Fault determination method and device, electronic equipment, storage medium and product

Abstract

The invention provides a fault determination method and device, electronic equipment, a storage medium and a product. The method comprises the following steps: acquiring first network coverage quality data of each space unit in a plurality of space units; performing iterative training on the network quality prediction model based on the first network coverage quality data to obtain a network quality evaluation model; acquiring second network coverage quality data of each space unit in the plurality of space units, and determining a network quality evaluation result corresponding to each space unit by using the network quality evaluation model; and determining the fault information in the plurality of space units based on the network quality evaluation result, thereby being capable of actively discovering hidden faults and improving the accuracy of fault positioning.

Description

Technical Field

[0001] This disclosure relates to the field of wireless technology, and in particular to a fault determination method, apparatus, electronic device, storage medium, and product. Background Technology

[0002] With the continuous expansion of indoor network coverage, indoor distributed antenna system (DAS) sites have become a core scenario for user traffic operations, and their operational quality directly affects network reputation and user experience. Current troubleshooting methods primarily rely on user complaints, performance degradation, explicit hardware alarms, and on-site investigations to identify problems. For latent faults, empirical judgment is needed, combining backend analysis with extensive frontend testing. This approach suffers from delayed problem detection and low accuracy in fault location, failing to meet the demands for efficient indoor network operation and maintenance and quality optimization. Summary of the Invention

[0003] This disclosure provides a fault determination method, apparatus, electronic device, storage medium, and product to solve the problems of delayed problem detection and low fault location accuracy in related technologies.

[0004] A first aspect of this disclosure provides a fault determination method, the method comprising: Obtain the first network coverage quality data for each of the multiple spatial units; Based on the first network coverage quality data, the network quality prediction model is iteratively trained to obtain the network quality assessment model. The second network coverage quality data of each spatial cell in multiple spatial cells is obtained, and the network quality assessment result corresponding to each spatial cell is determined by using a network quality assessment model; Based on the network quality assessment results, fault information in multiple spatial units was identified.

[0005] In one embodiment, acquiring first network coverage quality data for each of a plurality of spatial cells includes: Obtain raw network coverage quality data within the target building space; The target building space is divided into multiple spatial units according to the preset dimensions; Based on the original network coverage quality data and multiple spatial units, the first network coverage quality data for each spatial unit is determined.

[0006] In one embodiment, the raw network coverage quality data includes at least wireless measurement data, Internet data service data, and scalable data detection data.

[0007] In one embodiment, based on the original network coverage quality data and multiple spatial cells, determining the first network coverage quality data for each spatial cell includes: Obtain the three-dimensional coordinate range corresponding to each spatial unit; Based on the three-dimensional coordinate range corresponding to each spatial unit, determine the network coverage quality data within the same spatial unit from the original network coverage quality data; Network coverage quality data located within the same spatial cell are aggregated to obtain the first network coverage quality data for each spatial cell.

[0008] In one embodiment, fault information in multiple spatial units is determined based on network quality assessment results, including: Obtain network status metrics associated with the fault; Based on the values ​​of network status indicators, multiple consecutive value ranges are determined, and each value range corresponds to a network quality level. By comparing the network quality assessment results with the network quality level, fault information in multiple spatial units can be identified.

[0009] In one embodiment, based on first network coverage quality data, the network quality prediction model is iteratively trained to obtain a network quality assessment model, including: The first network coverage quality data is divided into training data and validation data; The network quality prediction model is iteratively trained using training data; The trained network quality prediction model was validated using validation data to obtain the validation accuracy. When the verification accuracy reaches a preset threshold, a network quality assessment model is obtained.

[0010] A second aspect of this disclosure provides a fault determination apparatus, the apparatus comprising: The acquisition unit is used to acquire the first network coverage quality data of each of the multiple spatial units. The training unit is used to iteratively train the network quality prediction model based on the first network coverage quality data to obtain the network quality evaluation model. The first determining unit is used to acquire the second network coverage quality data of each spatial unit in the multiple spatial units, and to determine the network quality assessment result corresponding to each spatial unit using the network quality assessment model. The second determining unit is used to determine fault information in multiple spatial units based on network quality assessment results.

[0011] A third aspect of this disclosure provides an electronic device comprising: At least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the methods described in the first aspect of this disclosure.

[0012] A fourth aspect of this disclosure provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to perform the methods described in the first aspect of this disclosure.

[0013] A fifth aspect of this disclosure provides a computer program product including a computer program that, when executed by a processor, implements the methods described in the first aspect of this disclosure.

[0014] In summary, this disclosure proposes a fault determination method, which includes: acquiring first network coverage quality data for each spatial unit in a plurality of spatial units; iteratively training a network quality prediction model based on the first network coverage quality data to obtain a network quality evaluation model; acquiring second network coverage quality data for each spatial unit in a plurality of spatial units, and using the network quality evaluation model to determine the network quality evaluation result corresponding to each spatial unit; and determining fault information in the plurality of spatial units based on the network quality evaluation results.

[0015] According to the solution provided in this disclosure, by acquiring the first network coverage quality data and the second network coverage quality data of each spatial unit in multiple spatial units, and using the network quality evaluation model trained with the first network coverage quality data to determine the network quality of each spatial unit corresponding to the second network coverage quality data, it is possible to proactively discover hidden faults and improve the accuracy of fault location.

[0016] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure, and are not intended to unduly limit this disclosure.

[0018] Figure 1 A flowchart illustrating a fault determination method provided in an embodiment of this disclosure; Figure 2 A flowchart illustrating a method for obtaining raw network coverage quality data according to an embodiment of this disclosure; Figure 3 This is a schematic diagram of the horizontal coverage range of an antenna provided in an embodiment of the present disclosure; Figure 4 This is a schematic diagram of an antenna elevation angle range provided in an embodiment of the present disclosure; Figure 5 A flowchart illustrating a method for determining the network quality assessment result corresponding to each spatial unit, provided in an embodiment of this disclosure; Figure 6 This is a schematic diagram illustrating neuronal information transmission according to an embodiment of the present disclosure; Figure 7 A schematic diagram of a fault determination process provided in an embodiment of this disclosure; Figure 8 A schematic diagram illustrating another fault determination process provided in this embodiment of the disclosure; Figure 9 This is a schematic diagram of the structure of a fault determination device provided in an embodiment of the present disclosure; Figure 10 This is a schematic diagram of the hardware composition structure of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0019] Embodiments of this disclosure are described in detail below. Examples of these embodiments are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this disclosure, and should not be construed as limiting this disclosure.

[0020] To facilitate a better understanding of the technical solutions described in the embodiments of this disclosure by those skilled in the art, the technical terms in the embodiments of this disclosure are explained as follows before introducing the embodiments of this disclosure.

[0021] As the scale of network coverage continues to expand, various indoor coverage scenarios have gradually become the main force in user traffic operation. With the surge in the scale of indoor network distribution, the operational quality of indoor distribution sites plays a more important supporting role in the overall network operation reputation.

[0022] Currently, the existing technical means for troubleshooting indoor distributed antenna system (DAS) sites mainly include user complaints, performance degradation, explicit hardware alarms, checking the rationality of parameter settings, and on-site problem investigation. Among the above solutions, the above solutions are more effective for explicit problems in indoor networks. However, for implicit faults, it is necessary to make empirical judgments by combining backend performance analysis with a large number of frontend tests. This usually results in difficulties in analysis and monitoring and low efficiency in fault location.

[0023] Meanwhile, on-site troubleshooting requires a lot of manpower and resources, which can significantly impact network operation quality and user experience.

[0024] Therefore, the relevant solutions mainly have the following defects: Problem detection relies primarily on user complaints, key performance indicators (KPIs), and alerts, resulting in a delayed and reactive response.

[0025] Fault location is determined by a series of layer-by-layer troubleshooting, which consumes a lot of manpower and resources and cannot achieve accurate location in a timely manner. The identification of latent faults is highly subjective, inefficient, inaccurate, and costly.

[0026] To address the shortcomings in related technologies, this disclosure obtains first and second network coverage quality data for each spatial unit in multiple spatial units, and uses a network quality assessment model trained on the first network coverage quality data to determine the network quality of each spatial unit corresponding to the second network coverage quality data. This approach can proactively detect hidden faults and improve the accuracy of fault location.

[0027] The present disclosure will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0028] The fault determination method provided in this disclosure can be applied to network fault location and operation and maintenance optimization scenarios in complex indoor wireless environments, such as network quality optimization in dense indoor scenarios like residential buildings and school dormitories. The execution subject of the method can be a hardware device, software system, or a platform combining hardware and software with data processing, model training, and analysis and decision-making capabilities, such as an operator's network operation and maintenance management platform or edge computing device.

[0029] like Figure 1 As shown, Figure 1 This is a flowchart illustrating the fault determination method provided in this embodiment of the disclosure. The fault determination method provided in this embodiment of the disclosure includes the following steps: Step 101: Obtain the first network coverage quality data for each spatial cell in the multiple spatial cells; In one embodiment, a spatial unit refers to an independent analysis unit that divides a target indoor area, such as a building, into units of precision. Spatial units are typically three-dimensional grids, such as 5m×5m×5m cube grids.

[0030] In one embodiment, the first network coverage quality data is sample data from multiple dimensions, including coverage network signal, device operation, and user perception.

[0031] In one embodiment, the first network coverage quality data of each spatial unit in multiple spatial units can be automatically obtained through the operator's network operation and maintenance platform.

[0032] In one embodiment, the first network coverage quality data of each of the multiple spatial units can also be obtained from actual field measurements.

[0033] Step 102: Based on the first network coverage quality data, iteratively train the network quality prediction model to obtain the network quality assessment model; In one embodiment, the network quality prediction model is used to determine the network status based on spatial cell network coverage quality data.

[0034] In one embodiment, the network quality prediction model can be a recurrent neural network (RNN) model, a gradient boosting tree (GBDT) model, or a deep learning hybrid model, such as a hybrid model of convolutional neural network (CNN) and RNN.

[0035] In one embodiment, the network quality assessment model is the final model formed by iteratively training, validating and optimizing the network quality prediction model using sample data, and has the ability to accurately assess the quality of spatial unit networks.

[0036] In one embodiment, before iteratively training the network quality prediction model, the first network coverage quality data can be normalized to obtain a format that can be used for analysis applications, and the normalized data can be inserted into the corresponding database for storage.

[0037] Step 103: Obtain the second network coverage quality data for each spatial unit in the multiple spatial units, and use the network quality assessment model to determine the network quality assessment result corresponding to each spatial unit; In one embodiment, the second network coverage quality data is current data used for network quality assessment and has the same data dimensions and format as the first network coverage quality data.

[0038] In one embodiment, the second network coverage quality data of each spatial unit in multiple spatial units can be obtained from the current network status data of the target area, such as by real-time data push from the operator's core network interface or base station equipment.

[0039] In one embodiment, the second network coverage quality data of each of the multiple spatial units can be acquired at a fixed period, which can be three hours or five hours.

[0040] In one embodiment, when abnormal fluctuations in key network indicators are detected, second network coverage quality data for each of the multiple spatial units can also be obtained, such as signal anomalies.

[0041] In one embodiment, the collected second network coverage quality data can be uploaded to a cloud server, and the network quality assessment model inference can be completed using cloud computing power to obtain the network quality assessment result corresponding to each spatial unit.

[0042] In one embodiment, the network quality assessment model can be deployed on edge devices near indoor scenes to perform data processing and assessment locally, thereby reducing network transmission latency.

[0043] Step 104: Based on the network quality assessment results, determine the fault information in multiple spatial units.

[0044] In one embodiment, the network quality assessment result is the result obtained by the network quality assessment model after analyzing the second network coverage quality data. It can quantitatively reflect the network coverage level of each spatial unit, such as signal strength, coverage rate, and interference level.

[0045] In one embodiment, the fault information is network anomaly-related information determined based on network quality assessment results, including fault location, i.e., which specific spatial unit it is, fault type, such as macro station intrusion, local spatial fault, and fault severity, such as mild, moderate, severe, etc.

[0046] In one embodiment, a box-and-whisker diagram analysis method can be used to statistically sort fault-related indicators, such as coverage and signal receiving power, and divide them into multiple quality level numerical ranges, such as five level numerical ranges, including very good, good, average, poor, and very poor. Poor and below numerical ranges are judged as faults.

[0047] In one embodiment, fixed thresholds can also be set based on industry standards and operational experience, such as a signal reception power below -110dBm being a fault, or a coverage rate below 75% being a fault. Meeting the threshold conditions is considered a fault.

[0048] In one embodiment, baseline values ​​for different time periods and regions can be calculated by combining historical data. When the evaluation result is lower than a preset percentage of the baseline value, such as 20%, it is determined to be a fault, thus adapting to the dynamic changes in network status.

[0049] Effective fault identification is achieved by collecting, training, and evaluating network coverage quality data for spatial units. First, initial network coverage quality data for each of the multiple spatial units is acquired. This initial data can encompass various relevant data reflecting network signal propagation, equipment operation, and user perception. Based on this initial data, a network quality prediction model is iteratively trained, and model parameters are optimized through data learning to ultimately obtain a network quality assessment model capable of accurately evaluating network quality. Secondary network coverage quality data is then input into the network quality assessment model, which outputs the network quality assessment results for each spatial unit. Finally, based on the network assessment results, it is determined whether faults exist in the multiple spatial units and what specific fault-related information they contain.

[0050] This application avoids excessive reliance on manual inspection, enabling rapid and accurate location of faults in multiple spatial units, thus improving the efficiency and accuracy of fault identification.

[0051] In one embodiment, acquiring first network coverage quality data for each of a plurality of spatial cells includes: Obtain raw network coverage quality data within the target building space; The target building space is divided into multiple spatial units according to the preset dimensions; Based on the original network coverage quality data and multiple spatial units, the first network coverage quality data for each spatial unit is determined.

[0052] In one embodiment, the target building space refers to the indoor area where network fault location and quality assessment need to be performed, including but not limited to the interior space of enclosed or semi-enclosed buildings such as office buildings, commercial complexes, transportation hubs, and residential buildings.

[0053] In one embodiment, the raw network coverage quality data is directly collected, unprocessed, multi-dimensional network status data, covering information such as network signaling interaction, wireless signal transmission, and user application perception.

[0054] In one embodiment, by communicating with the operator's core network equipment, the system automatically collects raw code stream data from key interfaces such as S1-MME and S1-U, including signaling interaction content such as user identifiers, network node information, tunnel identifiers, and access point information. It also automatically connects to base station equipment through a network operation and maintenance platform to collect data generated during mobile radio operation, such as base station identifiers, Received Signal Reference Power (RSRP), Signal to Interference plus Noise Ratio (SINR), frequency band information, and location-related parameters. Furthermore, using an automatic, seamless triggering method, it connects to typical internet apps, mobile terminal manufacturers, professional testing software vendors, and third-party suppliers to collect anonymized user terminal parameters, network signal perception, terminal location, and Wi-Fi information.

[0055] In one embodiment, the preset size refers to a pre-defined spatial unit division standard based on a balance between the physical structure of the target building space, the required analytical accuracy, and computational efficiency. This standard is used to discretize continuous building spaces into independently analyzable units.

[0056] In one embodiment, a 5m×5m×5m three-dimensional cubic grid can be divided based on a high-precision map and physical structure corresponding to the target building space, such as walls and floors.

[0057] In one embodiment, spatial units can be adaptively divided according to the coverage range of the base station signal. Specifically, the unit size of the strong signal coverage area can be expanded, and the unit size of the weak coverage area can be reduced. For example, a 10m×10m×10m cube grid can be used in an open office area, and a 3m×3m×3m cube grid can be used in corridors, elevator halls, etc.

[0058] In one embodiment, location information in the original network coverage quality data, such as latitude and longitude and floor height, can be filtered based on the coordinate range of the spatial unit to achieve matching between the original network coverage quality data and the spatial unit.

[0059] When acquiring raw network coverage quality data in the target building space, for the complex and enclosed indoor wireless environment, multi-source data covering network signaling, Internet application perception, and wireless operation are collected to ensure that the data covers multiple dimensions of information such as network operation status and user perception. When dividing the space into units according to preset dimensions, the target building space is divided into 5m×5m×5m cubic grids based on the physical structural characteristics of the indoor building, such as floor height and wall layout, with each grid serving as an independent spatial unit to ensure positioning accuracy. When determining the first network coverage quality data based on the raw network coverage quality data and multiple spatial units, a positional association mapping between each spatial unit and the raw data is first established, and then key quality parameters such as network signal strength and coverage range corresponding to each spatial unit are extracted to form the first network coverage quality data exclusive to each spatial unit.

[0060] By precisely dividing the space into units that are adapted to the indoor environment and combining the collection of raw data from multiple sources, we ensure that the first network coverage quality data accurately corresponds to the actual indoor space.

[0061] In one embodiment, the raw network coverage quality data includes at least wireless measurement data, Internet data service data, and scalable data detection data.

[0062] In one embodiment, the wireless measurement data includes start time, base station identifier (gNodeBID), cell identifier (CellID), Mobility Management Entity (MME) code (MmeCode), MME group identifier (MmeGroupID), user identifier (MmeUeS1apID) of the S1 application protocol of the mobile network key interface between the MME and the user equipment, reference signal received power (NRScRSRP) of the 5G serving cell, and 5G... The data includes: Reference Received Power of Neighboring Cell (NRNcRSRP), Reference Received Quality of 5G Serving Cell (NRScRSRQ), Reference Received Quality of 5G Neighboring Cell (NRNcRSRQ), Evolved General Terrestrial Radio Access Frequency Number of 5G Serving Cell (NRScEarfcn), Physical Cell Identifier of 5G Serving Cell (NRScPci), Evolved General Terrestrial Radio Access Frequency Number of 5G Neighboring Cell (NRNcEarfcn), Physical Cell Identifier of 5G Neighboring Cell (NRNcPci), Buffer Status Report of 5G Serving Cell (NRScBSR), Round-Trip Time Delay of 5G Serving Cell (NRScRTTD), Timing Advance of 5G Serving Cell (NRScTadv), Geographic Coordinates of 5G Serving Cell (NRScPOI), Power Margin Report of 5G Serving Cell (NRScPHR), Received Interference Power of 5G Serving Cell (NRScRIP), and Uplink Signal-to-Interference-plus-Noise Ratio of 5G Serving Cell (NRScSinrUL).

[0063] In one embodiment, the Internet data service data comes from various sources, including typical Internet applications, mobile terminal manufacturers, professional testing software manufacturers, and other third-party suppliers. It adopts an automatic triggering and seamless approach, mainly including de-identified temporary user stamps, timestamps, base station information, network affiliation, network signal, user terminal parameters, terminal location, WIFI, and other information about the user's perception of the local network and competing wireless networks in a 4G / 5G network environment.

[0064] In one embodiment, the scalable data detection data includes key mobile network interfaces S1-MME, S1-U interface, and user plane raw code stream information. Key information includes: signaling data length, city where the service occurred, signaling interface type, unique identifier of the scalable data record (xDR ID), radio access technology type (RAT), International Mobile Subscriber Identity (IMSI), International Mobile Equipment Identity (IMEI), Mobile Subscriber Integrated Services Digital Network Number (MSISDN), local province, local city, owner province, owner city, roaming type, machine IP address type, serving gateway / gateway GPRS support node IP address (SGW / GGSN IP Add), 5G base station / serving GPRS support node IP address (gNB / SGSN), IP Add, packet data network gateway IP address (PGWAdd), SGW / GGSN port number, and gNB / SGSN port number. Port), PGW port number (PGW Port), gNB / SGSN GTP tunnel endpoint identifier (gNB / SGSN GTP-TEID), SGW / GGSN GTP tunnel endpoint identifier (SGW / GGSN GTP), tracking area code (TEID), (TAC), cell ID, access point name (APN http), etc.

[0065] In one embodiment, such as Figure 2 As shown, Figure 2 This is a flowchart illustrating a method for obtaining raw network coverage quality data according to an embodiment of the present disclosure. The data includes MRO data (i.e., the aforementioned wireless measurement data), OTT data (i.e., the aforementioned Internet data service data), and XDR signaling data (i.e., the aforementioned scalable data detection data). The collected data is stored in the corresponding database.

[0066] The raw network coverage quality data includes at least wireless measurement data, Internet data service data, and scalable data detection data, which ensures the comprehensiveness of data collection and reflects the indoor network coverage status from multiple dimensions.

[0067] In one embodiment, based on the original network coverage quality data and multiple spatial cells, determining the first network coverage quality data for each spatial cell includes: Obtain the three-dimensional coordinate range corresponding to each spatial unit; Based on the three-dimensional coordinate range corresponding to each spatial unit, determine the network coverage quality data within the same spatial unit from the original network coverage quality data; Network coverage quality data located within the same spatial cell are aggregated to obtain the first network coverage quality data for each spatial cell.

[0068] In one embodiment, the spatial coordinate range refers to the three-dimensional position boundary of each spatial unit within the target building, defined by the start and end values ​​of coordinate intervals, such as the x-axis, y-axis, and z-axis.

[0069] In one embodiment, network coverage quality data within the same spatial cell refers to a subset of the original network coverage quality data whose location information falls within the coordinate range of a certain spatial cell, and can directly reflect the network status of that spatial cell.

[0070] In one embodiment, aggregation processing refers to the process of statistical analysis or feature extraction of multi-source, discrete raw data within the same spatial unit, used to transform discrete data into standardized indicators that can characterize the network quality of the spatial unit.

[0071] In one embodiment, the three-dimensional coordinate range corresponding to each spatial unit can be extracted from a high-precision geographic information map of the target building.

[0072] In one embodiment, location information such as latitude and longitude, floor height, and indoor relative coordinates in the network coverage quality data can be parsed and compared with the coordinate range of each spatial unit. If the location information falls within the coordinate range, it is determined to be data within that spatial unit.

[0073] In one embodiment, based on high-precision geographic information, the signal strength of the corresponding base station in each grid (i.e., the aforementioned spatial unit) is simulated and calculated using a propagation model based on 3D ray tracing. The signal strength refers to the network coverage quality data in each spatial unit, and the grid is marked with a unique identifier according to the field strength of each cell in the grid.

[0074] In one embodiment, location information in the original network coverage quality data can be matched with the POI of the spatial unit by using geographic coordinate points (POIs) to achieve precise binding between the original network coverage quality data and the unit. The POI accuracy information is then converted into angle percentage data, such as ≤5°@90%, which means that the measurement error of POIs at more than 90% of the sampling points is less than or equal to 5°.

[0075] Specifically, the horizontal POI is the antenna sector facing the base station cell. The vertical normal direction of the antenna cross-section is defined as 0°, and it rotates clockwise 360°. The horizontal coverage range of the antenna sector is 120°, with a main lobe of 60 degrees and side lobes of 60 degrees. Therefore, the horizontal coverage range of the antenna is [0, 60°] ⊆ [300°, 360°]. Figure 3 As shown, Figure 3 This is a schematic diagram of the horizontal coverage range of an antenna provided in an embodiment of this disclosure. The vertical POI is the front view of the AAU sector, with the normal direction of the cross-section being 0°. Values ​​above the normal are negative (i.e., negative on the side facing upwards) and values ​​below the normal are positive (i.e., positive on the side facing downwards). Figure 4 As shown, Figure 4 This is a schematic diagram of an antenna elevation angle range provided in an embodiment of the present disclosure, wherein the elevation angle range is [-90°, 90°], and the network coverage quality data of the common indoor location can be determined based on the above horizontal POI and vertical POI.

[0076] In one embodiment, based on the three-dimensional coordinate range of a spatial cell, the original network coverage quality data falling within that coordinate range are filtered out, and then aggregated, such as by calculating the average value and coverage rate, to obtain the first network coverage quality data of the corresponding cell.

[0077] When obtaining the three-dimensional coordinate range corresponding to each spatial unit, a unique three-dimensional spatial coordinate interval is assigned to each cubic grid unit after division based on the high-precision geographic information of the target building, clarifying the specific location boundary of each spatial unit within the building; when filtering data according to the three-dimensional coordinate range, the location association information in the original network coverage quality data is parsed, and the data is matched with the corresponding spatial coordinates to filter out the network coverage quality data that falls within the coordinate range of each spatial unit; when aggregating data within the same spatial unit, statistical analysis methods are used to calculate key indicators such as the average network signal strength, coverage, and occupancy rate within the unit, and these statistical results are used as the first network coverage quality data for that spatial unit.

[0078] By using coordinate matching to associate data with spatial units, and combining this with aggregation processing to extract key indicators, we can ensure that the first network coverage quality data of each spatial unit can accurately reflect the actual network coverage situation in that area.

[0079] In one embodiment, fault information in multiple spatial units is determined based on network quality assessment results, including: Obtain network status metrics associated with the fault; Based on the values ​​of network status indicators, multiple consecutive value ranges are determined, and each value range corresponds to a network quality level. By comparing the network quality assessment results with the network quality level, fault information in multiple spatial units can be identified.

[0080] In one embodiment, the network state indicators associated with a fault refer to quantitative parameters that can directly reflect the network coverage quality and operating status and are highly correlated with the occurrence of faults, and may include network coverage rate (such as in-building MR coverage rate), base station occupancy rate (such as in-building occupancy rate, macro station occupancy rate), signal reception power (such as average RSRP), and signal-to-interference-plus-noise ratio (such as average SINR).

[0081] In one embodiment, since there may be multiple network state indicators associated with a fault, the values of each indicator need to be arranged in descending order, and the upper edge, upper quartile, median, lower quartile, lower edge, and minimum outlier are calculated.

[0082] Exemplarily, as shown in Table 1, Table 1 is network information.

[0083]

[0084] [[ID=1']]Table 1 Network Information The five-color spectrum network quality level is determined through a box plot, as shown in Table 2, including: very good: green, upper quartile < X ≤ upper edge; better: blue, median < X ≤ upper quartile; average: yellow, lower middle quartile < X ≤ median; poor: orange, lower edge < X ≤ lower quartile; very poor: red, minimum outlier ≤ X ≤ lower edge, where X refers to the value of the network state indicator corresponding to each spatial unit.

[0085]

[0086] Table 2 Network Quality Level In one embodiment, multiple consecutive numerical intervals refer to non-overlapping and continuous numerical ranges formed based on the numerical distribution characteristics of network state indicators, and are used to classify and grade the indicator values; the network quality level is the network state evaluation level corresponding one-to-one with the consecutive numerical intervals, and is used to intuitively characterize the network coverage quality level of each spatial unit, and is usually divided into multiple levels from excellent to poor in terms of quality.

[0087] In one embodiment, key indicators can be set, such as average RSRP and network coverage rate, as the core judgment basis. If the network quality level corresponding to this indicator is poor or very poor, it is directly determined that there is a fault in this spatial unit, and the severity of the fault is consistent with the indicator level.

[0088] In one embodiment, the grade results of all fault-associated indicators can also be synthesized. If it exceeds the preset quantity, such as 3 indicators are in the poor or very poor grade, it is determined as a serious fault; 1-2 indicators are in this interval, it is determined as a minor fault; all indicators are very good, better, or average, it is determined as no fault.

[0089] In one embodiment, different weights can be assigned to each fault-related indicator, such as network coverage weight of 0.3, average RSRP weight of 0.3, average SINR weight of 0.2, and base station occupancy weight of 0.2. The indicator level weighted score of each spatial unit is calculated, and the severity of the fault is determined according to the score range. The lower the score, the more severe the fault.

[0090] For example, the evaluation results of spatial unit A are as follows: indoor MR coverage rate 72%, indoor occupancy rate 78%, macro station occupancy rate 22%, average RSRP -112dBm, and average SINR -4dB. Comparing the evaluation results of spatial unit A with the network quality level, all indicators are in the red to very poor level. Based on the characteristics of the indicator combination, it is determined that unit A has a serious fault. The fault type is a local area passive device fault or a radio frequency remote unit fault. The fault location is the three-dimensional coordinates of unit A, and the fault degree is serious.

[0091] When acquiring network status indicators associated with faults, key indicators such as network coverage, base station occupancy, signal received power, and signal-to-interference-plus-noise ratio are specifically selected. These indicators directly reflect network coverage and operational quality. When determining multiple continuous numerical intervals, a box-and-whisker diagram analysis method is used to sort the values ​​of each network status indicator from largest to smallest, and calculate the upper edge, upper quartile, median, lower quartile, lower edge, and minimum outlier. Based on this, five continuous quality level intervals are divided, corresponding to five network quality levels: very good, good, average, poor, and very poor. When comparing and determining fault information, the network quality assessment result of each spatial unit is compared with the above quality level intervals. The network quality level of the corresponding spatial unit is determined according to the interval to which the assessment result belongs, thereby identifying the spatial units with faults and clarifying the severity of the faults.

[0092] By using the above network quality level classification method and fault judgment logic, the fault identification process is standardized and quantified, which can avoid errors caused by subjective judgment and improve the accuracy and reliability of fault identification.

[0093] In one embodiment, based on first network coverage quality data, the network quality prediction model is iteratively trained to obtain a network quality assessment model, including: The first network coverage quality data is divided into training data and validation data; The network quality prediction model is iteratively trained using training data; The trained network quality prediction model was validated using validation data to obtain the validation accuracy. When the verification accuracy reaches a preset threshold, a network quality assessment model is obtained.

[0094] In one embodiment, the training data is sample data extracted from the first network coverage quality data for learning the parameters of the network quality prediction model. Its core function is to enable the model to grasp the correlation between spatial unit network quality data and network state. The validation data is sample data extracted from the first network coverage quality data for testing the training effect of the model and for evaluating the generalization ability and prediction accuracy of the model.

[0095] In one embodiment, iterative training of the network quality prediction model refers to the process of continuously adjusting parameters such as model weights and biases to continuously optimize the fitting effect of the network quality prediction model until the prediction error of the network quality prediction model on the training data tends to stabilize.

[0096] In one embodiment, when iteratively training the network quality prediction model using training data, a stochastic gradient descent algorithm can be used, with training data input in batches, loss function values ​​calculated after each batch of training, and model weights and bias parameters adjusted along the gradient descent direction; alternatively, an adaptive moment estimation algorithm can be used to dynamically adjust the learning rate, accelerate parameter update speed in the early stages of training, and gradually reduce the step size in the later stages to avoid model oscillation and improve convergence efficiency.

[0097] In one embodiment, when iteratively training the network quality prediction model using training data, a fixed number of training rounds can be set, such as 100 rounds, and training can be stopped after the preset number of rounds is completed; alternatively, a loss function threshold can be set, such as 0.01, and training can be stopped when the loss function value of the training data is lower than 0.01 and there is no significant decrease for several consecutive rounds, such as 5 rounds.

[0098] In one embodiment, the verification accuracy is the degree to which the model's output, after predicting the verification data, matches the actual network quality status in the verification data.

[0099] In one embodiment, the verification accuracy can be determined based on the proportion of samples whose predicted values ​​match the actual values ​​to the total number of verification samples; alternatively, the verification accuracy can be determined by calculating the mean of the squared differences between the predicted values ​​and the actual values.

[0100] In one embodiment, the preset threshold is a pre-set verification accuracy standard used to determine whether the model training has met the standard.

[0101] In one embodiment, if the verification accuracy does not reach a preset threshold, iterative training is performed again until the verification accuracy meets the target.

[0102] For example, when dividing training and validation data, samples are selected from the first network coverage quality data in a 7:3 ratio, with 70% of the samples used as training data for model training and 30% used as validation data for model performance validation. During iterative training using the training data, a recurrent neural network algorithm can be used to construct the network quality prediction model. This algorithm preserves the sequence dependencies of the data and continuously adjusts the model's weights and bias parameters by inputting the training data into the model, thereby optimizing the model's predictive performance. During model validation, validation data is input into the trained model, and the degree of matching between the model's output and the actual data is calculated to obtain the validation accuracy. A preset threshold can be set to 90%. When the validation accuracy reaches 90%, iterative training is stopped, and the model at this point is determined as the final network quality evaluation model.

[0103] In one embodiment, such as Figure 5 As shown, Figure 5 This is a flowchart illustrating a method for determining the network quality assessment result corresponding to each spatial unit, provided in an embodiment of this disclosure. The method includes a building indoor distributed antenna system (DAS) spatial grid network information database (i.e., the first network coverage quality data of each spatial unit among the aforementioned multiple spatial units), model building, selecting 70% of the data in the network information database as the training basis for subsequent machine learning to establish a building network information model; using machine learning algorithms to build a model of the network information, employing a specialized RNN recurrent neural network algorithm, where each neuron, in addition to inputting, processing, and outputting the current information, also passes the processed data to the next neuron, continuously optimizing and improving the final data analysis quality. The process is as follows: Figure 6 As shown, Figure 6 This is a schematic diagram of neuronal information transmission provided in an embodiment of this disclosure. The final simplified form is the same as that of a general neuron, where the input information is multiplied by weights and then a bias is added. Where W is the information weight and b is the bias value. 30% of the data in the building network information database is selected as test sequences for testing the established model. If the test accuracy is less than 90%, the model is optimized by constructing a building distribution model; otherwise, the building network information model is determined. When the test sequence's accuracy is less than 90%, the model is returned to the training sequence for calibration until the test sequence's accuracy is greater than 90%. The model with a test accuracy greater than 90% is selected as the building network information model.

[0104] In one embodiment, such as Figure 7 As shown, Figure 7This is a flowchart illustrating a fault determination process according to an embodiment of this disclosure. Macro-site intrusion includes good building or indoor MR coverage, good average RSRP, a downward shift in the macro-site occupancy rate chromatogram, and a simultaneous downward shift in the indoor occupancy rate chromatogram; all associated indicators above the building or indoor MR grid are in the orange or red spectrum. This is determined to be due to excessive macro-site coverage or abnormal parameter and neighbor cell configurations between the macro-site and the indoor MR cell. Building-wide faults include a downward shift in the building's indoor MR coverage rate chromatogram, or a downward shift in the average SINR chromatogram, with all indoor MR grid chromatograms in the building shifting downwards; the building's indoor MR coverage rate or average SINR is in the orange or red spectrum. This is determined to be a latent fault in the entire building's indoor MR system, and the parameters and fault conditions of its associated indoor MR cells are checked, such as handover parameter adjustments, cell backbone faults, combiner faults, etc. Localized building faults include a downward shift in the MR coverage rate chromatogram or average SINR chromatogram of some indoor MR grids in the building; the building's partial indoor MR coverage rate or average SINR is in the orange or red spectrum. The problem is determined to be a fault in a passive device or RRU in a local area of ​​the building. The problem in a certain branch is determined by combining the RRU granularity, floor and area.

[0105] By rationally dividing the sample data and setting verification thresholds, the predictive accuracy and reliability of the network quality assessment model are ensured, enabling it to accurately reflect the quality of indoor network coverage.

[0106] For example, such as Figure 8 As shown, Figure 8 This is a schematic diagram of another fault determination process provided in this embodiment of the disclosure. It utilizes a defined building network information model (i.e., the aforementioned network quality assessment model), imports the current building network information (i.e., the second network coverage quality data of each spatial unit in multiple spatial units), and realizes network monitoring and fault location of abnormal buildings. Specifically, it acquires the current MRO data of the building and generates a working input sequence. The working input sequence is imported into the building network model. The building network information model processes the working input sequence to generate indoor distribution signal index chromatograms, enabling the identification and location of building network faults. Simultaneously, the working results are analyzed and output. Finally, through summarizing experience, we continuously improved the accuracy of indoor positioning by fixing chromatographic problems.

[0107] In summary, the solution provided in this public disclosure is as follows: By acquiring the first and second network coverage quality data for each spatial unit in multiple spatial units, and using a network quality assessment model trained on the first network coverage quality data to determine the network quality of each spatial unit corresponding to the second network coverage quality data, latent faults can be proactively detected, improving the accuracy of fault location.

[0108] To implement the fault determination method provided in this disclosure, this disclosure also provides a fault determination device, such as... Figure 9 As shown. Figure 9 This is a schematic diagram of the structure of a fault determination device provided in an embodiment of the present disclosure. The fault determination device 900 includes: Acquisition unit 901 is used to acquire the first network coverage quality data of each spatial unit in multiple spatial units; Training unit 902 is used to iteratively train the network quality prediction model based on the first network coverage quality data to obtain the network quality evaluation model. The first determining unit 903 is used to acquire the second network coverage quality data of each spatial unit in the multiple spatial units, and to determine the network quality assessment result corresponding to each spatial unit using the network quality assessment model. The second determining unit 904 is used to determine fault information in multiple spatial units based on network quality assessment results.

[0109] In one embodiment, the acquisition unit 901 is specifically used for: Obtain raw network coverage quality data within the target building space; The target building space is divided into multiple spatial units according to the preset dimensions; Based on the original network coverage quality data and multiple spatial units, the first network coverage quality data for each spatial unit is determined.

[0110] In one embodiment, the raw network coverage quality data includes at least wireless measurement data, Internet data service data, and scalable data detection data.

[0111] In one embodiment, the first determining unit 903 is specifically used for: Obtain the three-dimensional coordinate range corresponding to each spatial unit; Based on the three-dimensional coordinate range corresponding to each spatial unit, determine the network coverage quality data within the same spatial unit from the original network coverage quality data; Network coverage quality data located within the same spatial cell are aggregated to obtain the first network coverage quality data for each spatial cell.

[0112] In one embodiment, the second determining unit 904 is specifically used for: Obtain network status metrics associated with the fault; Based on the values ​​of network status indicators, multiple consecutive value ranges are determined, and each value range corresponds to a network quality level. By comparing the network quality assessment results with the network quality level, fault information in multiple spatial units can be identified.

[0113] In one embodiment, the training unit 902 is specifically used for: The first network coverage quality data is divided into training data and validation data; The network quality prediction model is iteratively trained using training data; The trained network quality prediction model was validated using validation data to obtain the validation accuracy. When the verification accuracy reaches a preset threshold, a network quality assessment model is obtained.

[0114] It should be noted that the fault determination device provided in the above embodiments is only illustrated by the division of the above-described program modules when performing fault determination. In practical applications, the above processing can be assigned to different program modules as needed, that is, the internal structure of the fault determination device can be divided into different program modules to complete all or part of the processing described above. In addition, the fault determination device provided in the above embodiments and the fault determination method provided in the embodiments of this disclosure belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.

[0115] Figure 10 This is a schematic diagram of the hardware composition structure of the electronic device provided in the embodiments of this disclosure, such as... Figure 10 As shown, the electronic device 1000 includes at least one processor 1002; and a memory 1001 communicatively connected to the at least one processor 1002; wherein the memory 1001 stores instructions executable by the at least one processor 1002, the instructions being executed by the at least one processor 1002 to implement the steps of the fault determination method of the present disclosure embodiment.

[0116] Optionally, the electronic device may specifically be a fault determination device according to the embodiments of this application, and the electronic device may implement the corresponding processes implemented by the fault determination device in the various methods of the embodiments of this application. For the sake of brevity, it will not be described in detail here.

[0117] It is understood that the electronic device also includes a communication interface 1003. Various components in the electronic device are coupled together via a bus system 1004. It is understood that the bus system 1004 is used to implement communication between these components. In addition to a data bus, the bus system 1004 also includes a power bus, a control bus, and a status signal bus. However, for clarity, in... Figure 10 The general labeled all buses as Bus System 1004.

[0118] It is understood that memory 1001 can be volatile memory or non-volatile memory, or both. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), ferromagnetic random access memory (FRAM), flash memory, magnetic surface memory, optical disc, or compact disc read-only memory (CD-ROM); magnetic surface memory can be disk storage or magnetic tape storage. Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Synchronous Static Random Access Memory (SSRAM), Dynamic Random Access Memory (DRAM), Synchronous Dynamic Random Access Memory (SDRAM), Double Data Rate Synchronous Dynamic Random Access Memory (DDRSDRAM), Enhanced Synchronous Dynamic Random Access Memory (ESDRAM), SyncLink Dynamic Random Access Memory (SLDRAM), and Direct Rambus Random Access Memory (DRRAM).The memory 1001 described in the embodiments of the present invention is intended to include, but is not limited to, these and any other suitable types of memory.

[0119] The methods disclosed in the above embodiments can be applied to or implemented by the processor 1002. The processor 1002 may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above methods can be completed by integrated logic circuits in the hardware of the processor 1002 or by instructions in software form. The processor 1002 may be a general-purpose processor, a DSP, or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor 1002 can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. A general-purpose processor may be a microprocessor or any conventional processor, etc. The steps of the methods disclosed in the embodiments of this invention can be directly manifested as execution by a hardware decoding processor, or execution by a combination of hardware and software modules in the decoding processor. The software modules may be located in a storage medium, which is located in the memory 1001. The processor 1002 reads information from the memory 1001 and, in conjunction with its hardware, completes the steps of the aforementioned methods.

[0120] In an exemplary embodiment, the electronic device may be implemented by one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), FPGAs, general-purpose processors, controllers, MCUs, microprocessors, or other electronic components to perform the aforementioned method.

[0121] This disclosure also provides a non-transitory computer-readable storage medium storing computer instructions, which are used to cause a computer to execute the steps of the fault determination method of the present invention.

[0122] Optionally, the computer-readable storage medium can be applied to the fault determination apparatus in the embodiments of this application, and the computer instructions cause the computer to execute the corresponding processes implemented by the fault determination apparatus in the various methods of the embodiments of this application. For the sake of brevity, they will not be described in detail here.

[0123] This disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the fault determination method provided in this embodiment of the invention.

[0124] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0125] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0126] In addition, in the various embodiments of the present invention, each functional unit can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0127] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

[0128] Alternatively, if the integrated units of this invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

[0129] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A fault determination method, characterized in that, include: Obtain the first network coverage quality data for each of the multiple spatial units; Based on the first network coverage quality data, the network quality prediction model is iteratively trained to obtain the network quality assessment model. The second network coverage quality data of each spatial unit in multiple spatial units is obtained, and the network quality assessment result corresponding to each spatial unit is determined by the network quality assessment model. Based on the network quality assessment results, fault information in the multiple spatial units is determined.

2. The method according to claim 1, characterized in that, The acquisition of the first network coverage quality data for each of the multiple spatial cells includes: Obtain raw network coverage quality data within the target building space; The target building space is divided into multiple spatial units according to preset dimensions; Based on the original network coverage quality data and the multiple spatial units, the first network coverage quality data for each spatial unit is determined.

3. The method according to claim 2, characterized in that, The raw network coverage quality data includes at least wireless measurement data, Internet data service data, and scalable data detection data.

4. The method according to claim 2, characterized in that, The step of determining the first network coverage quality data for each spatial unit based on the original network coverage quality data and the plurality of spatial units includes: Obtain the three-dimensional coordinate range corresponding to each spatial unit; Based on the three-dimensional coordinate range corresponding to each spatial unit, determine the network coverage quality data located within the same spatial unit from the original network coverage quality data; The network coverage quality data located in the same spatial unit are aggregated to obtain the first network coverage quality data for each spatial unit.

5. The method according to claim 1, characterized in that, The step of determining fault information in the plurality of spatial units based on the network quality assessment results includes: Obtain network status metrics associated with the fault; Based on the values ​​of the network status indicators, multiple consecutive value intervals are determined, and each value interval corresponds to a network quality level. The network quality assessment results are compared with the network quality level to determine the fault information in the multiple spatial units.

6. The method according to claim 1, characterized in that, The step of iteratively training the network quality prediction model based on the first network coverage quality data to obtain a network quality assessment model includes: The first network coverage quality data is divided into training data and validation data; The network quality prediction model is iteratively trained using the training data. The trained network quality prediction model was validated using the validation data to obtain the validation accuracy. When the verification accuracy reaches a preset threshold, the network quality assessment model is obtained.

7. A fault determination device, characterized in that, include: The acquisition unit is used to acquire the first network coverage quality data of each of the multiple spatial units. The training unit is used to iteratively train the network quality prediction model based on the first network coverage quality data to obtain the network quality evaluation model. The first determining unit is used to acquire the second network coverage quality data of each spatial unit in the multiple spatial units, and to determine the network quality assessment result corresponding to each spatial unit using the network quality assessment model. The second determining unit is used to determine fault information in the plurality of spatial units based on the network quality assessment results.

8. An electronic device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to perform the method according to any one of claims 1 to 6.

10. A computer program product comprising a computer program that, when executed by a processor, implements the method of any one of claims 1 to 6.

Images