be seen, as the number of iterations increases, the block accuracy rate gradually rises to 85.3%, and the model prediction results become more and more accurate.

[0133] Figure 6 is the station accuracy rate after the training of ResNet-18 is completed, and different points represent different distributed antenna sectors. As Figure 6 can be seen, there are 120 stations in total for testing, the number of stations meeting 3 / 5N is 115, and the station accuracy rate is 95.83%.

[0134] Figure 7 is the confusion matrix of the station accuracy rate after the training of ResNet-18 is completed, which shows the specific sample distribution in the test set. As Figure 7 can be seen, the total number of samples is 120, among which the number of positive and negative samples is 61 and 59 respectively; 56 samples are correctly predicted with label 0 and 59 samples are correctly predicted with label 1.

[0135] In the embodiments of the present application, by combining the data statistical characteristics in the financial field with wireless communication technology and deep learning, a monitoring method for an AI in-building distribution system based on trendiness is proposed. By calculating indicators such as the true fluctuation amplitude, trend value, and trend index of the standing wave ratio of the in-building distribution system, this method effectively screens out systems with hidden obstacles and solves the problem of difficult dataset acquisition caused by wireless environment differences. At the same time, an improved residual network (ResNet) is used to train the behavioral characteristic data of selected users. This not only overcomes the misjudgment problem of traditional monitoring methods under the user tidal effect, but also significantly improves the accuracy of the model by introducing a target loss function and a parameter adjustment mechanism. It can detect abnormalities in the in-building distribution system before user complaints and repair them in a timely manner, significantly improving network quality and user experience. Especially in indoor scenarios such as hospitals, large shopping malls, and office buildings, it has extremely high monitoring and early warning capabilities for network quality degradation caused by passive device problems, thus optimizing the maintenance strategy of operators, reducing operation and maintenance costs, and enhancing the stability and reliability of the in-building distribution system.

[0136] According to the embodiments of the present application, a monitoring device for an in-building distribution system is provided. It should be noted that the monitoring device for the in-building distribution system in the embodiments of the present application can be used to execute the monitoring method for the in-building distribution system provided in the embodiments of the present application. The following introduces the monitoring device for the in-building distribution system provided in the embodiments of the present application.

[0137] Figure 8 is a structural diagram of a monitoring device for an in-building distribution system provided according to the embodiments of the present application. As shown in Figure X, the device includes:

[0138] An acquisition module 80, configured to acquire the standing wave ratio data of the in-building distribution system;

[0139] A determination module 82, configured to determine a first in-building distribution system according to the standing wave ratio data, where the first in-building distribution system is a part of the in-building distribution system in an abnormal state;

[0140] A training module 84, configured to determine the user characteristic data of the first in-building distribution system and a second in-building distribution system, and train a residual network according to the user characteristic data to obtain an in-building distribution monitoring model, where the second in-building distribution system is a part of the in-building distribution system in a normal state;

[0141] A monitoring module 86, configured to monitor the in-building distribution system through the in-building distribution monitoring model to obtain a monitoring result of the in-building distribution system, where the monitoring result is used to reflect the abnormal state of the in-building distribution system.

[0142] Through the acquisition module, determination module, training module and monitoring module in the monitoring device of the above-mentioned in-building distribution system, the purpose of accurately identifying and locating the obvious and hidden obstacles in the in-building distribution system is achieved, thus realizing early warning and rapid response, and effectively improving the technical effect of the indoor mobile network service quality. Furthermore, the technical problem that it is difficult for the in-building distribution monitoring technology in the related art to effectively identify the signal quality degradation caused by minor changes or loose connections of passive devices, especially prone to false alarms or missed alarms in an environment with significant user tidal effects and large index fluctuations, affecting network maintenance efficiency and user experience, is solved.

[0143] In the monitoring device of the in-building distribution system provided in the embodiment of the present application, the determination module is further configured to determine the true fluctuation amplitude, trend value, trend index and trend strength of the standing wave ratio of the in-building distribution system according to the standing wave ratio data; determine the average trend index of the in-building distribution system according to the true fluctuation amplitude, trend value, trend index and trend strength of the standing wave ratio; and determine the first in-building distribution system according to the average trend index and the first preset threshold.

[0144] In the monitoring device of the in-building distribution system provided in the embodiment of the present application, the training module is further configured to determine the target users in the first in-building distribution system and the second in-building distribution system, where the target users are used to represent the users who can reflect the overall performance of the in-building distribution system within a preset time period; obtain the behavior characteristic data of the target users, where the behavior characteristic data includes at least one of the following: the reference signal received power, signal-to-noise ratio and power headroom report of the target users; and convert the behavior characteristic data into the form of a feature matrix to obtain the user characteristic data.

[0145] In the monitoring device of the in-building distribution system provided in the embodiment of the present application, the training module is further configured to determine the abnormal data in the first in-building distribution system and the second in-building distribution system; remove the abnormal data, and determine the single-day call duration of all users in the first in-building distribution system and the second in-building distribution system, where the single-day call duration is used to represent the time length of the user's effective communication with the network within a day; and determine the users with a single-day call duration lower than the second preset threshold as the target users.

[0146] In the monitoring device of the in-building distribution system provided in the embodiment of the present application, the training module is further configured to preliminarily process the user characteristic data through the convolutional layer in the residual network to obtain the first feature map; perform deep feature learning on the first feature map through multiple residual blocks in the residual network to obtain the second feature map; and perform fusion processing on the second feature map through the fully connected layer in the residual network to obtain the prediction result corresponding to the user characteristic data, where the prediction result includes the first classification label for judging the abnormal state of the in-building distribution system.

[0147] In the monitoring device of the in-building distribution system provided by the embodiments of the present application, the training module is further configured to determine the difference value between the first classification label and the second classification label through a target loss function, where the second classification label is a true label manually marked to represent the state of the in-building distribution system; and adjust the model parameters of the residual network according to the difference value to obtain an in-building distribution monitoring model.

[0148] In the monitoring device of the in-building distribution system provided by the embodiments of the present application, the training module is further configured to determine the true positives, false positives, true negatives, and false negatives of the in-building distribution system, where the true positives are used to represent the number of in-building distribution systems with abnormal states correctly identified by the in-building distribution monitoring model; the false positives are used to represent the number of in-building distribution systems with normal states misidentified as abnormal by the in-building distribution monitoring model; the true negatives are used to represent the number of in-building distribution systems with normal states correctly identified by the in-building distribution monitoring model; the false negatives are used to represent the number of in-building distribution systems with abnormal states misidentified as normal by the in-building distribution monitoring model; and determine the block accuracy of the user feature data according to the true positives, false positives, true negatives, and false negatives, where the block accuracy is used to represent the classification accuracy of each feature matrix in the user feature data.

[0149] In the monitoring device of the in-building distribution system provided by the embodiments of the present application, the training module is further configured to determine the station accuracy of the in-building distribution system according to the block accuracy, where the station accuracy is used to represent the overall classification accuracy of the in-building distribution system.

[0150] The embodiments of the present application further provide an electronic device, including: a memory and a processor, where the memory is used to store program instructions; the processor is connected to the memory and is used to execute the monitoring method of the in-building distribution system described above.

[0151] It should be noted that the above electronic device is used to execute Figure 2 the monitoring method of the in-building distribution system shown, so the relevant explanations in the above monitoring method of the in-building distribution system also apply to this electronic device, and will not be elaborated here.

[0152] The embodiments of the present application further provide a non-volatile storage medium, which includes a stored computer program, where the device where the non-volatile storage medium is located executes the monitoring method of the in-building distribution system by running the computer program.

[0153] It should be noted that the above non-volatile storage medium is used to execute Figure 2 the monitoring method of the in-building distribution system shown, so the relevant explanations in the above monitoring method of the in-building distribution system also apply to this non-volatile storage medium, and will not be elaborated here.

[0154] The embodiments of the present application further provide a computer program product, including computer instructions, and the computer instructions implement the monitoring method of the in-building distribution system when executed by a processor.

[0155] It should be noted that the above computer program product is used to execute Figure 2 the monitoring method of the in-building distribution system shown. Therefore, the relevant explanations in the above monitoring method of the in-building distribution system also apply to this computer program product, and will not be elaborated here.

[0156] The serial numbers of the embodiments of the present application above are only for description and do not represent the advantages or disadvantages of the embodiments.

[0157] In the above embodiments of the present application, the descriptions of each embodiment have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0158] In several embodiments provided by the present application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only illustrative. For example, the division of the units can be a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some interfaces. The indirect coupling or communication connection of units or modules can be in an electrical or other form.

[0159] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place, or can be distributed to multiple units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0160] In addition, in each embodiment of the present application, the functional units can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0161] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), mobile hard disks, magnetic disks, or optical discs.

[0162] The above are only the preferred embodiments of this application. It should be noted that for those of ordinary skill in the art, without departing from the principle of this application, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of this application.

Claims

1. A monitoring method for a room distribution system, characterized in that: include: Obtain standing wave ratio data of indoor distribution system; Determine a first room-divided system according to the standing wave ratio data, wherein the first room-divided system is a room-divided system in which part of the room-divided system is in an abnormal state; Determine user feature data of the first room-based system and the second room-based system, and train the residual network according to the user feature data to obtain a room-based monitoring model, wherein the second room-based system is a room-based system in a normal state among the room-based systems; The room-divided system is monitored by the room-divided monitoring model to obtain a monitoring result of the room-divided system, wherein the monitoring result is used to reflect an abnormal state of the room-divided system.

2. The method according to claim 1, characterized in that Determining the first room subsystem according to the standing wave ratio data includes: Determine the actual fluctuation amplitude, trend value, trend index and trend strength of the standing wave ratio of the indoor distribution system according to the standing wave ratio data; Determining an average trend index of the room distribution system according to the actual fluctuation amplitude of the standing wave ratio, the trend value, the trend index and the trend intensity; The first compartment system is determined based on the average trend index and a first preset threshold.

3. The method according to claim 1, characterized in that: Determining user characteristic data of the first room-based system and the second room-based system includes: Determine target users in the first distributed room system and the second distributed room system, wherein the target users are used to represent users who can reflect the overall performance of the distributed room system within a preset time period; Acquire behavior characteristic data of the target user, wherein the behavior characteristic data includes at least one of the following: reference signal received power, signal-to-noise ratio, and power headroom report of the target user; The behavior feature data is converted into a feature matrix form to obtain the user feature data.

4. The method according to claim 3, characterized in that: Determining target users in the first room-based system and the second room-based system includes: Determining abnormal data in the first compartment system and the second compartment system; Remove the abnormal data, and determine the single-day call bill duration of all users in the first room-based system and the second room-based system, wherein the single-day call bill duration is used to indicate the length of time that the user effectively communicates with the network in one day; The user whose single-day call duration is lower than a second preset threshold is determined as the target user.

5. The method according to claim 1, characterized in that The residual network is trained according to the user feature data, including: Preliminarily processing the user feature data through a convolutional layer in the residual network to obtain a first feature map; Performing deep feature learning on the first feature map through multiple residual blocks in the residual network to obtain a second feature map; The second feature map is fused through the fully connected layer in the residual network to obtain a prediction result corresponding to the user feature data, wherein the prediction result includes a first classification label for determining an abnormal state of the room distribution system.

6. The method according to claim 5, characterized in that The method further comprises: Determine the difference value between the first classification label and the second classification label by using the target loss function, wherein the second classification label is a real label manually marked to represent the state of the room distribution system; The model parameters of the residual network are adjusted according to the difference value to obtain the room monitoring model.

7. The method according to claim 1, characterized in that The method further comprises: Determine the true positives, false positives, true negatives and false negatives of the room separation system, wherein the true positives are used to indicate the number of room separation systems correctly identified as abnormal by the room separation monitoring model; the false positives are used to indicate the number of room separation systems in normal state that are incorrectly identified as abnormal by the room separation monitoring model; the true negatives are used to indicate the number of room separation systems correctly identified as normal by the room separation monitoring model; and the false negatives are used to indicate the number of room separation systems in abnormal state that are incorrectly identified as normal by the room separation monitoring model; The block accuracy of the user feature data is determined according to the true positive examples, the false positive examples, the true negative examples, and the false negative examples, wherein the block accuracy is used to represent the classification accuracy of each feature matrix in the user feature data.

8. The method according to claim 7, characterized in that The method further comprises: The station accuracy of the indoor distributed system is determined according to the block accuracy, wherein the station accuracy is used to represent the overall classification accuracy of the indoor distributed system.

9. A monitoring device for a room distribution system, characterized in that: include: An acquisition module is used to acquire standing wave ratio data of the indoor distribution system; A determination module, configured to determine a first room-divided system according to the standing wave ratio data, wherein the first room-divided system is a room-divided system in which part of the room-divided system is in an abnormal state; A training module, used to determine user feature data of the first room-based system and the second room-based system, and train the residual network according to the user feature data to obtain a room-based monitoring model, wherein the second room-based system is a room-based system in a normal state among the room-based systems; The monitoring module is used to monitor the divided room system through the divided room monitoring model to obtain the monitoring result of the divided room system, wherein the monitoring result is used to reflect the abnormal state of the divided room system.

10. An electronic device, characterized in that: include: A memory and a processor, wherein the memory is used to store program instructions; The processor is connected to the memory and is used to execute the monitoring method of the indoor distribution system described in any one of claims 1 to 8.

11. A non-volatile storage medium, characterized in that: The non-volatile storage medium includes a stored computer program, wherein the device where the non-volatile storage medium is located executes the monitoring method for the indoor distributed system according to any one of claims 1 to 8 by running the computer program.

12. A computer program product comprising computer instructions, characterized in that: When the computer instructions are executed by the processor, the monitoring method of the indoor distribution system described in any one of claims 1 to 8 is implemented.

Images