A method and apparatus for wireless coverage fault identification and intelligent optimization

Abstract

This application discloses a method and apparatus for wireless coverage fault identification and intelligent optimization, belonging to the field of wireless network technology. The method includes: acquiring multiple spatial heat maps and multiple time-series curves in a railway scenario; stitching the multiple spatial heat maps together according to their acquisition time to obtain a stitched image; inputting the stitched image and the time-series curves into a fault identification model, and using the fault identification model to identify network fault information in the railway scenario based on at least one of the stitched image and the time-series curves. According to the embodiments of this application, the efficiency of network fault determination can be improved.

Description

Technical Field

[0001] This application belongs to the field of wireless network technology, and in particular relates to a method and apparatus for wireless coverage fault identification and intelligent optimization. Background Technology

[0002] As society develops, users' demand for networks is constantly increasing, which also increases the probability of network failures during use.

[0003] However, in existing technologies, network operators need to identify network faults in a railway scenario based on network performance images such as spatial heat maps and time-series curves. Therefore, the discovery of network faults is highly dependent on the experience of the operators; at the same time, because operators need to spend a lot of time identifying network faults, the efficiency of fault identification is low. Summary of the Invention

[0004] This application provides a method and apparatus for wireless coverage fault identification and intelligent optimization, which can solve the technical problem of low efficiency in determining existing network faults.

[0005] In a first aspect, embodiments of this application provide a method for wireless coverage fault identification and intelligent optimization, the method comprising:

[0006] Obtain multiple spatial heat maps and multiple time-series curves in the railway scenario;

[0007] Based on the acquisition time of the multiple spatial heat maps, the multiple spatial heat maps are stitched together to obtain a stitched image;

[0008] The stitched image and the time-series curve are input into the fault identification model, and the fault identification model identifies network fault information of the railway scenario based on at least one of the stitched image and the time-series curve.

[0009] In some embodiments, stitching together the multiple spatial heat maps according to their acquisition time to obtain a stitched image includes:

[0010] According to the time sequence of the acquisition time of the multiple spatial heat maps, the multiple spatial heat maps are stitched together along a first direction perpendicular to the railway line to obtain a stitched image, wherein the stitched image includes multiple linear images that extend along the railway line and are arranged in parallel in the time sequence direction.

[0011] In some embodiments, stitching together the multiple spatial heat maps according to their acquisition time to obtain a stitched image includes:

[0012] For any pixel in any spatial heatmap, obtain the position information and color information of the pixel;

[0013] Obtain the acquisition time of the spatial heatmap of the pixel;

[0014] The feature vector of the pixel is determined based on the acquisition time of the spatial heat map, the position information and color information of the pixel;

[0015] The feature vectors of all pixels in the multiple spatial heatmaps are fused to obtain the stitched image.

[0016] In some embodiments, before inputting the stitched image and the time-series curve into a fault identification model, and before the fault identification model identifies network fault information of the railway scenario based on at least one of the stitched image and the time-series curve, the method further includes:

[0017] Obtain a trained computer vision model;

[0018] By adding a network structure to the input of the computer vision model, the computer vision model is fine-tuned to obtain a fault identification model.

[0019] The fault identification model is trained to obtain a converged fault identification model.

[0020] In some embodiments, after inputting the stitched image into the fault identification model to obtain the network fault information of the railway scenario output by the fault identification model, the method further includes:

[0021] If the network fault information includes at least two network faults in the railway scenario, the at least two network faults are displayed, wherein the network faults include transient network faults, short-term interference faults, and base station faults;

[0022] Receive a first input for the first network fault among the at least two network faults;

[0023] In response to the first input, the first network fault corresponding to the first input is input into the scheme prediction model, and the target solution for the first network fault is determined through the scheme prediction model.

[0024] In some embodiments, after determining the target solution for the first network fault using the scheme prediction model, the method further includes:

[0025] A simulation scenario is constructed based on the target solution and the railway scenario;

[0026] The propagation of wireless signals under the simulated scenario is simulated to obtain a simulated spatial heat map and a simulated time series curve under the simulated scenario;

[0027] If the fault identification model determines that there is no network fault in the simulation scenario based on the simulated spatial heatmap and the simulated time series curve, it outputs the target solution to the user.

[0028] Secondly, embodiments of this application provide a wireless coverage fault identification and intelligent optimization device, the device comprising:

[0029] The first acquisition module is used to acquire multiple spatial heat maps and multiple time-series curves in the railway scenario;

[0030] The stitching module is used to stitch together the multiple spatial heat maps according to their acquisition time to obtain a stitched image;

[0031] The first identification module is used to input the stitched image and the time-series curve into the fault identification model, and to identify network fault information of the railway scenario based on at least one of the stitched image and the time-series curve.

[0032] Thirdly, embodiments of this application provide a wireless coverage fault identification and intelligent optimization device, the device including: a processor and a memory storing computer program instructions;

[0033] The processor implements the above-mentioned wireless coverage fault identification and intelligent optimization method when executing computer program instructions.

[0034] Fourthly, embodiments of this application provide a computer storage medium storing computer program instructions, which, when executed by a processor, implement the aforementioned wireless coverage fault identification and intelligent optimization method.

[0035] Fifthly, embodiments of this application provide a computer program product, which includes computer program instructions. When the computer program instructions are executed by a processor, they implement the above-mentioned wireless coverage fault identification and intelligent optimization method.

[0036] In this application, multiple time-series thermal images of railway scenes are time-series stitched together to obtain a stitched image. This stitched image and the time-series curve are then input into a fault identification model, enabling the model to identify fault information in the network based on at least one factor from both the stitched image and the time-series curve. Compared to traditional methods relying on manual image-by-image analysis, this approach automatically identifies fault types through a model. Furthermore, by jointly analyzing the stitched thermal image and the time-series curve, the model can quickly and accurately determine fault types in different times and spaces, thereby significantly improving the efficiency of fault identification and intelligent optimization for network wireless coverage. Attached Figure Description

[0037] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 This is a flowchart illustrating a wireless coverage fault identification and intelligent optimization method provided in an embodiment of this application;

[0039] Figure 2 This is a schematic diagram of the structure of a wireless coverage fault identification and intelligent optimization device provided in an embodiment of this application;

[0040] Figure 3 This is a schematic diagram of the hardware structure of a wireless coverage fault identification and intelligent optimization device provided in an embodiment of this application. Detailed Implementation

[0041] The features and exemplary embodiments of various aspects of this application will be described in detail below. To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain this application and not to limit it. For those skilled in the art, this application can be implemented without some of these specific details. The following description of the embodiments is merely to provide a better understanding of this application by illustrating examples of this application.

[0042] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0043] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. The embodiments will now be described in detail with reference to the accompanying drawings.

[0044] Current wireless network optimization primarily relies on two methods: expert experience and intelligent optimization algorithms. Expert experience methods adjust the network using measurement data from dynamic testing vehicles and empirical models, but suffer from low efficiency, unstable optimization results, and difficulty in quickly responding to dynamic changes in the network environment. Intelligent optimization algorithms, based on particle swarm optimization and ant colony optimization, perform global statistical feature optimization, but cannot accurately handle specific problems such as over-coverage and weak coverage caused by signal differences between base stations, resulting in limited optimization effectiveness in railway scenarios.

[0045] The railway scenario exhibits unique network characteristics: long, strip-shaped coverage areas, the Doppler effect caused by high-speed train movement, and frequent base station handovers. These characteristics pose significant challenges to conventional network fault diagnosis methods: on the one hand, fault judgment logic based on key signaling interactions in static scenarios is no longer applicable in railway scenarios; on the other hand, traditional heatmap analysis methods struggle to effectively identify abnormal signal patterns under dynamic conditions.

[0046] The existing technology has the following prominent problems: First, it cannot accurately distinguish the signal anomaly patterns unique to railway scenarios, such as the difference between signal attenuation caused by the Doppler effect and real network faults; second, it lacks the ability to jointly analyze time-series characteristics, making it difficult to identify progressive faults such as base station aging; and third, the solution generation process is disconnected from fault diagnosis, making it impossible to achieve closed-loop optimization.

[0047] Specifically, in order to address the problems of the prior art, embodiments of this application provide a method and apparatus for wireless coverage fault identification and intelligent optimization. The method for wireless coverage fault identification and intelligent optimization provided in this application embodiment will be described first below.

[0048] Figure 1A flowchart illustrating a wireless coverage fault identification and intelligent optimization method according to an embodiment of this application is shown. The method includes the following steps:

[0049] S110 acquires multiple spatial heat maps and multiple time-series curves in the railway scenario.

[0050] In this embodiment, a spatial heatmap refers to a two-dimensional image generated from wireless signal measurement data, used to characterize the spatial distribution of signal strength or delay. The spatial heatmap may include a coverage heatmap, a carrier-to-interference ratio (C / I) heatmap, and a timing advance (TA) heatmap. The coverage heatmap can be used to represent the spatial distribution of received signal strength (RSRP) at various locations along the railway line, using color to reflect the quality of signal coverage. The C / I heatmap can be used to represent the distribution of signal quality (C / I, i.e., carrier-to-interference ratio) at various locations along the railway line, used to determine the interference intensity. The TA heatmap can represent the delay between various locations along the line and the base station, reflecting the distance or propagation delay between the user and the base station.

[0051] A time-series graph is a line graph that shows the trend of a network performance indicator over time or distance, used to analyze dynamic fluctuations in network status. Time-series graphs can include coverage graphs, carrier-to-interference ratio (C / I) graphs, and transfer rate (TA) graphs. Coverage graphs show the trend of signal strength over time or distance during train operation, used to analyze coverage stability. Carrier-to-interference ratio (C / I) graphs show the trend of signal quality (C / I) along the train's path, used to identify interference fluctuations. TA graphs show the trend of TA values ​​during train operation, used to determine distance changes or handover anomalies.

[0052] In a railway scenario, a spatial heat map refers to a heat map of a single line segment collected along the railway line trajectory, reflecting the distribution of network performance indicators at various locations on the railway at a specific moment. These network performance indicators can include signal coverage strength, interference level, and transfer rate (TA) value.

[0053] Since each spatial heatmap corresponds to a different acquisition time, and the acquisition trajectory and format of each spatial heatmap are fixed and uniform, possessing time-series characteristics, multiple spatial heatmaps acquired in the railway scenario can be stitched together in chronological order to form a two-dimensional mosaic image with a joint time-space representation. This two-dimensional mosaic image can be used to observe the evolution of network states over time. Both the first and second stitched images are the aforementioned two-dimensional mosaic images.

[0054] S120, based on the acquisition time of the multiple spatial heat maps, the multiple spatial heat maps are stitched together to obtain a stitched image;

[0055] In this embodiment, since a single spatial heatmap typically refers to a spatial heatmap of a single line segment collected along a railway track, reflecting the distribution of network performance indicators at various locations on the railway at a certain moment, a single spatial heatmap can only express the spatial signal distribution. Multiple spatial heatmaps collected in the railway scenario can be stitched together in chronological order based on their acquisition time to form a two-dimensional stitched image with a combined temporal and spatial representation. This two-dimensional stitched image can then represent both the spatial and temporal distribution of signals.

[0056] For example, multiple spatial heat maps taken within a day can be stitched together to obtain a stitched image. This stitched image can be obtained by horizontally stitching together spatial heat maps collected every hour within a day. This stitched image then contains the signal conditions in the railway scenario for that day, and the fault identification model can use this stitched image to determine whether there are network faults in the railway scenario during that day.

[0057] S130, the stitched image and the time-series curve are input into the fault identification model, and the fault identification model identifies the network fault information of the railway scenario based on at least one of the stitched image and the time-series curve.

[0058] In this embodiment, the fault identification model is a large-scale AI model based on computer vision. It can automatically identify network faults such as weak coverage, interference, and handover failures in wireless networks based on at least one of the input spatial heatmaps, stitched images, and time-series curves by learning from a large number of labeled coverage heatmaps, load-to-interference ratio heatmaps, TA heatmaps, and related time-series curves. This model can be built by extending a pre-trained general-purpose vision model (such as YOLO). By adding network structures to the computer vision model and fine-tuning it using labeled data, a dedicated AI model with network problem identification capabilities can be formed.

[0059] During the training of the fault identification model, multiple historical spatial heatmaps and multiple historical time-series curves can be acquired. Then, at least two images are selected from the multiple historical spatial heatmaps and stitched together each time to obtain a historical stitched image. Fault labels are added to the historical spatial heatmaps, historical time-series curves, and historical stitched images based on the fault types of network faults in the multiple historical spatial heatmaps, multiple historical time-series curves, and historical stitched images. The images with added fault labels are determined as the first training sample set. Then, the fault identification model is trained using the first training sample set until the fault identification model meets the preset first convergence condition, at which point the training of the fault identification model is stopped, resulting in a converged fault identification model.

[0060] Then, the stitched image and the time-series curve can be input into the fault identification model, and the fault identification model can identify the network fault information of the railway scenario based on at least one of the stitched image and the time-series curve.

[0061] In this embodiment, multiple time-series thermal images of railway scenes are stitched together to obtain a stitched image. This stitched image and the time-series curve are then input into a fault identification model, enabling the model to identify fault information in the network based on at least one factor from both the stitched image and the time-series curve. Compared to traditional methods relying on manual image-by-image analysis, this approach automatically identifies fault types through a model. Furthermore, by jointly analyzing the stitched thermal image and the time-series curve, the model can quickly and accurately determine fault types in different times and spaces, thereby significantly improving the efficiency of network wireless coverage fault identification and intelligent optimization.

[0062] Specifically, in some embodiments, since a single spatial heatmap reflects the network status along the railway track at a certain moment, the fault identification model can identify transient, location-related instantaneous network faults such as weak coverage and over-coverage through the spatial heatmap. Instantaneous network faults refer to local network performance anomalies reflected by at least one spatial heatmap collected at the same time within a specific moment, such as weak coverage faults, over-coverage faults, or faults caused by sudden interference points, as well as handover failure faults.

[0063] For example, a fault identification model can detect that the signal strength power at a certain location in the coverage heatmap is greater than -75 dBm, and that the TA value of this area is relatively large compared to the area corresponding to the base station, thus identifying it as an over-coverage fault. Alternatively, if the fault identification model detects that the signal strength power in a certain area is less than -95 dBm, accompanied by a high TA value, it can identify it as a weak coverage fault. Furthermore, if the fault identification model detects that the C / I value on a certain trajectory in the carrier-to-interference ratio heatmap is less than 10 dB, it can identify that there is co-channel or adjacent-channel interference fault on that trajectory.

[0064] Furthermore, the fault identification model can also perform joint analysis on multiple spatial heat maps collected at the same time. For example, if the fault identification model finds a sudden drop boundary in the load-to-dryness ratio heat map, and a significant jump at the same point in the TA heat map, with the two boundaries coinciding, then it is marked as a handover failure fault.

[0065] In some specific embodiments, the spatial heatmap includes a coverage heatmap and a TA heatmap, and the transient network fault includes signal quality anomalies and transient signal interference;

[0066] The step of determining the instantaneous network fault within the time corresponding to each spatial heatmap based on each spatial heatmap includes:

[0067] In the railway scenario, if a first region of the coverage heatmap at a first moment is a first color and a first region of the TA heatmap at the first moment is a second color, it is determined that there is a signal quality anomaly in the first region at the first moment. The first color in the coverage heatmap is used to characterize that the signal strength is greater than a first threshold, and the second color in the TA heatmap is used to characterize that the delay is greater than a second threshold.

[0068] If, in the railway scenario, the color in the second region of the coverage heatmap changes abruptly along the railway direction at the second moment, it is determined that there is transient signal interference in the second region at the second moment.

[0069] In this embodiment, the coverage heatmap is a two-dimensional image generated from signal strength measurement data, used to reflect the signal coverage strength at different locations along the railway line. Therefore, a color-coded method can be used to generate the coverage heatmap; for example, red areas in the coverage heatmap represent areas with signal strength greater than -75dBm, and green areas represent areas with signal strength less than -95dBm. The TA heatmap is a two-dimensional image generated from delay measurement data, used to reflect the distribution of signal transmission delay in a railway scenario. Therefore, a color gradient can be used to represent the delay value; for example, red areas in the TA heatmap represent areas with delays exceeding 3ms. Therefore, signal quality anomalies can be identified based on the combination of the first and second colors.

[0070] Specifically, the data for the coverage heatmap and TA heatmap are collected through wireless network measurement equipment. This allows for real-time recording of signal strength and latency information along the railway line, generating the coverage heatmap and TA heatmap. When a first region in the coverage heatmap displays the first color at a certain moment, it indicates that the signal strength in that region exceeds a preset first threshold, which can be -75dBm. Simultaneously, when the first region in the TA heatmap displays the second color, it indicates that the latency in that region exceeds a preset second threshold, which can be 3ms. In this case, due to the unusual combination of high signal strength and high latency, the fault identification model can determine that there is an anomaly in signal quality in that region.

[0071] Furthermore, the fault identification model can also detect color abrupt changes along the railway direction in the coverage heatmap. If a sudden color change occurs in the second region of the coverage heatmap at a second time point, such as a sudden change from red to green within a continuous 100-meter range, it indicates a drastic change in signal strength along that path, possibly caused by transient interference from a train passing through a tunnel or high-voltage power line. By simultaneously analyzing the spatial distribution characteristics and color change patterns of the spatial heatmap, it is possible to accurately distinguish between normal signal fluctuations and anomalies caused by faults.

[0072] Through the above technical solution, this application can accurately identify transient network faults in railway scenarios caused by abnormal base station distribution or dynamic environmental changes. Specifically, the judgment of signal quality anomalies avoids misjudgment based on a single indicator through joint analysis of coverage and TA heatmaps; the detection of transient signal interference solves the problem of insufficient sensitivity of traditional methods in dynamic scenarios by capturing color jumps along the railway direction. This technical solution is particularly suitable for high-speed moving railway environments, effectively distinguishing normal signal fluctuations from real faults, and improving the accuracy and real-time performance of network fault diagnosis.

[0073] In some embodiments, multiple spatial heat maps within a first time period in a railway scenario can be acquired, wherein the first time period includes multiple first time periods. Then, at least two spatial heat maps within each first time period are stitched together to obtain at least one first stitched image. The fault identification model determines the short-term interference fault within the first time period corresponding to the first stitched image based on at least one of the first stitched images and the time-series curve.

[0074] In this embodiment, the first stitched image consists of at least two spatial heat maps collected within a first time period. Multiple spatial heat maps collected in the railway scenario are stitched together in chronological order to form a two-dimensional stitched image representing a combined time-space representation. The first stitched image can be a combined image formed by stitching together multiple spatial heat maps of the same area within consecutive short time periods along a temporal direction. For example, if the first time period is one day, the first stitched image is a stitched image obtained by horizontally stitching together spatial heat maps collected every hour within that day. The fault identification model can use the first stitched image to determine whether a short-term interference fault exists in the railway scenario within the first time period. A short-term interference fault refers to a sudden interference problem that occurs within a relatively short period of time.

[0075] In some specific embodiments, the first stitched image includes a first overlay stitched image, which is obtained by stitching together at least two overlay heat maps within a first time period in chronological order;

[0076] The step of determining the short-term interference fault within the first time period corresponding to each of the first stitched images using the fault identification model includes:

[0077] In the railway scenario, if the color in the third region of the first overlay mosaic changes in the temporal direction, it is determined that there is a short-term interference fault in the third region during the first time period.

[0078] In this embodiment, the first overlay image refers to a two-dimensional image formed by horizontally or vertically stitching together overlay heat maps collected in a continuous first time period under the same railway scene in chronological order. Thus, the first stitched image can reflect both spatial and temporal features.

[0079] Color jumps in the temporal direction refer to discontinuous changes in the signal strength value of a specific third region in a mosaic image over time. If a fault identification model, when analyzing the first coverage mosaic image in a railway scenario, finds a color jump in the third region in the temporal direction—that is, a significant color change in this region between different acquisition times—it indicates a sudden change in the signal strength value within the third region between adjacent acquisition times. Based on this, the fault identification model can determine that a short-term interference fault exists in the third region within the corresponding time period. This color jump phenomenon indicates that the network signal deteriorates rapidly or improves rapidly within a short period, usually lacking continuity, and is often caused by external temporary interference sources (such as temporary electromagnetic equipment or construction interference), thus being identified as a short-term interference fault.

[0080] Specifically, in the railway scenario, mobile terminals encounter transient interference sources during high-speed operation. Conventional single-frame heatmap analysis cannot capture such transient phenomena. By stitching together the coverage heatmaps collected in the first time period to form a two-dimensional image containing time dimension information, the model can extract feature patterns along the temporal direction using a convolutional neural network. When a color change occurs in a certain area of ​​the stitched image, it indicates that a sudden change in signal strength has occurred at that location within the corresponding time period. For example, the detection of alternating red and yellow stripes in the third area of ​​the stitched image indicates that there is a short-term signal fluctuation in that area caused by temporary obstacles or electromagnetic interference. This temporal stitching method allows the model to directly learn the temporal correlation features of interference events, avoiding the complexity of manually extracting temporal features in traditional methods.

[0081] Through the above technical solution, this application solves the problem of difficulty in capturing the dynamic characteristics of short-term interference faults in the time dimension in railway scenarios, and achieves accurate identification of temporary interference events. This method transforms the dynamic process into a two-dimensional spatial pattern through time-series stitching, effectively reducing the complexity of model processing time-series data, improving the real-time performance and accuracy of fault detection, and providing a reliable technical means for rapid fault location in railway mobile communication systems.

[0082] In some embodiments, multiple spatial heat maps within a first time period in a railway scenario can be acquired. The first time period includes at least one second time period, the duration of which is longer than that of the first time period. At least two spatial heat maps within each second time period are stitched together to obtain at least one second stitched image. The base station fault within the second time period corresponding to the second stitched image is determined by the fault identification model based on at least one of the second stitched images and the time series curve.

[0083] In this embodiment, the second stitched image consists of at least two spatial heat maps collected within the second time period. Multiple spatial heat maps collected in the railway scenario are stitched together chronologically to form a two-dimensional stitched image representing a combined time-space representation. The second stitched image can be a combined image formed by stitching multiple spatial heat maps of the same area over a continuous long period along a temporal direction. For example, if the second time period is one month, the second stitched image could be a stitched image obtained by horizontally or vertically stitching together a spatial heat map collected every day within that month. The fault identification model can use the second stitched image to determine whether a base station fault exists in the railway scenario during the second time period. A base station fault refers to a stability problem caused by a decrease in base station power or equipment aging, determined over a long period by observing the continuous signal attenuation trend in the stitched spatial heat maps collected over multiple days, combined with changes in the time-series curve.

[0084] In this embodiment, multiple spatial heat maps and time-series curves within a first time period in a railway scenario can be acquired, and a fault identification model can be used to identify different types of network faults in stages. Specifically, instantaneous network faults can be identified based on a single spatial heat map, or short-term interference faults and base station faults can be identified by stitching together multiple spatial heat maps and combining them with time-series curves. Compared to the traditional method that relies on manual image-by-image analysis, this solution automatically identifies fault types through a model and classifies problems according to the time dimension, making the fault identification process more systematic and automated. Through the joint analysis of spatial heat maps and time-series curves, the model can quickly and accurately determine the fault types at different time scales, thereby significantly improving the efficiency of network fault location and confirmation.

[0085] In some specific embodiments, the second stitched image includes a second coverage stitched image, which is obtained by stitching together at least two coverage heatmaps within a second time period in chronological order, and the base station fault includes base station aging problems;

[0086] The step of determining the base station fault within the second time period corresponding to each of the second stitched images using the fault identification model includes:

[0087] In the railway scenario, if the fourth region, composed of the first color in the second coverage mosaic, shrinks along the time sequence to form an inverted trapezoidal distribution, it is determined that there is a base station aging problem in the second time period.

[0088] In this embodiment, the second coverage mosaic refers to a two-dimensional image formed by stitching together coverage heatmaps continuously collected within a second time period in chronological order along a first direction. For example, spatial heatmaps collected monthly can be arranged from top to bottom, so that the horizontal dimension represents spatial extension and the vertical dimension represents the passage of time. The duration of the second time period is set to be longer than that of the first time period, for example, the second time period can be several weeks or several months, to capture the long-term trend of base station performance. The inverted trapezoidal distribution refers to the first color area exhibiting a geometric shape that is wider at the top and narrower at the bottom in the temporal dimension. This shrinkage feature reflects the physical phenomenon that the coverage area of ​​the base station gradually decreases over time.

[0089] Specifically, the first color in the coverage heatmap is used to characterize high-power areas where the signal strength exceeds a certain threshold. In a railway scenario, because trains move along fixed tracks, the ideal coverage area of ​​a base station should exhibit a continuous, strip-like distribution. As base station equipment ages and its transmission power decreases, its coverage area shrinks daily. By stitching multiple coverage heatmaps from the second time period along a temporal sequence, the shrinkage process of high-power areas is transformed into an inverted trapezoidal geometric feature in a two-dimensional image. This feature differs significantly from the local anomalies caused by random jumps or transient faults resulting from short-term interference, enabling the fault identification model to accurately determine base station aging issues through morphological analysis.

[0090] Compared to existing technologies, traditional methods rely on single heatmaps or short-term statistical data, failing to distinguish between gradual failures caused by equipment aging and transient interference. This solution uses time-series stitching to form a two-dimensional spatiotemporal distribution map covering the heatmap, converting temporal information into spatially identifiable geometric features. This allows for the visualization and accurate detection of long-term performance degradation issues.

[0091] Through the above technical solution, this application achieves automated identification of base station aging problems in railway scenarios, effectively avoiding missed faults caused by the low efficiency of manual inspections. By capturing the inverted trapezoidal distribution characteristics, it can accurately distinguish between equipment aging and short-term interference, providing a reliable basis for base station maintenance decisions and avoiding redundant maintenance costs caused by misjudgment.

[0092] As an optional embodiment, the step of stitching together the multiple spatial heat maps according to their acquisition time to obtain a stitched image includes:

[0093] According to the time sequence of the acquisition time of the multiple spatial heat maps, the multiple spatial heat maps are stitched together along a first direction perpendicular to the railway line to obtain a stitched image, wherein the stitched image includes multiple linear images that extend along the railway line and are arranged in parallel in the time sequence direction.

[0094] In this embodiment, a method for stitching together multiple spatial heat maps specifically includes acquiring the multiple spatial heat maps and the acquisition time of each spatial heat map, and stitching the multiple spatial heat maps together along a first direction according to the time sequence of the acquisition time.

[0095] The acquisition time refers to the timestamp of each spatial heatmap, which can be automatically recorded through system logs or the data acquisition module to ensure the accuracy of the time sequence information. The first direction can be horizontal or vertical. Stitching along the first direction means arranging spatial heatmaps from different times in chronological order of acquisition along the horizontal or vertical direction. For example, heatmaps from adjacent times can be connected sequentially along the horizontal axis to form a continuous time-series stitched image.

[0096] Specifically, this method can establish a temporal correlation between multiple spatial heatmaps by extracting their timestamps within the same time period. For example, in a railway scenario, a coverage heatmap is collected every hour, and then 24 consecutively collected heatmaps are stitched together from top to bottom in chronological order into a single stitched image. This stitched image includes multiple linear images that extend along the railway line and are arranged parallel to each other in the temporal direction.

[0097] Compared to existing technologies, conventional network fault identification methods typically process a single spatial heatmap, failing to capture the temporal correlation of fault features in dynamic scenarios. For example, existing technologies only focus on spatial distribution in heatmap analysis of static scenes, while in railway scenarios, train movement causes signal quality to change rapidly over time. This solution stitches multiple spatial heatmaps together temporally to form a composite image containing temporal evolution information. This composite image is then used as input to a fault identification model, enabling the model to simultaneously analyze both spatial distribution and temporal variation patterns, thus more accurately identifying network faults in railway scenarios.

[0098] Through the above technical solution, this application solves the problem that short-term interference faults in railway scenarios are difficult to capture due to dynamic changes. The time-series stitching method transforms discrete spatial heatmaps into continuous time-series images, enabling the fault identification model to directly learn dynamic features such as signal jumps and diffusion, thereby improving the accuracy of network wireless coverage fault identification and intelligent optimization.

[0099] As an optional embodiment, the step of stitching together the multiple spatial heat maps according to their acquisition time to obtain a stitched image includes:

[0100] For any pixel in any spatial heatmap, obtain the position information and color information of the pixel;

[0101] Obtain the acquisition time of the spatial heatmap of the pixel;

[0102] The feature vector of the pixel is determined based on the acquisition time of the spatial heat map, the position information and color information of the pixel;

[0103] The feature vectors of all pixels in the multiple spatial heatmaps are fused to obtain the stitched image.

[0104] In this embodiment, in the original spatial heatmap, each pixel typically consists of its two-dimensional location information (X, Y) and color information (such as RGB values) forming a three-dimensional feature vector, used to represent the point's geographical location on the railway line and its corresponding network signal status (such as coverage strength, interference level, etc.). However, traditional three-dimensional features cannot reflect the image acquisition time, and therefore cannot reflect the evolution of network status over time.

[0105] Therefore, for each pixel in each spatial heatmap, based on the original (X, Y, RGB) features, the temporal information can be encoded into the pixel's color channel by combining the acquisition time of the spatial heatmap to which the pixel belongs. For example, the acquisition time can be normalized and then fused or concatenated with the RGB channels. This results in a feature vector (X, Y, RGB_T) that is still three-dimensional but implicitly contains the temporal dimension. This feature vector not only describes the spatial location and signal state of the pixel but also expresses the time point in time when that state occurred.

[0106] Finally, the three-dimensional feature vectors of all pixels in the spatial heatmaps collected from multiple different times can be fused to obtain a matrix, which is a stitched image that can comprehensively express spatial location, signal strength and temporal changes.

[0107] The above scheme integrates spatial, color, and temporal information into a unified stitched image, enabling a spatiotemporal representation of network status in railway scenarios. This helps the model accurately identify network faults that evolve over time, improving the accuracy and efficiency of fault diagnosis.

[0108] In some embodiments, before inputting the stitched image and the time-series curve into a fault identification model, and before the fault identification model identifies network fault information of the railway scenario based on at least one of the stitched image and the time-series curve, the method further includes:

[0109] Obtain a trained computer vision model;

[0110] By adding a network structure to the input of the computer vision model, the computer vision model is fine-tuned to obtain a fault identification model.

[0111] The fault identification model is trained to obtain a converged fault identification model.

[0112] In this embodiment, the trained computer vision model refers to a deep neural network model such as YOLO or CNN that has the ability to extract general image features. Its main function is to extract the spatial structural features of the image.

[0113] However, since computer vision models are natively designed for static images, they cannot directly adapt to the stitched images with fused temporal information in this application. Therefore, structural fine-tuning is necessary. Specifically, a new network structure, such as a temporal embedding module or a multi-channel fusion layer, can be added to the input of the fault recognition model to encode the acquisition time of each image. This new structure, together with the original image features, forms a temporally enhanced input vector, thereby enabling the fault recognition model to perceive spatiotemporal joint features. The fine-tuned fault recognition model is then trained until convergence.

[0114] In this way, by fine-tuning the input of the model, the fault identification model can be equipped with the ability to process temporal spatial images, which can significantly improve the accuracy of fault identification and intelligent optimization of dynamic wireless coverage in railway networks.

[0115] As an optional embodiment, after inputting the stitched image into the fault identification model to obtain the network fault information of the railway scene output by the fault identification model, the method further includes:

[0116] If the network fault information includes at least two network faults in the railway scenario, the at least two network faults are displayed, wherein the network faults include transient network faults, short-term interference faults, and base station faults;

[0117] Receive a first input for the first network fault among the at least two network faults;

[0118] In response to the first input, the first network fault corresponding to the first input is input into the scheme prediction model, and the target solution for the first network fault is determined through the scheme prediction model.

[0119] In this embodiment, the solution prediction model refers to a machine learning model built based on expert experience and historical optimization cases. Specifically, network faults in text form can be input into the solution prediction model, which will analyze the input network faults and output the target solutions. During the training process of the solution prediction model, a second training sample set can be obtained. This second training sample set includes multiple historical network fault data and the corresponding solutions for each historical network fault. The solution prediction model is trained using the second training sample set until it meets a preset second convergence condition, at which point training stops, resulting in a converged solution prediction model.

[0120] In this embodiment, when the fault identification model detects a network fault in the current railway scenario, it can generate a text description of the network fault, i.e., network fault information. If the network fault information includes only one network fault, it can be directly input into the solution prediction model, which will then input a solution for that network fault.

[0121] If the network fault information includes at least two network faults, then the at least two network faults can be displayed through a visual interface. The user can select the first network fault to be processed from the displayed at least two network faults by means of a first input, wherein the first input can be a touch operation or a voice command.

[0122] After the user selects the first network fault through the first input, the fault information of the selected first network fault can be transformed into a feature vector and then input into the solution prediction model. The solution prediction model can perform solution matching based on the fault information of the first network fault and generate the target solution for the first network fault.

[0123] Compared to existing technologies, traditional solutions rely on manual fault screening and experience manuals for solution development, resulting in low processing efficiency and poor adaptability. This solution, through automated fault information integration and intelligent model prediction, achieves rapid response to multiple concurrent faults in dynamic railway scenarios, avoiding the problem of easily overlooked temporally related faults in manual processing.

[0124] Through the above technical solution, this application effectively solves the problem of collaborative processing of multi-source network faults in railway scenarios, significantly improving the response speed and accuracy of network optimization in complex dynamic environments. This solution ensures the complete preservation of multi-dimensional fault characteristics through structured fault information integration, optimizes fault processing priorities using an interactive selection mechanism, and guarantees the practicality and scenario adaptability of the solution by combining a prediction model trained with railway-specific features.

[0125] As an optional embodiment, after determining the target solution for the first network fault through the scheme prediction model, the method further includes:

[0126] A simulation scenario is constructed based on the target solution and the railway scenario;

[0127] The propagation of wireless signals under the simulated scenario is simulated to obtain a simulated spatial heat map and a simulated time series curve under the simulated scenario;

[0128] If the fault identification model determines that there is no network fault in the simulation scenario based on the simulated spatial heatmap and the simulated time series curve, it outputs the target solution to the user.

[0129] In this embodiment, after determining the target solution for the first network fault through the scheme prediction model, a simulation scenario can be constructed based on the target solution and the railway scenario. The propagation of wireless signals under the simulation scenario is simulated to obtain a simulated spatial heat map and a simulated time series curve. If the fault identification model confirms that there is no network fault in the simulation scenario based on the simulation data, the target solution output by the scheme prediction model is considered to be effective, and therefore the target solution can be output to the user.

[0130] Among them, the simulation scenario refers to a virtual test environment constructed based on the environmental characteristics of the target solution and the actual railway scenario. Specifically, it can be realized by using a three-dimensional geographic information system combined with base station layout data, and by importing data on railway track alignment, train speed and distribution of surrounding buildings to construct a dynamic test scenario.

[0131] Specifically, once the target solution is generated, the system automatically extracts base station location parameters, train trajectory data, and environmental obstacle data from the railway scenario to construct a linear coverage simulation scenario that includes time-varying parameters. The system then uses a ray tracing algorithm to simulate the signal propagation process along the railway track under optimized parameters, generating a simulated spatial heatmap and simulated time-series curve consistent with the measured data format. A fault identification model performs a secondary analysis of the simulated spatial heatmap and simulated time-series curve. If no fault characteristics such as over-coverage or interference are detected, the target solution is deemed effective.

[0132] Compared with existing technologies, traditional optimization schemes rely solely on empirical parameter adjustments and lack verification mechanisms, making them prone to failure due to dynamic environmental changes. This solution constructs a dedicated railway simulation environment, employs high-precision channel modeling technology to recreate signal propagation characteristics, and combines this with a fault identification model to form a closed-loop verification mechanism, effectively avoiding the subjectivity of manual debugging and the blindness of traditional algorithms.

[0133] Through the above technical solution, this application can ensure the validity verification of the generated optimization parameters before actual deployment in complex railway scenarios, solve the problem of insufficient reliability of optimization suggestions in the prior art, prevent network performance degradation caused by improper parameter settings, and improve the security and decision credibility of the automated optimization system.

[0134] Based on the wireless coverage fault identification and intelligent optimization method provided in the above embodiments, this application also provides specific implementation methods of the wireless coverage fault identification and intelligent optimization device. Please refer to the following embodiments.

[0135] First see Figure 2 The wireless coverage fault identification and intelligent optimization device 200 provided in this application embodiment includes the following modules:

[0136] The first acquisition module 201 is used to acquire multiple spatial heat maps and multiple time-series curves in the railway scenario;

[0137] The stitching module 202 is used to stitch together the multiple spatial heat maps according to the acquisition time of the multiple spatial heat maps to obtain a stitched image;

[0138] The first identification module 203 is used to input the stitched image and the time-series curve into the fault identification model, and to identify the network fault information of the railway scenario based on at least one of the stitched image and the time-series curve.

[0139] The device can time-series stitch together multiple spatial heat maps of railway scenes collected over time to obtain a stitched image. This stitched image, along with the time-series curve, is then input into a fault identification model. The model can then identify fault information in the network based on at least one factor from either the stitched image or the time-series curve. Compared to traditional methods that rely on manual, image-by-image analysis, this solution automatically identifies fault types through a model. Furthermore, by jointly analyzing the stitched heat map and the time-series curve, the model can quickly and accurately determine fault types in different times and spaces, thereby significantly improving the efficiency of fault identification and intelligent optimization for network wireless coverage.

[0140] As one implementation of this application, the above-mentioned splicing module 202 can be specifically used for:

[0141] According to the time sequence of the acquisition time of the multiple spatial heat maps, the multiple spatial heat maps are stitched together along a first direction perpendicular to the railway line to obtain a stitched image, wherein the stitched image includes multiple linear images that extend along the railway line and are arranged in parallel in the time sequence direction.

[0142] As one implementation of this application, the splicing module 202 described above can also be specifically used for:

[0143] For any pixel in any spatial heatmap, obtain the position information and color information of the pixel;

[0144] Obtain the acquisition time of the spatial heatmap of the pixel;

[0145] The feature vector of the pixel is determined based on the acquisition time of the spatial heat map, the position information and color information of the pixel;

[0146] The feature vectors of all pixels in the multiple spatial heatmaps are fused to obtain the stitched image.

[0147] As one implementation of this application, the wireless coverage fault identification and intelligent optimization device 200 may further include:

[0148] The second acquisition module is used to acquire the trained computer vision model;

[0149] The fine-tuning module is used to fine-tune the computer vision model by adding a network structure to the input end of the computer vision model to obtain a fault recognition model.

[0150] The training module is used to train the fault identification model to obtain a fault identification model that has been trained to convergence.

[0151] As one implementation of this application, the wireless coverage fault identification and intelligent optimization device 200 may further include:

[0152] The display module is used to display at least two network faults in the railway scenario when the network fault information includes at least two network faults, wherein the network faults include transient network faults, short-term interference faults, and base station faults.

[0153] The receiving module is configured to receive a first input regarding the first network fault among the at least two network faults;

[0154] The prediction module is used to respond to the first input, input the first network fault corresponding to the first input into the scheme prediction model, and determine the target solution for the first network fault through the scheme prediction model.

[0155] As one implementation of this application, the wireless coverage fault identification and intelligent optimization device 200 may further include:

[0156] A construction module is used to build a simulation scenario based on the target solution and the railway scenario;

[0157] The simulation module is used to simulate the propagation of wireless signals in the simulation scenario and obtain a simulated spatial heat map and a simulated time series curve in the simulation scenario.

[0158] The output module is used to output the target solution to the user when the fault identification model determines that there is no network fault in the simulation scenario based on the simulated spatial heat map and the simulated time series curve.

[0159] The wireless coverage fault identification and intelligent optimization device provided in this embodiment of the invention can implement the steps in the above method embodiments, and will not be repeated here to avoid repetition.

[0160] Figure 3 A schematic diagram of the hardware structure of the wireless coverage fault identification and intelligent optimization device provided in an embodiment of this application is shown.

[0161] The wireless coverage fault identification and intelligent optimization device may include a processor 1001 and a memory 1002 storing computer program instructions.

[0162] Specifically, the processor 1001 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0163] Memory 1002 may include mass storage for data or instructions. For example, and not limitingly, memory 1002 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 1002 may include removable or non-removable (or fixed) media. Where appropriate, memory 1002 may be internal or external to the integrated gateway disaster recovery device. In a particular embodiment, memory 1002 is non-volatile solid-state memory.

[0164] Memory may include read-only memory (ROM), random access memory (RAM), disk storage media devices, optical storage media devices, flash memory devices, and electrical, optical, or other physical / tangible memory storage devices. Therefore, typically, memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the methods according to one aspect of this disclosure.

[0165] The processor 1001 reads and executes computer program instructions stored in the memory 1002 to implement any of the wireless coverage fault identification and intelligent optimization methods in the above embodiments.

[0166] In one example, the wireless coverage fault identification and intelligent optimization device may further include a communication interface 1003 and a bus 1010. For example, Figure 3 As shown, the processor 1001, memory 1002, and communication interface 1003 are connected through bus 1010 and complete communication with each other.

[0167] The communication interface 1003 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.

[0168] Bus 1010 includes hardware, software, or both, that couples components of a wireless coverage fault identification and intelligent optimization device together. For example, and not limitingly, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 1010 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, this application contemplates any suitable bus or interconnect.

[0169] The wireless coverage fault identification and intelligent optimization device can be based on the above embodiments to realize the combination of the above-described wireless coverage fault identification and intelligent optimization method and apparatus.

[0170] Furthermore, in conjunction with the wireless coverage fault identification and intelligent optimization methods in the above embodiments, this application embodiment can provide a computer storage medium for implementation. This computer storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement any one of the wireless coverage fault identification and intelligent optimization methods in the above embodiments and achieve the same technical effect. To avoid repetition, further details are omitted here. The aforementioned computer-readable storage medium may include non-transitory computer-readable storage media, such as read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks, etc., and is not limited thereto.

[0171] In addition, this application also provides a computer program product, including computer program instructions, which, when executed by a processor, can implement the steps and corresponding content of the aforementioned method embodiments.

[0172] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0173] The functional blocks shown in the above block diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. Programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0174] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0175] The aspects of this disclosure have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus, and computer program products according to embodiments of this disclosure. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by special-purpose hardware performing the specified functions or actions, or can be implemented by a combination of special-purpose hardware and computer instructions.

[0176] The above are merely specific embodiments of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the protection scope of this application.

Claims

1. A method for wireless coverage fault identification and intelligent optimization, characterized in that, The method includes: Obtain multiple spatial heat maps and multiple time-series curves in the railway scenario; Based on the acquisition time of the multiple spatial heat maps, the multiple spatial heat maps are stitched together to obtain a stitched image. The stitched image is a two-dimensional image formed by stitching together the spatial heat maps acquired in the same railway scene in a continuous time period in chronological order, either horizontally or vertically. The stitched image and the time-series curve are input into the fault identification model, and the fault identification model identifies network fault information of the railway scenario based on the stitched image and the time-series curve. The spatial heatmap includes coverage heatmap and TA heatmap; the network faults include transient network faults, short-term interference faults, and base station faults; the transient network faults include abnormal signal quality and transient signal interference; and the base station faults include base station aging issues. The step of identifying network fault information in the railway scenario based on the stitched image and the time-series curve includes: In the railway scenario, if a first region of the coverage heatmap at a first moment is a first color and a first region of the TA heatmap at the first moment is a second color, it is determined that there is a signal quality anomaly in the first region at the first moment. The first color in the coverage heatmap is used to characterize that the signal strength is greater than a first threshold, and the second color in the TA heatmap is used to characterize that the delay is greater than a second threshold. If, in the railway scenario, the color in the second region of the coverage heatmap changes abruptly along the railway direction at the second moment, it is determined that there is transient signal interference in the second region at the second moment. In the railway scenario, if the color in the third region of the first coverage mosaic map changes chronologically, it is determined that a short-term interference fault exists in the third region during the first time period; the first coverage mosaic map is obtained by chronologically stitching together at least two coverage heatmaps during the first time period. In the railway scenario, if the fourth region composed of the first color in the second coverage mosaic map shrinks along the time sequence to form an inverted trapezoidal distribution, it is determined that there is a base station aging problem in the second time period; the second coverage mosaic map is obtained by stitching together at least two coverage heat maps in the second time period in time sequence. The step of stitching together the multiple spatial heat maps according to their acquisition time to obtain a stitched image includes: According to the time sequence of the acquisition time of the multiple spatial heat maps, the multiple spatial heat maps are stitched together along a first direction perpendicular to the railway line to obtain a stitched image, wherein the stitched image includes multiple linear images that extend along the railway line and are arranged in parallel in the time sequence direction.

2. The wireless coverage fault identification and intelligent optimization method according to claim 1, characterized in that, The step of stitching together the multiple spatial heat maps according to their acquisition time to obtain a stitched image includes: For any pixel in any spatial heatmap, obtain the position information and color information of the pixel; Obtain the acquisition time of the spatial heatmap of the pixel; The feature vector of the pixel is determined based on the acquisition time of the spatial heat map, the position information and color information of the pixel; The feature vectors of all pixels in the multiple spatial heatmaps are fused to obtain the stitched image.

3. The wireless coverage fault identification and intelligent optimization method according to claim 1, characterized in that, Before inputting the stitched image and the time-series curve into the fault identification model, and before the fault identification model identifies the network fault information of the railway scenario based on the stitched image and the time-series curve, the method further includes: Obtain a trained computer vision model; By adding a network structure to the input of the computer vision model, the computer vision model is fine-tuned to obtain a fault identification model. The fault identification model is trained to obtain a converged fault identification model.

4. The wireless coverage fault identification and intelligent optimization method according to claim 1, characterized in that, After inputting the stitched image into the fault identification model to obtain the network fault information of the railway scene output by the fault identification model, the method further includes: If the network fault information includes at least two network faults in the railway scenario, the at least two network faults are displayed, wherein the network faults include transient network faults, short-term interference faults, and base station faults; Receive a first input for the first network fault among the at least two network faults; In response to the first input, the first network fault corresponding to the first input is input into the scheme prediction model, and the target solution for the first network fault is determined through the scheme prediction model.

5. The wireless coverage fault identification and intelligent optimization method according to claim 4, characterized in that, After determining the target solution for the first network fault using the scheme prediction model, the method further includes: A simulation scenario is constructed based on the target solution and the railway scenario; The propagation of wireless signals under the simulated scenario is simulated to obtain a simulated spatial heat map and a simulated time series curve under the simulated scenario; If the fault identification model determines that there is no network fault in the simulation scenario based on the simulated spatial heatmap and the simulated time series curve, it outputs the target solution to the user.

6. A wireless coverage fault identification and intelligent optimization device, characterized in that, The device includes: The first acquisition module is used to acquire multiple spatial heat maps and multiple time-series curves in the railway scenario; The stitching module is used to stitch together the multiple spatial heat maps according to the acquisition time of the multiple spatial heat maps to obtain a stitched image. The stitched image is a two-dimensional image formed by stitching together the spatial heat maps acquired in the same railway scene in a continuous time period in chronological order, either horizontally or vertically. The first identification module is used to input the stitched image and the time-series curve into the fault identification model, and to identify network fault information of the railway scenario through the fault identification model based on the stitched image and the time-series curve. The spatial heatmap includes a coverage heatmap and a TA heatmap. The network faults include transient network faults, short-term interference faults, and base station faults. The transient network faults include signal quality anomalies and transient signal interference. The base station faults include base station aging issues. The first identification module is specifically used to: determine that there is a signal quality anomaly in the first area at the first moment when, in the railway scenario, a first area of ​​the coverage heatmap at a first moment is a first color, and a first area of ​​the TA heatmap at the first moment is a second color; wherein, the first color in the coverage heatmap is used to characterize a signal strength greater than a first threshold, and the second color in the TA heatmap is used to characterize a delay greater than a second threshold; determine that there is transient signal interference in the second area at the second moment when, in the railway scenario, a color jumps along the railway direction in the second area of ​​the coverage heatmap at a second moment; determine that there is a short-term interference fault in the third area within the first time period when, in the railway scenario, a color jumps in the temporal direction in the third area of ​​the first coverage stitched image; the first coverage stitched image is obtained by stitching at least two coverage heatmaps within the first time period in temporal order. In the railway scenario, if the fourth region composed of the first color in the second coverage mosaic map shrinks along the time sequence to form an inverted trapezoidal distribution, it is determined that there is a base station aging problem in the second time period; the second coverage mosaic map is obtained by stitching together at least two coverage heat maps in the second time period in time sequence. The stitching module is specifically used to: stitch the multiple spatial heat maps together along a first direction perpendicular to the railway line according to the time sequence of their acquisition time, to obtain a stitched image, wherein the stitched image includes multiple linear images that extend along the line direction and are arranged parallel to each other in the time sequence direction.

7. A wireless coverage fault identification and intelligent optimization device, characterized in that, The wireless coverage fault identification and intelligent optimization device includes: a processor and a memory storing computer program instructions; When the processor executes the computer program instructions, it implements the wireless coverage fault identification and intelligent optimization method as described in any one of claims 1-5.

8. A computer storage medium, characterized in that, The computer storage medium stores computer program instructions, which, when executed by a processor, implement the wireless coverage fault identification and intelligent optimization method as described in any one of claims 1-5.

9. A computer program product, characterized in that, The computer program product includes computer program instructions, which, when executed by a processor, implement the wireless coverage fault identification and intelligent optimization method as described in any one of claims 1-5.

Images