CN120812631A

CN120812631A - Method and device for wireless coverage fault identification and intelligent optimization

Info

Publication number: CN120812631A
Application number: CN202511241655.2A
Authority: CN
Inventors: 官科; 何丹萍; 于俊逸
Original assignee: Beijing Qianjing Technology Co ltd; Beijing Jiaotong University
Current assignee: Beijing Qianjing Technology Co ltd; Beijing Jiaotong University
Priority date: 2025-09-02
Filing date: 2025-09-02
Publication date: 2025-10-17

Abstract

The application discloses a method and a device for identifying wireless coverage faults and intelligently optimizing the wireless coverage faults, and belongs to the technical field of wireless networks. The method comprises the steps of obtaining a plurality of space heat maps and a plurality of time sequence graphs under a railway scene, splicing the plurality of space heat maps according to the acquisition time of the plurality of space heat maps to obtain spliced images, inputting the spliced images and the time sequence graphs into a fault identification model, and identifying network fault information of the railway scene according to at least one of the spliced images and the time sequence graphs through the fault identification model. According to the embodiment of the application, the efficiency of network fault determination can be improved.

Description

Method and device for wireless coverage fault identification and intelligent optimization

Technical Field

The application belongs to the technical field of wireless networks, and particularly relates to a method and a device for wireless coverage fault identification and intelligent optimization.

Background

With the development of society, the demand of users for networks is continuously increased, and meanwhile, the probability of network faults in the using process is increased.

However, in the prior art, it is required that an operator of a network recognize a network failure in a railway scene from a network performance image such as a space heat map, a time sequence graph, etc. of the railway scene. Therefore, the discovery of network faults has great dependence on the experience of operation and maintenance personnel, and meanwhile, the operation and maintenance personnel have low efficiency of determining the faults because a great amount of time is required to be spent in determining the network faults.

Disclosure of Invention

The embodiment of the application provides a method and a device for identifying and intelligently optimizing wireless coverage faults, which can solve the technical problem of low determination efficiency of the existing network faults.

In a first aspect, an embodiment of the present application provides a method for identifying and intelligently optimizing a wireless coverage failure, where the method includes:

acquiring a plurality of space heat maps and a plurality of time sequence graphs in a railway scene;

Splicing the plurality of space heat maps according to the acquisition time of the plurality of space heat maps to obtain a spliced image;

inputting the spliced image and the time sequence curve graph into a fault identification model, and identifying network fault information of the railway scene according to at least one of the spliced image and the time sequence curve graph through the fault identification model.

In some embodiments, the stitching the plurality of space heat maps according to the acquisition time of the plurality of space heat maps to obtain a stitched image includes:

And splicing the plurality of space heat maps along a first direction perpendicular to the railway line according to the time sequence of the acquisition time of the plurality of space heat maps to obtain spliced images, wherein the spliced images comprise a plurality of linear images which extend along the line direction and are arranged in parallel in the time sequence direction.

For any pixel point in any space heat map, acquiring the position information and the color information of the pixel point;

Acquiring acquisition time of a space heat map where the pixel points are located;

determining a feature vector of the pixel point according to the acquisition time of the space heat map, the position information and the color information of the pixel point;

And fusing the feature vectors of all the pixel points in the plurality of space heat maps to obtain the spliced image.

In some embodiments, before the inputting the stitched image and the time series plot into a fault identification model, identifying network fault information of the railway scene from at least one of the stitched image and the time series plot by the fault identification model, the method further comprises:

acquiring a trained computer vision model;

Adding a network structure at the input end of the computer vision model to finely tune the computer vision model to obtain a fault identification model;

and training the fault recognition model to obtain the fault recognition model trained to be converged.

In some embodiments, after the 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:

Displaying at least two network faults in the railway scene under the condition that the network fault information comprises the at least two network faults, wherein the network faults comprise instantaneous network faults, short-time interference faults and base station faults;

Receiving a first input for a first network failure of the at least two network failures;

and responding to the first input, inputting a first network fault corresponding to the first input into a scheme prediction model, and determining a target solution of the first network fault through the scheme prediction model.

In some embodiments, after the determining the target solution to the first network failure by the solution prediction model, the method further comprises:

constructing a simulation scene according to the target solution and the railway scene;

Simulating the propagation of the wireless signal in the simulation scene to obtain a simulation space heat map and a simulation time sequence curve diagram in the simulation scene;

And outputting the target solution to a user under the condition that the fault identification model determines that no network fault exists in the simulation scene according to the simulated space heat diagram and the simulated time sequence diagram.

In a second aspect, an embodiment of the present application provides a device for identifying and intelligently optimizing a wireless coverage failure, where the device includes:

The first acquisition module is used for acquiring a plurality of space heat maps and a plurality of time sequence graphs in a railway scene;

the splicing module is used for splicing the plurality of space heat maps according to the acquisition time of the plurality of space heat maps to obtain spliced images;

The first recognition module is used for inputting the spliced image and the time sequence curve graph into a fault recognition model, and recognizing network fault information of the railway scene according to at least one of the spliced image and the time sequence curve graph through the fault recognition model.

In a third aspect, an embodiment of the present application provides a wireless coverage fault identification and intelligent optimization device, the device including a processor and a memory storing computer program instructions;

the wireless coverage fault identification and intelligent optimization method is realized when the processor executes the computer program instructions.

In a fourth aspect, an embodiment of the present application provides a computer storage medium, where computer program instructions are stored, where the computer program instructions, when executed by a processor, implement a method for identifying and intelligently optimizing a wireless coverage failure as described above.

In a fifth aspect, embodiments of the present application provide a computer program product comprising computer program instructions which, when executed by a processor, implement a wireless coverage fault identification and intelligent optimization method as above.

According to the application, a plurality of railway scene space heat maps acquired according to time are subjected to time sequence splicing to obtain spliced images, and the spliced images and the time sequence curve graph are input into a fault identification model together, so that the fault identification model identifies fault information in a network according to at least one of the spliced images and the time sequence curve graph. Compared with the traditional mode of manually analyzing by graph, the scheme automatically identifies the fault type through the model, and the model can rapidly and accurately judge the fault type under different time and space through joint analysis of the spliced heat map and the time sequence graph, so that the network wireless coverage fault identification and intelligent optimization efficiency is remarkably improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the embodiments of the present application will be briefly described below, and it is obvious that the drawings described below are only some embodiments of the present application, and other drawings may be obtained according to these drawings without inventive effort to a person of ordinary skill in the art.

Fig. 1 is a schematic flow chart of a wireless coverage fault recognition and intelligent optimization method according to an embodiment of the present application;

FIG. 2 is a schematic diagram of a wireless coverage fault recognition and intelligent optimization device according to an embodiment of the present application;

fig. 3 is a schematic hardware structure of a wireless coverage fault recognition and intelligent optimization device according to an embodiment of the present application.

Detailed Description

Features and exemplary embodiments of various aspects of the present application will be described in detail below, and in order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be described in further detail below with reference to the accompanying drawings and the detailed embodiments. It should be understood that the particular embodiments described herein are meant to be illustrative of the application only and not limiting. It will be apparent to one skilled in the art that the present application may be practiced without some of these specific details. The following description of the embodiments is merely intended to provide a better understanding of the application by showing examples of the application.

It is noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising" does not exclude the presence of additional identical elements in a process, method, article, or apparatus that comprises an element.

It should be noted that, without conflict, the embodiments of the present application and features of the embodiments may be combined with each other. The embodiments will be described in detail below with reference to the accompanying drawings.

Current wireless network optimization mainly relies on two ways of expert experience and intelligent optimization algorithm. The expert experience method carries out network adjustment through the dynamic detection vehicle measurement data and the experience model, has the problems of low efficiency, unstable optimization effect and the like, and is difficult to quickly respond to the dynamic change of the network environment. The intelligent optimization algorithm performs global statistical feature optimization based on algorithms such as particle swarm and ant colony, but cannot accurately process specific problems such as over coverage and weak coverage caused by signal difference among base stations, and has limited optimization effect in railway scenes.

The railway scene has unique network characteristics such as long-strip coverage, doppler effect caused by high-speed movement of the train, frequent base station switching and the like. These features make conventional network fault diagnosis methods have significant challenges, on one hand, fault judgment logic based on key signaling interaction in static scenes is not applicable in railway scenes, and on the other hand, signal abnormal modes in motion states are difficult to effectively identify by conventional thermodynamic diagram analysis methods.

The prior art has the following outstanding problems that firstly, the special signal abnormal mode of a railway scene, such as the difference between signal attenuation and real network faults caused by Doppler effect, cannot be accurately distinguished, secondly, the joint analysis capability of time sequence characteristics is lacking, progressive faults such as base station aging and the like are difficult to identify, thirdly, the solution generation process is disjointed with fault diagnosis, and closed loop optimization cannot be realized.

Specifically, in order to solve the problems in the prior art, the embodiment of the application provides a method and a device for wireless coverage fault identification and intelligent optimization. The following first describes a wireless coverage fault identification and intelligent optimization method provided by the embodiment of the application.

Fig. 1 is a schematic flow chart of a wireless coverage fault recognition and intelligent optimization method according to an embodiment of the present application. The method comprises the following steps:

S110, acquiring a plurality of space heat maps and a plurality of time sequence graphs in a railway scene.

In this embodiment, the space heat map refers to a two-dimensional image generated by measuring data through wireless signals, and is used for representing spatial distribution of signal intensity or time delay, where the space heat map may include a coverage heat map, a carrier-to-heat map, and a TA (TIMING ADVANCE ) heat map, where the coverage heat map may be used to represent spatial distribution of signal receiving intensity (RSRP) at each position along a railway, and reflect signal coverage with color, the carrier-to-heat map may be used to represent signal quality (C/I, i.e., carrier-to-interference ratio) distribution at each position along the railway, and be used to determine interference intensity, and the TA heat map may be used to represent time delay between each position along the railway and a base station, and reflect distance between a user and the base station or propagation delay.

The time sequence graph refers to a line graph showing the change trend of a certain network performance index with time or distance, and is used for analyzing the dynamic fluctuation of the network state. The timing graphs may include coverage graphs, carrier to interference ratio graphs, TA graphs. The coverage graph is used for representing the change trend of the signal intensity with time or distance in the running process of the train and analyzing the coverage stability. The carrier-to-interference ratio graph can be used for showing the change trend of signal quality (C/I) along with the running path of the train and identifying the interference fluctuation point. The TA graph can be used for showing the TA value change trend of the train in the running process and judging the distance change or the switching abnormality.

The space heat map under the railway scene refers to the space heat map of a single line segment acquired along the railway track, and reflects the distribution condition of network performance indexes at various positions at a certain moment when the train runs on the railway. The network performance indicators may include, among other things, signal coverage strength, interference level, TA value, etc.

Because each space heat map corresponds to different acquisition moments, the acquisition track of each space heat map is fixed, the format is uniform, and the time sequence characteristics are provided, a plurality of space heat maps acquired under a railway scene can be spliced in time sequence, so that a two-dimensional splice map which is represented by time-space combination is formed, and the two-dimensional splice map can be used for observing the evolution of a network state along with time. The first spliced image and the second spliced image are the two-dimensional spliced image.

S120, splicing the plurality of space heat maps according to the acquisition time of the plurality of space heat maps to obtain a spliced image;

In this embodiment, since the single Zhang Kongjian heat map is often a space heat map of a single line segment collected along a railway track, the distribution situation of network performance indexes at various positions at a certain moment when a train runs on the railway is reflected. Therefore, a single space heat map can only express signal distribution in space, and a plurality of space heat maps acquired in a railway scene can be spliced in time sequence according to the acquisition time of the plurality of space heat maps, so that a two-dimensional spliced image of time-space joint representation is formed. The two-dimensional stitched image may represent both the spatial and the temporal distribution of the signal.

For example, a plurality of space heat maps in a day may be spliced to obtain a spliced image, where the plurality of space heat maps may be spliced images obtained by transversely splicing the space heat maps acquired every 1 hour in a day. Then, the spliced image contains the signal condition of the railway scene in the day, and the fault identification model can judge whether the network fault exists in the railway scene in the day through the spliced image.

S130, inputting the spliced image and the time sequence curve graph into a fault recognition model, and recognizing network fault information of the railway scene according to at least one of the spliced image and the time sequence curve graph through the fault recognition model.

In this embodiment, the fault identification model is an AI large model based on computer vision, and can automatically identify network faults such as weak coverage, interference, switching failure and the like in the wireless network based on at least one of the input space heat map, the spliced image and the time sequence graph by learning a large number of marked coverage heat maps, carrier heat maps, TA heat maps and related time sequence graphs. The model can be constructed by extending a pre-trained general vision model (such as YOLO), and a special AI model with network problem identification capability is finally formed by adding a network structure into a computer vision model and performing fine tuning training by combining labeling data.

In the training process of the fault identification model, a plurality of historical space heat maps and a plurality of historical time sequence graphs can be obtained, then at least two historical spliced images are selected from the plurality of historical space heat maps each time to obtain a historical spliced image, fault labels are added to the historical space heat maps, the historical time sequence graphs and the historical spliced image according to the plurality of historical space heat maps, the plurality of historical time sequence graphs and the fault types of network faults in the historical spliced image, the images added with the fault labels are determined to be a first training sample set, then the fault identification model is trained by adopting the first training sample set until the fault identification model meets a preset first convergence condition, and training of the fault identification model is stopped to obtain the fault identification model trained to convergence.

The stitched image and the time series graph may then be input into a fault identification model by which network fault information of the railway scene is identified from at least one of the stitched image and the time series graph.

In this embodiment, by performing time sequence stitching on a plurality of railway scene space heat maps acquired in time, a stitched image is obtained, and the stitched image and the time sequence graph are input to a fault identification model together, so that the fault identification model identifies fault information in a network according to at least one of the stitched image and the time sequence graph. Compared with the traditional mode of manually analyzing by graph, the scheme automatically identifies the fault type through the model, and the model can rapidly and accurately judge the fault type under different time and space through joint analysis of the spliced heat map and the time sequence graph, so that the network wireless coverage fault identification and intelligent optimization efficiency is remarkably improved.

Specifically, in some embodiments, because the single Zhang Kongjian heatmap reflects the network state along the railroad track at a certain point in time, the fault identification model may identify transient, location-dependent transient network faults such as weak coverage, over-coverage, etc. through the spatial heatmap. Transient network faults refer to local network performance anomalies reflected by at least one spatial heat map with the same acquisition time within a specific time, such as weak coverage faults, over coverage faults or interference burst point faults, switching failure faults and the like.

For example, the fault recognition model may detect that the signal strength power of a certain position in the coverage heat map is greater than-75 dBm, and the TA value of the area far from the corresponding area of the base station is greater, so that an over-coverage fault can be determined, or the fault recognition model may detect that if the signal strength power of a certain area is less than-95 dBm, and accompanied with a high TA value, a weak-coverage fault can be determined. The fault identification model can also detect that the C/I value on a certain section of track in the carrier-to-interference heat map is lower than 10dB, and can judge that the section of track has the same-frequency or adjacent-frequency interference fault.

In addition, the fault identification model can perform joint analysis on a plurality of space heat maps acquired at the same moment. For example, the fault recognition model finds that a dip boundary exists in the carrier-to-carrier heat map, meanwhile, a significant jump exists in the TA heat map, and the two boundaries coincide, so that the fault is marked as a switching failure fault.

In some specific embodiments, the spatial heat map comprises a coverage heat map and a TA heat map, and the transient network failure comprises signal quality anomalies and transient signal disturbances;

The determining, according to each space heat map, the transient network fault in the time corresponding to the space heat map includes:

Determining that signal quality abnormality exists in a first area of a coverage heat map at a first moment under the railway scene, wherein the first area of the coverage heat map at the first moment is a first color, and the first area of a TA heat map at the first moment is a second color, wherein the first color of the coverage heat map is used for representing that signal intensity is larger than a first threshold, and the second color of the TA heat map is used for representing that time delay is larger than a second threshold;

and under the condition that the color in a second area of the coverage heat map at a second moment in the railway scene jumps in the railway direction, determining that the transient signal interference exists in the second area at the second moment.

In this embodiment, since the overlay heat map refers to a two-dimensional image generated from signal strength measurement data, it is used to reflect signal overlay strengths at different locations along the railway. Thus, in particular, the overlay heat map may be generated in a color-coded manner, e.g., with red areas representing areas with signal intensities greater than-75 dBm and green areas representing areas with signal intensities less than-95 dBm. The TA heat map refers to a two-dimensional image generated by time delay measurement data and is used for reflecting the distribution of signal transmission time delay under a railway scene, so that a color gradient can be used to represent a time delay value, for example, a red area in the TA heat map represents an area with a time delay exceeding 3 ms. Accordingly, the identification signal quality abnormality can be judged based on the combination of the first color and the second color.

Specifically, the data of the coverage heat map and the TA heat map are collected through the wireless network measuring equipment, so that the signal intensity and the time delay information in a railway scene can be recorded in real time when the railway moves along the railway, and the coverage heat map and the TA heat map are generated. When a certain first area in the coverage heat map at a certain moment is displayed as a first color, the signal intensity of the first area can be indicated to exceed a preset first threshold value, the first threshold value can be 75dBm, and meanwhile, the first area in the TA heat map is displayed as a second color, the time delay of the first area is indicated to exceed a preset second threshold value, and the second threshold value can be 3ms. At this time, the fault recognition model may determine that there is a signal quality abnormality in the area due to an abnormal combination of high signal strength and high delay.

In addition, the fault recognition model can detect color jump along the railway direction in the coverage heat map, if the coverage heat map has color jump in a second area at a second moment, for example, from red to green in a continuous 100-meter range, the fault recognition model indicates that the signal strength has a severe change on the path, and transient interference can be caused when a train passes through a tunnel or a high-voltage transmission line. By simultaneously analyzing the spatial distribution characteristics and the color change modes of the spatial heat map, the abnormality caused by normal signal fluctuation and the abnormality caused by faults can be accurately distinguished.

Through the technical scheme, the method and the device can accurately identify the transient network faults caused by abnormal distribution of the base stations or dynamic environment changes in the railway scene. The method comprises the steps of determining signal quality abnormality, performing coverage and TA heat map joint analysis, avoiding single index misjudgment, and detecting instantaneous signal interference, wherein the signal quality abnormality is determined by the coverage and TA heat map joint analysis, and the problem of insufficient sensitivity of the traditional method in a dynamic scene is solved by capturing color jump along the railway direction. The technical scheme is particularly suitable for a railway environment moving at a high speed, can effectively distinguish normal signal fluctuation from real faults, and improves accuracy and instantaneity of network fault diagnosis.

In some embodiments, a plurality of space heat maps in a first time period in a railway scene can be obtained, wherein the first time period comprises a plurality of first time periods, then at least two space heat maps in each first time period are spliced to obtain at least one first spliced image, and short-time interference faults in the first time period corresponding to the first spliced image are determined according to at least one of the first spliced images and the time sequence curve graph through the fault identification model.

In this embodiment, the first stitched image is at least two space heat maps acquired in the first period, and a plurality of space heat maps acquired in the railway scene are stitched in time sequence, so as to form a two-dimensional stitched image represented by time-space combination. The first stitched image may be a combined image formed by stitching multiple space heat maps of the same area in a continuous short period along a time sequence direction, for example, the first period is one day, and the first stitched image is a stitched image obtained by transversely stitching the space heat maps acquired every 1 hour in one day. The fault identification model can judge whether short-time interference faults exist in a first period of time under the railway scene through the first spliced image. Wherein, the short-time interference fault refers to the sudden interference problem which occurs in a short time period.

In some specific embodiments, the first stitched image includes a first overlay stitching graph, where the first overlay stitching graph is obtained by stitching at least two overlay heat maps in a first period according to a time sequence;

the determining, by the fault recognition model, a short-time interference fault in a first period corresponding to the first stitched image according to each first stitched image includes:

and under the condition that the color in a third area in the first coverage splice graph jumps in the time sequence direction in the railway scene, determining that a short-time interference fault exists in the third area in the first period.

In this embodiment, the first coverage mosaic refers to a two-dimensional image formed by performing transverse or longitudinal stitching on coverage heat maps acquired in a continuous first period in the same railway scene in a time sequence, so that the first stitched image can embody space features as well as time sequence features.

The color jump in the time sequence direction means that the signal intensity value of a specific third area in the spliced graph shows discontinuous change on a time axis, if the fault identification model is used for analyzing the first coverage spliced graph in the railway scene, if the color jump of the third area in the time sequence direction is found, that is, the area has obvious color change between different acquisition moments, the signal intensity value in the third area has abrupt change between adjacent acquisition moments, and the fault identification model can judge that the third area has short-time interference faults in a corresponding time period according to the signal intensity value. This color jump phenomenon indicates that the network signal gets worse from good or from bad in a short time, usually without persistence, often caused by external temporary interference sources (e.g. temporary electromagnetic equipment, construction interference) and is thus identified as a short-time interference fault.

In particular, in a railway scenario, mobile terminals may experience short-lived sources of interference during high speed operation. Conventional single frame thermogram analysis cannot capture such transients. And after the two-dimensional images containing time dimension information are formed by splicing the coverage heat maps acquired in the first time period according to time sequence, the model can extract the characteristic mode in the time sequence direction through the convolutional neural network. When color jump occurs in a certain area in the splice graph, the position is indicated to have abrupt change of signal intensity in a corresponding time period. For example, a stripe in which red and yellow alternate is detected in the third region of the mosaic, indicating that there is a short-time signal fluctuation caused by a temporary obstacle or electromagnetic interference in the region. The time sequence splicing mode enables the model to directly learn time correlation characteristics of interference events, and avoids complexity of manually extracting time sequence characteristics in the traditional method.

Through the technical scheme, the problem that the dynamic characteristics of the short-time interference fault in the time dimension are difficult to capture in the railway scene is solved, and the accurate identification of the temporary interference event is realized. According to the method, a dynamic process is converted into a two-dimensional space mode through time sequence splicing, so that complexity of processing time sequence data by a model is effectively reduced, instantaneity and accuracy of fault detection are improved, and a reliable technical means is provided for rapid fault positioning of a railway mobile communication system.

In some embodiments, a plurality of space heat maps in a first time period in a railway scene can be acquired, the first time period comprises at least one second time period, the duration of the second time period is longer than that of the first time period, at least two space heat maps in each second time period are spliced to obtain at least one second spliced image, and a base station fault in the second time period corresponding to the second spliced image is determined according to at least one of the second spliced images and the time sequence graph through the fault identification model.

In this embodiment, the second stitched image is at least two space heat maps acquired in the second period, and the plurality of space heat maps acquired in the railway scene are stitched in time sequence, so as to form a two-dimensional stitched image represented by time-space combination. The second stitched image may be a combined image formed by stitching multiple space heat maps of the same area in a continuous long period along a time sequence direction, for example, the second period is one month, and the second stitched image is a stitching image obtained by stitching one space heat map acquired every other day in one month in a transverse or longitudinal manner. The fault identification model can judge whether the base station fault exists in the second period of time under the railway scene through the second spliced image. The base station fault refers to the continuous attenuation trend of signals in a space heat map splice map acquired by multiple days in a longer period of time, and the stability problem caused by the power reduction of the base station or the equipment aging is judged by combining the change of a time sequence graph.

In the embodiment of the application, different types of network faults can be identified in stages by utilizing a fault identification model through acquiring a plurality of space heat maps and time sequence graphs in a first duration under a railway scene, wherein the method specifically comprises the steps of identifying an instantaneous network fault based on a single Zhang Kongjian heat map, and identifying a short-time interference fault and a base station fault by combining the plurality of space heat maps with the time sequence graphs. Compared with the traditional mode of manually analyzing the images, the fault type is automatically identified through the model, and the problem is judged in a grading manner according to the time dimension, so that the fault identification process is more systematic and automatic. Through the joint analysis of the space heat map and the time sequence graph, the model can rapidly and accurately judge the fault types under different time scales, thereby remarkably improving the efficiency of network fault positioning and confirmation.

In some specific embodiments, the second stitched image includes a second coverage stitching graph, where the second coverage stitching graph is obtained by stitching at least two coverage heat maps in a second period of time according to a time sequence, and the base station fault includes a base station aging problem;

the determining, by the fault recognition model, the base station fault in the second period corresponding to the second stitched image according to each second stitched image includes:

and under the condition that a fourth area formed by the first color in the second coverage splice graph is contracted along the time sequence direction under the railway scene to form inverted trapezoid distribution, determining that the base station ageing problem exists in the second period.

In this embodiment, the second coverage stitching graph refers to a two-dimensional image formed by stitching coverage heat graphs continuously acquired in the second period along the first direction according to a time sequence, for example, the spatial heat graphs acquired by month may be arranged in a direction from top to bottom, so that the lateral direction represents a spatial delay, and the longitudinal dimension represents a time lapse. The duration of the second period is set to be greater than the first period, for example, the second period may be weeks or months for capturing a long-term trend of change in the base station performance. The inverted trapezoidal distribution refers to a geometric form of the first color region with a wide upper part and a narrow lower part in a time sequence dimension, and the shrinkage characteristic reflects a physical phenomenon that the coverage area of the base station is gradually reduced along with time.

In particular, since the first color in the overlay heat map is used to characterize the high power region where the signal strength is above a certain threshold. In a railway scenario, the ideal coverage area of a base station should exhibit a continuous band-like distribution as the train moves along a fixed track. When the base station equipment ages and causes a decrease in transmit power, its coverage area will shrink day by day. By stitching the multiple overlay heatmaps along the timing during the second time period, the shrinkage process of the high power region is converted to an inverted trapezoidal geometry in the two-dimensional image. The characteristic has obvious difference with local abnormality generated by random jump or instantaneous fault caused by short-time interference, so that the fault identification model can accurately judge the aging problem of the base station through morphological analysis.

Compared with the prior art, the traditional method relies on Shan Zhangre diagrams or short-term statistical data, and gradual faults and instantaneous interference caused by equipment aging cannot be distinguished. According to the scheme, a two-dimensional space-time distribution diagram covering the heat map is formed through time sequence splicing, and time dimension information is converted into geometric features which can be identified in space, so that the problem of long-term performance degradation can be visually presented and accurately detected.

Through the technical scheme, the automatic identification of the aging problem of the base station in the railway scene is realized, and the fault missed detection caused by low manual inspection efficiency is effectively avoided. By capturing the inverted trapezoid distribution characteristics, equipment aging and short-time interference can be accurately distinguished, a reliable decision basis is provided for base station maintenance, and redundant maintenance cost caused by misjudgment is avoided.

As an optional embodiment, the stitching the plurality of space heat maps according to the acquisition time of the plurality of space heat maps to obtain a stitched image includes:

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

The acquisition time is a time stamp generated by each space heat map, and can be automatically recorded through a system log or a data acquisition module, so that the accuracy of time sequence information is ensured. If the first direction may be a horizontal direction or a vertical direction, then the stitching along the first direction means that the space heat maps at different moments are arranged along the horizontal direction or the vertical direction according to the sequence of the acquisition time, for example, the heat maps at adjacent moments are sequentially connected along the horizontal axis to form a continuous time sequence stitched image.

Specifically, the method can establish the time sequence association relation of a plurality of space heat maps by extracting the time stamps of the plurality of space heat maps in the same time period. For example, in a railway scene, a coverage heat map is acquired every 1 hour, and then 24 Zhang Retu acquired continuously are spliced from top to bottom in time sequence into a spliced image, wherein the spliced image comprises a plurality of linear images extending along the line direction and arranged in parallel in the time sequence direction.

Compared with the prior art, the conventional network fault identification method generally processes the single Zhang Kongjian heat map independently and cannot capture the time sequence relevance of fault characteristics in a dynamic scene. For example, prior art heat map analysis of static scenes only focuses on spatial distribution, while train motion in railway scenes results in rapid changes in signal quality over time. In the scheme, a plurality of space thermal maps can be spliced based on time sequence to form a spliced map containing time evolution information, and the spliced map is used as the input of a fault identification model, so that the fault identification model can analyze the space distribution and the time change rule at the same time, and network faults under a railway scene can be identified more accurately.

Through the technical scheme, the problem that short-time interference faults are difficult to capture due to dynamic changes in a railway scene is solved. The time sequence splicing method converts the space heat map at discrete time into a continuous time sequence image, so that the fault identification model can directly learn dynamic characteristics such as signal jump, diffusion and the like, and the precision of wireless coverage fault identification and intelligent optimization of the network is improved.

In this embodiment, in the original space heat map, each pixel point generally forms a three-dimensional feature vector with two-dimensional position information (X, Y) and color information (e.g. RGB values) for representing the geographic position of the point on the railway line and its corresponding network signal status (e.g. coverage intensity, interference level, etc.). However, the traditional three-dimensional features cannot reflect the acquisition time of the image, so that the evolution rule of the network state in the time dimension is difficult to reflect.

Therefore, on the basis of the original (X, Y, RGB) characteristics, each pixel point in each space heat map can be combined with the acquisition time of the space heat map to which the pixel point belongs, time information is encoded into a color channel of the pixel point, for example, the acquisition time can be normalized and then is subjected to function fusion or splicing with the RGB channel, and then, a characteristic vector (X, Y, RGB_T) with a three-dimensional structure but an implicit time dimension can be obtained. The feature vector not only describes the spatial position and signal state of the pixel point, but also expresses the point in time when the state occurs.

Finally, such three-dimensional feature vectors of all pixel points in the space heat map acquired from a plurality of different times can be fused to obtain a matrix, and the matrix is a spliced image capable of comprehensively expressing the changes of the space position, the signal intensity and the time.

By means of the scheme, spatial and temporal expression of network states in a railway scene can be achieved by fusing the spatial, color and time information into the unified spliced image, accurate identification of network faults evolving along with time by the model is facilitated, and fault diagnosis accuracy and efficiency are improved.

acquiring a trained computer vision model;

In this embodiment, the trained computer vision model refers to a deep neural network model with general image feature extraction capability, such as YOLO and CNN, and is mainly used for extracting spatial structural features of an image.

However, since the original design of the computer vision model is oriented to the static image, the method cannot be directly adapted to the spliced image fused with the time sequence information, and therefore, the structure of the spliced image is required to be finely adjusted. Specifically, a network structure, such as a time embedding module or a multi-channel fusion layer, is newly added at the input end of the fault recognition model and is used for encoding the acquisition time of each image, and an input vector with enhanced time sequence is formed together with the original image characteristics, so that the fault recognition model has the perception capability of time-sequence joint characteristics. And training the fine-tuned fault identification model until convergence.

In this way, by fine tuning the input end of the model, the fault identification model can have the capability of processing time sequence space images, and the accuracy of dynamic wireless coverage fault identification and intelligent optimization in the railway network is remarkably improved.

As an optional embodiment, after the inputting the stitched image into the fault identification model and obtaining the network fault information of the railway scene output by the fault identification model, the method further includes:

In this embodiment, the solution prediction model refers to a machine learning model constructed based on expert experience and historical optimization cases, and specifically, a text form network fault may be input into the solution prediction model, where the solution prediction model may analyze the input network fault and output a target solution of the network fault. In the training process of the scheme prediction model, a second training sample set can be obtained, wherein the second training sample set comprises a plurality of historical network fault data and solutions corresponding to the historical network fault data; and training the solution prediction model by adopting the second training sample set until the solution prediction model meets a preset second convergence condition, and stopping training the solution prediction model to obtain the solution prediction model trained to be converged.

In this embodiment, when the failure recognition model detects that there is a network failure in the current railway scenario, a text description of the network failure, i.e., network failure information, may be generated. If only one network failure is included in the network failure information, the network failure may be directly input into the solution prediction model, from which the solution of the network failure is input.

If the network fault information comprises at least two network faults, the at least two network faults can be displayed through a visual interface, and a user selects a first network fault to be processed from the displayed at least two network faults through a first input, wherein the first input can be touch operation or voice instruction.

After the user selects the first network fault through the first input, the fault information of the selected first network fault can be input into a scheme prediction model after the feature vector conversion, and the scheme prediction model can perform scheme matching based on the fault information of the first network fault to generate a target solution of the first network fault.

Compared with the prior art, the traditional scheme relies on manual fault screening and refers to an experience manual to formulate a solution, and has the problems of low processing efficiency and poor scheme suitability. According to the scheme, through automatic fault information integration and intelligent model prediction, rapid response during multi-fault concurrency under a railway dynamic scene is realized, and the problem that time sequence associated faults are easy to ignore in manual processing is avoided.

Through the technical scheme, the collaborative processing difficulty of the multi-source network fault in the railway scene is effectively solved, and the response speed and the scheme accuracy of network optimization in a complex dynamic environment are obviously improved. According to the scheme, the structural fault information is integrated to ensure the complete reservation of the multidimensional fault characteristics, the interactive selection mechanism is utilized to optimize the fault processing priority, and the practicality and scene adaptation degree of the solution are ensured by combining the prediction model trained by the special railway characteristics.

As an alternative embodiment, after the determining, by the solution prediction model, the target solution of the first network failure, the method further includes:

In this embodiment, after determining the target solution of the first network failure by the solution prediction model, a simulation scenario may be constructed according to the target solution and the railway scenario, propagation of the wireless signal under the simulation scenario may be simulated to obtain a simulated space heat map and a simulated time sequence graph, and in the case where the failure recognition model confirms that no network failure exists in the simulation scenario according to the simulated data, the target solution output by the solution prediction model is considered to be valid, and thus the target solution may be output to the user.

The simulation scene is a virtual test environment constructed based on the environmental characteristics of the target solution and the actual railway scene, and can be realized by combining a three-dimensional geographic information system with base station layout data, and a dynamic test scene is constructed by importing rail trend, train running speed and surrounding building distribution data.

Specifically, after the target solution is generated, the system automatically extracts the base station position parameters, the train running track and the environmental obstacle data in the railway scene, and constructs the linear coverage simulation scene containing time dimension changes. And simulating the propagation process of the signal along the rail under the optimized parameters by using a ray tracing algorithm, and generating a simulated space heat map and a simulated time sequence graph which are consistent with the measured data format. The fault recognition model carries out secondary analysis on the simulation space heat map and the simulation time sequence graph, and if fault characteristics such as over coverage, interference and the like are not detected, the target solution is judged to be effective.

Compared with the prior art, the traditional optimization scheme only depends on empirical parameter adjustment and lacks verification links, and is easy to cause optimization failure due to dynamic environment change. According to the scheme, the special simulation environment of the railway is constructed, the signal propagation characteristics are restored by adopting a high-precision channel modeling technology, and a closed-loop verification mechanism is formed by combining a fault identification model, so that subjectivity of manual debugging and blindness of a traditional algorithm are effectively avoided.

Through the technical scheme, the method and the device can ensure the validity verification of the generated optimization parameters before actual deployment in a complex railway scene, solve the problem of insufficient reliability of optimization suggestions in the prior art, prevent network performance deterioration caused by improper parameter setting, and improve the safety and decision reliability of an automatic optimization system.

Based on the wireless coverage fault identification and intelligent optimization method provided by the embodiment, correspondingly, the application further provides a specific implementation mode of the wireless coverage fault identification and intelligent optimization device. Please refer to the following examples.

Referring first to fig. 2, a wireless coverage fault recognition and intelligent optimization apparatus 200 provided in an embodiment of the present application includes the following modules:

a first obtaining module 201, configured to obtain a plurality of space heat maps and a plurality of time sequence graphs in a railway scene;

The stitching module 202 is configured to stitch the plurality of space heat maps according to the acquisition time of the plurality of space heat maps to obtain a stitched image;

The first identifying module 203 is configured to input the stitched image and the time sequence graph into a fault identifying model, and identify network fault information of the railway scene according to at least one of the stitched image and the time sequence graph through the fault identifying model.

The equipment can be used for carrying out time sequence splicing on a plurality of railway scene space heat maps acquired according to time to obtain spliced images, and inputting the spliced images and the time sequence curve graph into the fault identification model together so that the fault identification model can identify fault information in a network according to at least one of the spliced images and the time sequence curve graph. Compared with the traditional mode of manually analyzing by graph, the scheme automatically identifies the fault type through the model, and the model can rapidly and accurately judge the fault type under different time and space through joint analysis of the spliced heat map and the time sequence graph, so that the network wireless coverage fault identification and intelligent optimization efficiency is remarkably improved.

As an implementation manner of the present application, the splicing module 202 may specifically be used for:

As an implementation manner of the present application, the above-mentioned splicing module 202 may be specifically further used for:

As an implementation manner of the present application, the wireless coverage fault identification and intelligent optimization apparatus 200 may further include:

the second acquisition module is used for acquiring the trained computer vision model;

The fine tuning module is used for fine tuning the computer vision model by adding a network structure at the input end of the computer vision model to obtain a fault identification model;

and the training module is used for training the fault recognition model to obtain a fault recognition model trained to be converged.

The display module is used for displaying at least two network faults in the railway scene under the condition that the network fault information comprises the at least two network faults, wherein the network faults comprise instantaneous network faults, short-time interference faults and base station faults;

A receiving module for receiving a first input to a first network failure of the at least two network failures;

and the prediction module is used for responding to the first input, inputting the first network fault corresponding to the first input into a scheme prediction model, and determining a target solution of the first network fault through the scheme prediction model.

The construction module is used for constructing a simulation scene according to the target solution and the railway scene;

The simulation module is used for simulating the propagation of the wireless signal in the simulation scene to obtain a simulation space heat map and a simulation time sequence graph in the simulation scene;

and the output module is used for outputting the target solution to a user under the condition that the fault recognition model determines that no network fault exists in the simulation scene according to the simulation space heat map and the simulation time sequence graph.

The wireless coverage fault recognition and intelligent optimization device provided by the embodiment of the invention can realize each step in the method embodiment, and in order to avoid repetition, the description is omitted here.

Fig. 3 is a schematic hardware structure diagram of a wireless coverage fault recognition and intelligent optimization device according to an embodiment of the present application.

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

In particular, the processor 1001 may include a Central Processing Unit (CPU), or an Application SPECIFIC INTEGRATED Circuit (ASIC), or may be configured as one or more integrated circuits that implement embodiments of the present application.

Memory 1002 may include mass storage for data or instructions. By way of example, and not limitation, memory 1002 may include a hard disk drive (HARD DISK DRIVE, HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or universal serial bus (Universal Serial Bus, USB) drive, or a combination of two or more of the foregoing. The 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, where appropriate. In a particular embodiment, the memory 1002 is a non-volatile solid state memory.

The memory may include Read Only Memory (ROM), random Access Memory (RAM), magnetic disk storage media devices, optical storage media devices, flash memory devices, electrical, optical, or other physical/tangible memory storage devices. Thus, in general, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software comprising 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 methods in accordance with aspects of the present disclosure.

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

In one example, the wireless coverage fault identification and intelligent optimization device can also include a communication interface 1003 and a bus 1010. As shown in fig. 3, the processor 1001, the memory 1002, and the communication interface 1003 are connected to each other by a bus 1010, and communicate with each other.

The communication interface 1003 is mainly used for implementing communication among the modules, devices, units and/or apparatuses in the embodiment of the application.

Bus 1010 includes hardware, software, or both, coupling components of the wireless coverage fault identification and intelligent optimization device to each other. By way of example, and not limitation, the buses may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an infiniband interconnect, a Low Pin Count (LPC) bus, a memory bus, a micro channel 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 bus, or a combination of two or more of the above. Bus 1010 may include one or more buses, where appropriate. Although embodiments of the application have been described and illustrated with respect to a particular bus, the application contemplates any suitable bus or interconnect.

The wireless coverage fault identification and intelligent optimization device can be based on the embodiment, so that the wireless coverage fault identification and intelligent optimization method and device are combined.

In addition, in combination with the wireless coverage fault recognition and intelligent optimization method in the above embodiment, the embodiment of the application can be implemented by providing a computer storage medium. The computer storage medium stores computer program instructions, which when executed by the processor, implement any one of the wireless coverage fault identification and intelligent optimization methods in the above embodiments, and achieve the same technical effects, and are not repeated here. The computer readable storage medium may include a non-transitory computer readable storage medium, such as a read only memory (ReadOnly Memory, ROM for short), a random access memory (Random Access Memory, RAM for short), a magnetic disk or an optical disk, and the like, which are not limited herein.

In addition, the embodiment of the application also provides a computer program product, which comprises computer program instructions, wherein the computer program instructions can realize the steps and corresponding contents of the embodiment of the method when being executed by a processor.

It should be understood that the application is not limited to the particular arrangements and instrumentality described above and shown in the drawings. For the sake of brevity, a detailed description of known methods is omitted here. In the above embodiments, several specific steps are described and shown as examples. The method processes of the present application are not limited to the specific steps described and shown, but various changes, modifications and additions, or the order between steps may be made by those skilled in the art after appreciating the spirit of the present application.

The functional blocks shown in the above block diagrams may be implemented in hardware, software, firmware, or a combination thereof. When implemented in hardware, it may be, for example, an electronic circuit, an Application Specific Integrated Circuit (ASIC), suitable firmware, a plug-in, a function card, or the like. When implemented in software, the elements of the application are the programs or code segments used to perform the required tasks. The program or code segments may be stored in a machine readable medium or transmitted over transmission media or communication links by a data signal carried in a carrier wave. A "machine-readable medium" may include any medium that can store or transfer information. Examples of machine-readable media include electronic circuitry, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio Frequency (RF) links, and the like. The code segments may be downloaded via computer networks such as the internet, intranets, etc.

It should also be noted that the exemplary embodiments mentioned in this disclosure describe some methods or systems based on a series of steps or devices. The present application is not limited to the order of the above-described steps, that is, the steps may be performed in the order mentioned in the embodiments, or may be performed in a different order from the order in the embodiments, or several steps may be performed simultaneously.

Aspects of the present disclosure are described above with reference to flowchart illustrations and/or block diagrams of methods, apparatus and computer program products according to embodiments of the disclosure. It will be understood that each block of 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 may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions/acts specified in the flowchart and/or block diagram block or blocks. Such a processor may be, but is not limited to being, a general purpose processor, a special purpose processor, an application specific processor, or a field programmable logic circuit. It will also be understood that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware which performs the specified functions or acts, or combinations of special purpose hardware and computer instructions.

In the foregoing, only the specific embodiments of the present application are described, and it will be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working processes of the systems, modules and units described above may refer to the corresponding processes in the foregoing method embodiments, which are not repeated herein. It should be understood that the scope of the present application is not limited thereto, and any equivalent modifications or substitutions can be easily made by those skilled in the art within the technical scope of the present application, and they should be included in the scope of the present application.

Claims

1. A method for identifying and intelligently optimizing wireless coverage failures, comprising:

Obtain multiple spatial heat maps and multiple time series curve graphs in railway scenarios;

splicing the multiple spatial heat maps according to the acquisition time of the multiple spatial heat maps to obtain a spliced image;

The spliced image and the timing curve diagram are input into a fault identification model, and the fault identification model is used to identify network fault information of the railway scene according to at least one of the spliced image and the timing curve diagram.

2. The wireless coverage fault identification and intelligent optimization method according to claim 1, wherein the step of stitching the multiple spatial heat maps according to their acquisition time to obtain a stitched image comprises:

According to the time sequence of the acquisition time of the multiple spatial heat maps, the multiple spatial heat maps are spliced along a first direction perpendicular to the railway line to obtain a spliced image, wherein the spliced image includes multiple linear images extending along the line direction and arranged in parallel in the time sequence direction.

3. The wireless coverage fault identification and intelligent optimization method according to claim 1, wherein the step of stitching the multiple spatial heat maps according to their acquisition time to obtain a stitched image comprises:

For any pixel in any spatial heat map, obtain the position information and color information of the pixel;

Obtaining the acquisition time of the spatial heat map where the pixel point is located;

Determining a feature vector of the pixel point according to the acquisition time of the spatial heat map, the position information and the color information of the pixel point;

The feature vectors of all pixels in the multiple spatial heat maps are fused to obtain the spliced image.

4. The wireless coverage fault identification and intelligent optimization method according to claim 1, characterized in that before inputting the spliced image and the timing curve graph into a fault identification model and identifying network fault information of the railway scenario based on at least one of the spliced image and the timing curve graph using the fault identification model, the method further comprises:

Get the trained computer vision model;

By adding a network structure to the input end of the computer vision model, the computer vision model is fine-tuned to obtain a fault recognition model;

The fault identification model is trained to obtain a fault identification model that is trained to convergence.

5. The wireless coverage fault identification and intelligent optimization method according to claim 1, characterized in that after inputting the spliced image into a fault identification model to obtain network fault information of the railway scene output by the fault identification model, the method further comprises:

In a case where the network fault information includes at least two network faults in the railway scenario, displaying the at least two network faults, wherein the network faults include instantaneous network faults, short-term interference faults, and base station faults;

receiving a first input of a first network failure of the at least two network failures;

In response to the first input, a first network fault corresponding to the first input is input into a solution prediction model, and a target solution to the first network fault is determined by the solution prediction model.

6. The wireless coverage fault identification and intelligent optimization method according to claim 5, wherein after determining the target solution to the first network fault using the solution prediction model, the method further comprises:

constructing a simulation scenario based on the target solution and the railway scenario;

Simulating the propagation of wireless signals in the simulation scenario to obtain a simulated spatial heat map and a simulated timing curve graph in the simulation scenario;

When the fault identification model determines that no network fault exists in the simulation scenario according to the simulation space heat map and the simulation timing curve graph, the target solution is output to the user.

7. A wireless coverage fault identification and intelligent optimization device, characterized in that the device comprises:

The first acquisition module is used to obtain multiple spatial heat maps and multiple time series curve graphs in the railway scene;

a stitching module, configured to stitch the multiple spatial heat maps together according to the acquisition time of the multiple spatial heat maps to obtain a stitched image;

The first recognition module is configured to input the spliced image and the timing curve diagram into a fault recognition model, and recognize network fault information of the railway scene according to at least one of the spliced image and the timing curve diagram through the fault recognition model.

8. A wireless coverage fault identification and intelligent optimization device, characterized in that the wireless coverage fault identification and intelligent optimization device comprises: a processor and a memory storing computer program instructions;

When the processor executes the computer program instructions, the wireless coverage fault identification and intelligent optimization method according to any one of claims 1 to 6 is implemented.

9. A computer storage medium, characterized in that computer program instructions are stored on the computer storage medium, and when the computer program instructions are executed by a processor, the wireless coverage fault identification and intelligent optimization method according to any one of claims 1 to 6 is implemented.

10. A computer program product, characterized in that the computer program product comprises computer program instructions, and when the computer program instructions are executed by a processor, the wireless coverage fault identification and intelligent optimization method according to any one of claims 1 to 6 is implemented.