Train communication quality detection method and train

Abstract

The present disclosure provides a train communication quality detection method and a train, which can be applied to the technical field of network monitoring. The method comprises the following steps: projecting a plurality of running parameter groups of a target network device in a train at a plurality of time points in a first preset time period into a multi-dimensional vector space based on the running parameter groups, to obtain a plurality of running vectors of the plurality of time points; clustering the plurality of running vectors to obtain a clustering result; in response to obtaining a plurality of updated running parameter groups of the target network device at a plurality of time points in a second preset time period, updating at least one cluster based on a plurality of updated running vectors determined by the updated running parameter groups to obtain a clustering update result; and determining a communication quality detection result of each port in the target network device based on the clustering update result and the clustering result.

Description

Technical Field

[0001] This disclosure relates to the field of network monitoring technology, specifically to a train communication quality testing method and a train. Background Technology

[0002] During train operation, the multi-port communication quality of onboard network equipment directly affects the stable operation of critical services such as the train control system and passenger service system. Due to the complex train network environment, network equipment needs to operate continuously under harsh conditions such as vibration and electromagnetic interference. Ports are prone to anomalies such as data packet loss and increased latency. Therefore, the communication quality of each port is usually detected by comparing the operating parameters of each port with reference values, so as to promptly detect and locate problems and ensure the normal operation of the train.

[0003] In the process of implementing this disclosure, it was found that the relevant technology has at least the following problems: the ports of network devices are affected by the complex operating conditions they are in, and occasional fluctuations in communication quality are likely to occur. In this case, judging by reference values ​​will identify the port as an abnormal port and trigger the maintenance process, resulting in unnecessary waste of resources. Summary of the Invention

[0004] In view of the above problems, this disclosure provides a method for detecting train communication quality and a train.

[0005] According to a first aspect of this disclosure, a train communication quality detection method is provided, comprising: projecting multiple sets of operating parameters for target network devices in a train into a multi-dimensional vector space based on acquired sets of operating parameters for each time point within a first preset time period, thereby obtaining operating vectors for each time point, wherein the sets of operating parameters include operating parameters for multiple ports of the target network devices, and the dimension of the multi-dimensional vector space is determined according to the number of ports of the target network devices; clustering the multiple operating vectors to obtain clustering results, wherein the clustering results include at least one cluster and a cluster type, the cluster type indicating whether the ports of the target network devices belonging to the clusters are in a normal or abnormal state; in response to acquiring updated sets of operating parameters for each time point within a second preset time period, updating at least one cluster based on the multiple updated operating vectors determined by the updated sets of operating parameters, thereby obtaining a clustering update result; and determining the communication quality detection result for each port of the target network devices based on the clustering update result and the clustering result.

[0006] According to embodiments of this disclosure, each dimension in the multidimensional vector space corresponds to a port of the target network device; the running vector at each time point corresponds to a set of running parameters; multiple sets of running parameters are projected into the multidimensional vector space to obtain running vectors at multiple time points, including: for each set of running parameters, based on the mapping relationship between each port in the set of running parameters and the dimensions of the running vector, and the running parameters of each port, determining the label values ​​of multiple dimensions in the running vector.

[0007] According to embodiments of this disclosure, determining the communication quality detection result for each port in the target network device based on the clustering update result and the clustering result includes: when the number of update running vectors representing the updated clusters in the clustering update result is greater than a threshold, determining the communication quality detection result for each port in the target network device based on the cluster type of the updated clusters, wherein the update running vectors of the updated clusters include the running vectors added to the updated clusters relative to the unupdated clusters; wherein the cluster type of the updated clusters is the same as the cluster type of the unupdated clusters; the cluster type of the unupdated clusters is determined based on the running parameter group corresponding to the running vectors and the reference running parameter group.

[0008] According to embodiments of this disclosure, determining the communication quality detection results of each port in a target network device based on the cluster type of the updated cluster includes: generating a communication quality detection result indicating that all ports in the target network device are normal when the cluster type of the updated cluster indicates that the ports belonging to the target network device in the cluster are in a normal state; determining abnormal ports based on the abnormal dimension of the cluster and the correspondence between dimensions and ports when the cluster type of the updated cluster indicates that the ports belonging to the target network device in the cluster are in an abnormal state; and generating a communication quality detection result indicating that the abnormal ports of the target network device are in an abnormal state; wherein, the abnormal dimension of the updated cluster is determined by the following method: for multiple dimensions of the multidimensional vector space, the difference between the label values ​​of the updated cluster and the normal cluster in the dimension is determined respectively to obtain the operating parameter offsets of the updated cluster and the normal cluster in multiple dimensions, wherein the cluster type of the normal cluster indicates that the ports belonging to the target network device in the normal cluster are in a normal state; and determining the dimension where the operating parameter offset reaches the offset threshold as the abnormal dimension of the updated cluster.

[0009] According to embodiments of this disclosure, clustering multiple running vectors to obtain clustering results includes: repeating the following operations until all running vectors have been labeled: obtaining an unlabeled running vector from the multiple running vectors; determining the distance between the cluster center of the labeled cluster and the running vector in the multidimensional vector space as the labeling distance; determining the labeling result of the running vector based on the labeling distance and a preset distance; and determining at least one cluster based on the labeling results of the multiple running vectors, and determining the clustering result.

[0010] According to embodiments of this disclosure, determining the labeling result of a running vector based on a labeling distance and a preset distance includes: if there is a labeling distance less than the preset distance, labeling the running vector to the cluster to which the cluster center corresponding to the labeling distance less than the preset distance belongs; if there is no labeling distance less than the preset distance, determining the neighborhood range of the running vector with the running vector as the center and the preset distance as the radius; if the number of running vectors that have been labeled in the neighborhood range reaches a preset number, determining the cluster with the running vector as the cluster center; and if there is no labeling distance less than the preset distance and the number of running vectors that have been labeled in the neighborhood range does not reach the preset number, labeling the running vector as noise.

[0011] According to an embodiment of this disclosure, the train communication quality detection method further includes: when it is determined that the distance between the running vector and multiple labeled clusters is less than a preset distance, merging the multiple labeled clusters to obtain an updated cluster, and labeling the running vector into the updated cluster.

[0012] According to embodiments of this disclosure, at least one cluster is updated based on multiple update running vectors determined by an update running parameter group to obtain a clustering update result, including: determining the cluster center of each of the multiple clusters; for each update running vector, determining the distance between the update running vector and at least one cluster center to obtain at least one update distance; and updating the update running vector to the nearest cluster based on the at least one update distance to obtain a clustering update result.

[0013] According to embodiments of this disclosure, updating the update running vector to the nearest cluster based on at least one update distance to obtain a cluster update result includes: adding the update running vector to the cluster, and recalculating the average of multiple running vectors based on the update running vector and multiple running vectors in the cluster to update the cluster center of the cluster, thereby obtaining a cluster update result.

[0014] A second aspect of this disclosure provides a train, comprising: a target network device disposed within the train, the target network device including multiple ports; a memory for storing operating parameters of each port of the target network device; and a controller for retrieving the operating parameters of each port of the target network device from the memory and executing the steps of the above method based on the operating parameters of each port of the target network device to obtain a communication quality prediction result for each port in the target network device.

[0015] A third aspect of this disclosure provides a train communication quality detection device, comprising: a vector projection module, configured to project multiple sets of operating parameters for target network devices in a train into a multi-dimensional vector space based on the acquired sets of operating parameters for each of multiple times within a first preset time period, thereby obtaining operating vectors for each of the multiple times, wherein the sets of operating parameters include the operating parameters of multiple ports of the target network devices, and the dimension of the multi-dimensional vector space is determined according to the number of ports of the target network devices; a vector clustering module, configured to cluster the multiple operating vectors to obtain clustering results, wherein the clustering results include at least one cluster and a cluster type, the cluster type indicating whether the ports of the target network devices belonging to the clusters are in a normal or abnormal state; a clustering update module, configured to update at least one cluster based on the multiple updated operating vectors determined by the updated sets of operating parameters for each of the multiple times within a second preset time period, thereby obtaining a clustering update result; and a result determination module, configured to determine the communication quality detection result for each port of the target network devices based on the clustering update result and the clustering result.

[0016] A fourth aspect of this disclosure provides an electronic device comprising: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors perform the methods described above.

[0017] A fifth aspect of this disclosure also provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, cause the processor to perform the methods described above.

[0018] A sixth aspect of this disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0019] According to embodiments of this disclosure, the operating parameter groups of multiple ports of the target network device are projected into multi-dimensional operating vectors and clustered. The clustering results are then updated based on subsequent data. This enables dynamic detection of communication quality and independent evaluation of multiple ports of the target network device. The clustering operation determines clusters only after a certain number of operating vectors of the same type have been collected. Independently occurring operating vectors with offsets typically do not form clusters and are considered noise. This addresses occasional quality degradation caused by normal fluctuations in some interfaces of the target network device, preventing it from affecting the communication quality detection results and ensuring the accuracy of communication quality detection. The clustering update mechanism based on data from a subsequent second preset time period does not require re-clustering the entire dataset; it only optimizes the clustering results based on new data. While ensuring comprehensive detection, it improves real-time detection efficiency and adapts to the highly dynamic, multi-port operating scenarios of train network devices. Attached Figure Description

[0020] The foregoing contents, as well as other objects, features, and advantages of this disclosure, will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0021] Figure 1 The diagram illustrates a train communication quality detection method and a train application scenario according to embodiments of the present disclosure.

[0022] Figure 2 A flowchart illustrating a train communication quality detection method according to an embodiment of the present disclosure is shown schematically.

[0023] Figure 3A This illustration schematically shows intermediate results during the clustering process of the train communication quality detection method according to an embodiment of the present disclosure;

[0024] Figure 3B The illustration schematically shows intermediate results during the clustering process of a train communication quality detection method according to another embodiment of the present disclosure;

[0025] Figure 3C The illustration schematically shows intermediate results during the clustering process of a train communication quality detection method according to yet another embodiment of the present disclosure;

[0026] Figure 4 This diagram illustrates a method for determining abnormal ports according to an embodiment of the present disclosure;

[0027] Figure 5 A schematic diagram illustrating the structure of a train communication quality detection device according to an embodiment of the present disclosure is shown; and

[0028] Figure 6A block diagram schematically illustrates an electronic device suitable for implementing a train communication quality detection method according to an embodiment of the present disclosure. Detailed Implementation

[0029] The embodiments of the present disclosure will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the disclosure. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of the present disclosure for ease of explanation. However, it will be apparent that one or more embodiments may be practiced without these specific details. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concepts of the present disclosure.

[0030] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. The terms “comprising,” “including,” etc., as used herein indicate the presence of the stated features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0031] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0032] When using expressions such as "at least one of A, B and C", they should generally be interpreted in accordance with the meaning that is commonly understood by those skilled in the art (e.g., "a system having at least one of A, B and C" should include, but is not limited to, a system having A alone, a system having B alone, a system having C alone, a system having A and B, a system having A and C, a system having B and C, and / or a system having A, B and C, etc.).

[0033] This disclosure provides a train communication quality detection method, comprising: projecting multiple sets of operating parameters of a target network device in a train into a multi-dimensional vector space based on the acquired sets of operating parameters for multiple moments within a first preset time period, thereby obtaining operating vectors for each moment, wherein the operating parameter sets include operating parameters for multiple ports of the target network device, and the dimension of the multi-dimensional vector space is determined according to the number of ports of the target network device; clustering the multiple operating vectors to obtain clustering results, wherein the clustering results include at least one cluster and a cluster type, the cluster type indicating whether the port of the target network device belonging to the cluster is in a normal or abnormal state; in response to acquiring updated sets of operating parameters for multiple moments within a second preset time period, updating at least one cluster according to the multiple updated operating vectors determined by the updated operating parameter sets, thereby obtaining a clustering update result; and determining the communication quality detection result for each port of the target network device based on the clustering update result and the clustering result.

[0034] Figure 1 The diagram illustrates a train communication quality detection method and a train application scenario according to an embodiment of the present disclosure.

[0035] like Figure 1 As shown, the application scenario 100 according to this embodiment may include a train 110, which includes a target network device 111, a memory 112, and a controller 113.

[0036] The target network device 111 is installed inside the train 110, and the target network device 111 includes multiple ports.

[0037] The memory 112 is used to store the operating parameters of each port of the target network device 111.

[0038] The controller 113 is used to obtain the operating parameters of each port of the target network device 111 from the memory 112, and perform calculations and verifications based on the operating parameters of each port of the target network device 111 to obtain the communication quality prediction results of each port in the target network device 111.

[0039] It should be noted that the train communication quality detection method provided in this embodiment can generally be executed by the controller 113. Correspondingly, the train communication quality detection device provided in this embodiment can generally be located within the controller 113. The train communication quality detection method provided in this embodiment can also be executed by a server or server cluster that is different from the controller 113 and capable of communicating with the target network device 111, the memory 112, and / or the controller 113. Correspondingly, the train communication quality detection device provided in this embodiment can also be located in a server or server cluster that is different from the controller 113 and capable of communicating with the target network device 111, the memory 112, and / or the controller 113.

[0040] It should be understood that Figure 1 The number of trains, target network devices, memory, and controllers shown is merely illustrative. Any number of trains, target network devices, memory, and controllers can be used depending on implementation needs.

[0041] Figure 2 A flowchart illustrating a train communication quality detection method according to an embodiment of the present disclosure is shown schematically.

[0042] like Figure 2 As shown, the train communication quality detection method of this embodiment includes operations S210 to S240.

[0043] In operation S210, based on the acquired operating parameter groups of the target network devices in the train at multiple times within the first preset time period, the multiple operating parameter groups are projected into a multi-dimensional vector space to obtain the operating vectors at each time.

[0044] In operation S220, multiple running vectors are clustered to obtain clustering results. The clustering results include at least one cluster and the cluster type of the cluster. The cluster type indicates whether the port of the target network device belonging to the cluster is in a normal or abnormal state.

[0045] In operation S230, in response to obtaining the update operation parameter groups for multiple times within the second preset time period, at least one cluster is updated according to the multiple update operation vectors determined by the update operation parameter groups to obtain the clustering update result.

[0046] In operation S240, based on the clustering update results and clustering results, the communication quality detection results for each port in the target network device are determined.

[0047] The target network devices may include devices installed on the train that communicate and interact via the network through ports, such as Ethernet switches.

[0048] Taking an Ethernet switch as an example, when the communication quality of all ports of an Ethernet switch is normal, the parameter values ​​related to each port will remain stable or fluctuate within a certain range. However, when an abnormal communication quality occurs at a port, the parameter values ​​related to the port experiencing the abnormality will deviate from the normal range. Therefore, the communication quality of a target network device can be detected by monitoring the parameter values ​​of each port during operation.

[0049] Specifically, for the first preset period after the target network device starts working, the operating parameters of each port of the target network device at each moment in the first preset period can be collected as the operating parameter group at that moment for subsequent detection.

[0050] Taking an Ethernet switch as an example, an Ethernet switch has five ports: Ethernet port, aggregation port, access port, console port, and management port. At each moment, the operating parameters of the above five ports are obtained, and the obtained five operating parameters are used as the operating parameter group for that moment.

[0051] Since the failure status of multiple ports on a target network device is independent of each other, communication quality testing can be refined from the target network device to the port level. The operating parameter set includes the individual operating parameters of each port on the target network device. This set can be represented as operating vectors in a multi-dimensional vector space. This multi-dimensional vector space allows for independent analysis of the operating parameters and communication quality of each port. The dimensions of the multi-dimensional vector space are determined based on the number of ports on the target network device.

[0052] In the clustering results, if the cluster type indicates that all ports of the target network devices belonging to that cluster are in a normal state, the cluster can be identified as a normal cluster. If the cluster type indicates that the ports of the target network devices belonging to that cluster are in an abnormal state, the cluster can be identified as an abnormal cluster.

[0053] The system can continuously collect operating parameters from multiple ports, obtaining updated operating parameter sets at multiple time points. Based on these updated operating parameter sets within any second preset time period after obtaining the clustering results, at least one cluster in the clustering results can be updated to obtain the clustering update result. The clustering update result includes at least one updated cluster and the cluster type of the updated cluster.

[0054] In one example, the clustering method used when clustering multiple running vectors obtained by projecting the running parameter group of the first preset time period and the clustering method used when updating the clusters can be different. Preferably, a density-based clustering method can be used to cluster multiple running vectors, and a distance-based clustering method can be used to update the clusters.

[0055] Based on the clustering update results and clustering results, the situation of newly added update running vectors in each cluster can be determined, and the trend of update running parameter group in the second preset time period can be determined accordingly. Furthermore, it can be determined whether each port in the target network device represented by the above trend has a fault, and the communication quality detection result can be obtained.

[0056] According to embodiments of this disclosure, the operating parameter groups of multiple ports of the target network device are projected into multi-dimensional operating vectors and clustered. The clustering results are then updated based on subsequent data. This enables dynamic detection of communication quality and independent evaluation of multiple ports of the target network device. The clustering operation determines clusters only after a certain number of operating vectors of the same type have been collected. Independently occurring operating vectors with offsets typically do not form clusters and are considered noise. This addresses occasional quality degradation caused by normal fluctuations in some interfaces of the target network device, preventing it from affecting the communication quality detection results and ensuring the accuracy of communication quality detection. The clustering update mechanism based on data from a subsequent second preset time period does not require re-clustering the entire dataset; it only optimizes the clustering results based on new data. While ensuring comprehensive detection, it improves real-time detection efficiency and adapts to the highly dynamic, multi-port operating scenarios of train network devices.

[0057] According to embodiments of this disclosure, each dimension in the multidimensional vector space corresponds to a port of the target network device; the running vector at each time point corresponds to a set of running parameters; multiple sets of running parameters are projected into the multidimensional vector space to obtain running vectors at multiple time points, including: for each set of running parameters, based on the mapping relationship between each port in the set of running parameters and the dimensions of the running vector, and the running parameters of each port, determining the label values ​​of multiple dimensions in the running vector.

[0058] Each dimension of the multidimensional vector space can be used to represent the value of a port operating parameter of the target network device. Taking the Ethernet switch mentioned above as an example, the number of dimensions of the multidimensional vector space is the same as the number of ports, which is 5, and each dimension corresponds to one port. Therefore, the five dimensions can represent the operating parameters of the Ethernet port, aggregation port, access port, console port and management port in turn.

[0059] Therefore, for each time-based operating parameter group, the operating parameters of the Ethernet port, aggregation port, access port, console port, and management port in the operating parameter group can be marked as tag values ​​on the dimension corresponding to each port, and used as the components of the operating vector in that dimension.

[0060] According to embodiments of this disclosure, a one-to-one mapping between ports and vector dimensions is established, and vector label values ​​are directly determined through port operating parameters, enabling rapid structured transformation of multi-port operating data. This explicit mapping ensures that the operating vector accurately reflects the true operating status of each port, avoiding detection bias caused by parameter confusion, and providing a high-fidelity data foundation for subsequent clustering analysis, thus improving the accuracy of the detection results.

[0061] According to embodiments of this disclosure, clustering multiple running vectors to obtain clustering results includes: repeating the following operations until all running vectors have been labeled: obtaining an unlabeled running vector from the multiple running vectors; determining the distance between the cluster center of the labeled cluster and the running vector in the multidimensional vector space as the labeling distance; determining the labeling result of the running vector based on the labeling distance and a preset distance; and determining at least one cluster based on the labeling results of the multiple running vectors, and determining the clustering result.

[0062] The labeling result for each running vector can include one of the following: assigning the running vector to a cluster, determining a cluster with the running vector as the cluster center, or classifying the running vector as noise.

[0063] The labeling result can be determined based on the relationship between the labeling distance and the preset distance. Specifically, it can include: if there is a labeling distance smaller than the preset distance, labeling the running vector to the cluster to which the cluster center corresponding to the labeling distance smaller than the preset distance belongs; if there is no labeling distance smaller than the preset distance, determining the neighborhood range of the running vector with the running vector as the center and the preset distance as the radius; if the number of running vectors that have been labeled in the neighborhood range reaches a preset number, determining the cluster with the running vector as the cluster center; and if there is no labeling distance smaller than the preset distance and the number of running vectors that have been labeled in the neighborhood range does not reach the preset number, labeling the running vector as noise.

[0064] Figure 3A The illustration schematically shows intermediate results during the clustering process of the train communication quality detection method according to an embodiment of the present disclosure.

[0065] like Figure 3AAs shown, taking a two-dimensional vector space as an example, two clusters are obtained during the clustering process, namely the first cluster 301 and the second cluster 302. The first cluster 301 and the second cluster 302 each contain multiple running vectors. In addition, there are five running vectors that are marked as noise.

[0066] Taking running vector 303 as an example, in the two-dimensional vector space, it is determined that the marking distance between running vector 303 and the respective cluster centers of the first cluster 301 and the second cluster 302 is greater than the preset distance. Taking running vector 303 as the center and the preset distance as the radius, the neighborhood range of running vector 303 is determined, which is shown as a dashed circle in the figure. It can be determined that the number of running vectors that have been marked in the neighborhood range of running vector 303 is 2, which has not reached the preset number of 5. In this case, it is difficult to determine whether the deviation of running vector 303 is due to occasional fluctuations. Therefore, running vector 303 can be marked as noise.

[0067] Specifically, if a running vector is actually normal data, but because it is the first running vector of its type to be labeled and has not yet formed a cluster of that type, it will be labeled as noise. In subsequent labeling processes, running vectors will continuously be labeled in the multidimensional vector space, and those in its vicinity will be labeled as noise. During this continuous labeling process, there exists a range where there is no labeling distance less than a preset distance, and the number of labeled running vectors in the running vector's neighborhood reaches a preset number. The following is a summary of this process. Figure 3B The above-described vector marking process will be explained.

[0068] Figure 3B The illustration schematically shows intermediate results during the clustering process of a train communication quality detection method according to another embodiment of the present disclosure.

[0069] like Figure 3B As shown, the running vector 304 is marked in a two-dimensional vector space, and a neighborhood range is determined with running vector 304 as the center and a preset distance as the radius. At this time, it is determined that the number of running vectors that have been marked in the neighborhood range of running vector 304 has reached the preset number 5. That is, after inserting running vector 304, the running vector and other running vectors existing in its neighborhood range have met the requirements to form a cluster. Therefore, a new cluster, namely the third cluster 305, can be determined with running vector 304 as the cluster center, and the running vectors in the neighborhood range of the running vector are added to this cluster.

[0070] By identifying new clusters in the above manner, the labeling results of running vectors that were incorrectly classified as noise due to the limitation on the number of running vectors in the early stage can be corrected, thereby ensuring the correctness of the clustering results.

[0071] According to embodiments of this disclosure, an iterative clustering method that sequentially selects unlabeled vectors for labeling achieves unsupervised automatic clustering. It eliminates the need for pre-setting the number of clusters or complex prior rules, naturally dividing normal and abnormal clusters based on distance thresholds, thus reducing manual configuration costs. Through iterative operations, it ensures that all running vectors are reasonably classified, avoiding data omissions and improving the completeness of clustering results. Furthermore, during vector classification, hierarchical logic using distance judgment, neighborhood statistics, and noise labeling optimizes cluster generation and data filtering. Isolated running vectors can be labeled as noise, preventing them from interfering with the formation of normal clusters. Pre-setting a number of clusters ensures that newly generated clusters possess statistical significance, improving the reliability of clustering results and reducing detection errors caused by false clusters.

[0072] Specifically, when it is determined that the distance between the running vector and the labels of multiple labeled clusters is less than a preset distance, the multiple labeled clusters are merged to obtain an updated cluster, and the running vector is labeled into the updated cluster. The following is done through... Figure 3C The process of merging clusters is explained.

[0073] Figure 3C The illustration schematically shows intermediate results during the clustering process of the train communication quality detection method according to yet another embodiment of the present disclosure.

[0074] like Figure 3C As shown, the running vector 306 is marked in the two-dimensional vector space. It can be determined that the marking distance between the running vector 306 and the first cluster 301 and the second cluster 302 is less than the preset distance. Therefore, the first cluster 301 and the second cluster 302 can be merged to obtain the updated cluster 307, and the running vector 306 is marked in the updated cluster.

[0075] According to embodiments of this disclosure, cluster merging is performed on running vectors with overlapping clusters, resolving the problem of ambiguous data attribution. This avoids clustering confusion caused by the same running vector being repeatedly labeled by multiple clusters, resulting in a clearer cluster structure. Furthermore, since the merged clusters all cover the running vectors, they have high similarity. Merging these clusters allows the merged clusters to more comprehensively reflect the common characteristics of the device's operating status, improving the consistency of subsequent detection results.

[0076] According to embodiments of this disclosure, at least one cluster is updated based on multiple update running vectors determined by an update running parameter group to obtain a clustering update result, including: determining the cluster center of each of the multiple clusters; for each update running vector, determining the distance between the update running vector and at least one cluster center to obtain at least one update distance; and updating the update running vector to the nearest cluster based on the at least one update distance to obtain a clustering update result.

[0077] Distance-based clustering algorithms can be used to update clusters. This involves assigning the updated vector to the cluster with the closest distance. Specifically, an update distance threshold can be set; if the update distance between the updated vector and all clusters exceeds the threshold, the updated vector can be considered noise.

[0078] According to embodiments of this disclosure, clusters are updated based on the distance between the updated running vector and the cluster center, achieving efficient iteration of clusters. There is no need to recalculate the cluster centers of the entire dataset; only the new vector is assigned to the nearest cluster, significantly reducing the computational overhead of the update phase. Furthermore, clustering operations can be performed with only a small number of updated running vectors, enabling continuous monitoring of target network devices. The distance-first update logic ensures that new data is accurately integrated into the existing clustering system, maintaining the consistency of the clustering results.

[0079] According to embodiments of this disclosure, updating the update running vector to the nearest cluster based on at least one update distance to obtain a cluster update result includes: adding the update running vector to the cluster, and recalculating the average of multiple running vectors based on the update running vector and multiple running vectors in the cluster to update the cluster center of the cluster, thereby obtaining a cluster update result.

[0080] The average value in multiple dimensions can be determined based on the components of the updated running vector and multiple running vectors in the cluster in multiple dimensions, and the multidimensional vector determined based on the average value in multiple dimensions can be used as the cluster center.

[0081] According to embodiments of this disclosure, the average value of cluster centers is recalculated after adding and updating the running vector, achieving dynamic optimization of clusters. The cluster centers are adjusted in real time with new data, ensuring that the clusters always keep pace with the latest operating state of the equipment, avoiding detection lag caused by static cluster centers; the update method of the average value calculation enables the cluster centers to reflect the statistical characteristics of all data within the cluster, improving the representativeness of the clustering results and the accuracy of detection.

[0082] According to embodiments of this disclosure, determining the communication quality detection result for each port in the target network device based on the clustering update result and the clustering result includes: when the number of update running vectors representing the updated clusters in the clustering update result is greater than a threshold, determining the communication quality detection result for each port in the target network device based on the cluster type of the updated clusters, wherein the update running vectors of the updated clusters include the running vectors added to the updated clusters relative to the unupdated clusters; wherein the cluster type of the updated clusters is the same as the cluster type of the unupdated clusters; the cluster type of the unupdated clusters is determined based on the running parameter group corresponding to the running vectors and the reference running parameter group.

[0083] During the process of updating the clustering results, multiple update vectors corresponding to multiple update parameter groups within the second preset time period can be assigned to corresponding clusters to obtain updated clusters. Therefore, the number of update vectors in each updated cluster indicates how many update parameter groups conform to the cluster type of that updated cluster exist within the second preset time period.

[0084] During the second preset time period, the operating parameters of the target network device may also fluctuate occasionally. Therefore, a quantity threshold can be set. Only when the number of updated running vectors representing the updated clusters in the cluster update results is greater than the quantity threshold is it determined that the communication quality of the target network device during the second preset time period is the same as the communication quality represented by the updated clusters. That is, the communication quality detection results of each port in the target network device are determined based on the cluster type of the updated clusters.

[0085] The quantity threshold can be determined based on the number of moments within the second preset time period. For example, it can be set to 2 / 3 of the number of moments within the second preset time period. If the duration of the second preset time period is 60 seconds and the sampling period is 1 second, it can be determined that there are 60 moments within the second preset time period. Therefore, the quantity threshold can be set to 40. That is, if the number of update running vectors in the updated cluster represented by the cluster update result is greater than 40, it can be determined that the communication quality of the target network device is the same as the communication quality represented by the updated cluster in the second preset time period.

[0086] The cluster type of a cluster before the update can be determined based on the running parameter set and reference running parameter set corresponding to the running vectors in that cluster. Specifically, multiple running parameters included in the running parameter set and multiple reference running parameters included in the running parameter set can be identified, where there is a correspondence between the running parameters and reference running parameters for the same port. The running parameters corresponding to the reference running parameters are verified to identify anomalies in the multiple ports represented by the running parameter set. Based on the anomalies in the multiple ports, the cluster type can be determined.

[0087] For example, if among the multiple operating parameters in the operating parameter group, the operating parameters of the Ethernet port exceed the parameter range specified by the reference operating parameters of the Ethernet port, while the operating parameters of other ports are all within the parameter range specified by the reference operating parameters of each port, it can be determined that the operating parameter group indicates that the Ethernet port is abnormal. In this case, the cluster type can be used to indicate that the Ethernet port is in an abnormal state, while other ports are in a normal state.

[0088] According to embodiments of this disclosure, a quantity threshold is set, and it is determined whether the detection result is based on the updated clusters, thus balancing detection sensitivity and stability. The detection result is only updated when the newly added running vector reaches the threshold, avoiding misjudgments caused by a small amount of abnormal data. At the same time, it ensures that the clusters can reflect changes in the operating status of the target network device in a timely manner, making the communication quality detection results both reliable and dynamically adaptable.

[0089] After obtaining the updated clusters, independence tests can be performed based on the cluster types of the updated clusters to determine the communication quality of each port in the target network device represented by the cluster type, and thus obtain the communication quality test results of each port.

[0090] Specifically, when the cluster type of the updated cluster indicates that the ports of the target network devices belonging to the cluster are in a normal state, a communication quality detection result indicating that all ports of the target network devices are normal is generated. When the cluster type of the updated cluster indicates that the ports of the target network devices belonging to the cluster are in an abnormal state, the abnormal ports are determined based on the abnormal dimensions of the cluster and the correspondence between dimensions and ports. A communication quality detection result indicating that the abnormal ports of the target network devices are in an abnormal state is generated. The abnormal dimensions of the updated cluster are determined as follows: for multiple dimensions of the multidimensional vector space, the difference between the label values ​​of the updated cluster and the normal cluster in each dimension is determined to obtain the operating parameter offsets of the updated cluster and the normal cluster in multiple dimensions. The cluster type of the normal cluster indicates that the ports of the target network devices belonging to the normal cluster are in a normal state. If there is an operating parameter offset that reaches the offset threshold, the dimension where the operating parameter offset is located is determined as the abnormal dimension of the updated cluster.

[0091] Figure 4 The diagram illustrates a method for determining abnormal ports according to an embodiment of the present disclosure.

[0092] like Figure 4As shown, taking a three-dimensional vector space as an example, it includes four updated clusters: normal cluster 401, first updated cluster 402, second updated cluster 403, and third updated cluster 404. The three-dimensional vector space includes three dimensions: x, y, and z. The x-dimensional dimension is used to represent the operating parameters of the Ethernet port, the y-dimensional dimension is used to represent the operating parameters of the aggregation port, and the z-dimensional dimension is used to represent the operating parameters of the access port.

[0093] Depend on Figure 4 It can be seen that, compared with the normal cluster 401, the first updated cluster 402 has reached the offset threshold in the x-dimensional offset of the running parameters; compared with the normal cluster 401, the second updated cluster 403 has reached the offset threshold in the y-dimensional offset of the running parameters; and compared with the normal cluster 401, the third updated cluster 404 has reached the offset threshold in the z-dimensional offset of the running parameters.

[0094] After determining the abnormal dimensions of the first updated cluster 402, the second updated cluster 403, and the third updated cluster 404, the abnormal ports corresponding to each updated cluster can be determined according to the correspondence between ports and dimensions. That is, the abnormal port corresponding to the first updated cluster 402 is the Ethernet port, the abnormal port corresponding to the second updated cluster 403 is the aggregation port, and the abnormal port corresponding to the third updated cluster 404 is the access port.

[0095] Specifically, if the offset of the operating parameters in all dimensions reaches the offset threshold when the updated cluster is compared with the normal cluster, it can be determined that all ports of the target network device are abnormal, or that the target network device itself is abnormal. Since the target network device has a large number of ports, multiple ports typically do not malfunction simultaneously. Therefore, if the offset of the operating parameters in all dimensions reaches the offset threshold, it can be determined that the target network device itself is abnormal.

[0096] According to embodiments of this disclosure, abnormal dimensions are located by comparing the dimensional offsets of updated clusters with normal clusters, and abnormal ports are determined based on the correspondence between dimensions and ports, enabling precise source tracing of anomalies. For updated clusters with anomalies, communication quality detection results are generated according to cluster type and abnormal ports. For updated clusters without anomalies, the overall situation is directly fed back. This allows for precise identification of abnormal ports in abnormal states, improving the practicality of communication quality detection results. In normal states, communication quality detection results are output promptly, improving the feedback efficiency of communication quality detection.

[0097] Based on the above-described train communication quality detection method, this disclosure also provides a train communication quality detection device. The following will be combined with... Figure 5 The device is described in detail.

[0098] Figure 5 A schematic block diagram of a train communication quality detection device according to an embodiment of the present disclosure is shown.

[0099] like Figure 5 As shown, the train communication quality detection device 500 of this embodiment includes a vector projection module 510, a vector clustering module 520, a clustering update module 530, and a result determination module 540.

[0100] The vector projection module 510 is used to project multiple sets of operating parameters of the target network devices in the train into a multi-dimensional vector space based on the acquired sets of operating parameters for each time point within a first preset time period, thereby obtaining operating vectors for each time point. The sets of operating parameters include the operating parameters of multiple ports of the target network devices, and the dimensions of the multi-dimensional vector space are determined according to the number of ports of the target network devices. In one embodiment, the vector projection module 510 can be used to perform the operation S210 described above, which will not be repeated here.

[0101] The vector clustering module 520 is used to cluster multiple running vectors to obtain clustering results. The clustering results include at least one cluster and the cluster type, where the cluster type indicates whether the port of the target network device belonging to the cluster is in a normal or abnormal state. In one embodiment, the vector clustering module 520 can be used to perform the operation S220 described above, which will not be repeated here.

[0102] The clustering update module 530 is used to update at least one cluster in response to obtaining multiple update operation parameter groups for each time point within a second preset time period, and to update the clustering update result by determining multiple update operation vectors based on the update operation parameter groups. In one embodiment, the clustering update module 530 can be used to perform the operation S230 described above, which will not be repeated here.

[0103] The result determination module 540 is used to determine the communication quality detection result for each port in the target network device based on the clustering update result and the clustering result. In one embodiment, the result determination module 540 can be used to perform the operation S240 described above, which will not be repeated here.

[0104] According to embodiments of this disclosure, the vector projection module 510 includes a marker value determination submodule.

[0105] The tag value determination submodule is used to determine the tag values ​​of multiple dimensions in the running vector for each running parameter group, based on the mapping relationship between each port in the running parameter group and the dimensions of the running vector, and the running parameters of each port.

[0106] According to embodiments of this disclosure, the result determination module 540 includes a result determination submodule.

[0107] The result determination submodule is used to determine the communication quality detection results of each port in the target network device based on the cluster type of the updated cluster when the number of update running vectors of the updated cluster is greater than the number threshold. The update running vectors of the updated cluster include the running vectors added by the updated cluster relative to the unupdated cluster.

[0108] According to embodiments of this disclosure, the result determination submodule includes a first result determination unit, a port determination unit, and a second result determination unit.

[0109] The first result determination unit is used to generate a communication quality detection result indicating that all ports of the target network device are normal, provided that the cluster type of the updated cluster indicates that the ports of the target network device belonging to the cluster are in a normal state.

[0110] The port determination unit is used to determine the abnormal port based on the abnormal dimension of the cluster and the correspondence between the dimension and the port when the port of the target network device belonging to the cluster is in an abnormal state, according to the cluster type representation of the updated cluster.

[0111] The second result determination unit is used to generate communication quality detection results indicating that the abnormal port of the target network device is in an abnormal state.

[0112] According to embodiments of this disclosure, the vector clustering module 520 includes a vector acquisition submodule, a first distance determination submodule, a label determination submodule, and a vector clustering submodule.

[0113] The vector acquisition submodule is used to obtain an unlabeled running vector from multiple running vectors.

[0114] The first distance determination submodule is used to determine the distance between the cluster center of the labeled cluster and the running vector in the multidimensional vector space, as the label distance.

[0115] The label determination submodule is used to determine the labeling result of the running vector based on the label distance and the preset distance.

[0116] The vector clustering submodule is used to determine at least one cluster based on the labeling results of multiple running vectors, and to determine the clustering result.

[0117] According to embodiments of this disclosure, the label determination submodule includes a first labeling unit, a neighborhood determination unit, a cluster determination unit, and a second labeling unit.

[0118] The first labeling unit is used to label the running vector to the cluster to which the cluster center corresponding to the labeling distance less than the preset distance belongs when there is a labeling distance less than the preset distance.

[0119] The neighborhood determination unit is used to determine the neighborhood range of the running vector with the running vector as the center and the preset distance as the radius, in the absence of a marker distance less than the preset distance.

[0120] The cluster determination unit is used to determine a cluster by using the running vector as the cluster center when the number of marked running vectors in the neighborhood reaches a preset number.

[0121] The second marking unit is used to mark the running vector as noise when there is no marking distance less than a preset distance and the number of running vectors that have been marked in the neighborhood range has not reached the preset number.

[0122] According to embodiments of this disclosure, the train communication quality detection device 500 further includes a cluster merging module.

[0123] The cluster merging module is used to merge multiple labeled clusters to obtain an updated cluster when the distance between the running vector and multiple labeled clusters is less than a preset distance, and then mark the running vector into the updated cluster.

[0124] According to embodiments of this disclosure, the clustering update module 530 includes a center determination submodule, a second distance determination submodule, and a clustering update submodule.

[0125] The cluster center determination submodule is used to determine the cluster center of each of the multiple clusters.

[0126] The second distance determination submodule is used to determine the distance between the updated running vector and at least one cluster center for each updated running vector, thereby obtaining at least one updated distance.

[0127] The clustering update submodule is used to update the update running vector to the nearest cluster based on at least one update distance, so as to obtain the clustering update result.

[0128] According to embodiments of this disclosure, the clustering update submodule includes a clustering update unit.

[0129] The clustering update unit is used to add the update running vector to the cluster and, based on the update running vector and multiple running vectors in the cluster, recalculate the average of multiple running vectors to update the cluster centers and obtain the clustering update result.

[0130] According to embodiments of this disclosure, any plurality of modules among the vector projection module 510, vector clustering module 520, clustering update module 530, and result determination module 540 may be combined into one module, or any one of these modules may be split into multiple modules. Alternatively, at least part of the functionality of one or more of these modules may be combined with at least part of the functionality of other modules and implemented in one module. According to embodiments of this disclosure, at least one of the vector projection module 510, vector clustering module 520, clustering update module 530, and result determination module 540 may be at least partially implemented as hardware circuitry, such as a field-programmable gate array (FPGA), a programmable logic array (PLA), a system-on-a-chip, a system-on-a-substrate, a system-on-package, an application-specific integrated circuit (ASIC), or any other reasonable means of integrating or packaging circuitry, or implemented in software, hardware, or firmware, or in any suitable combination of any of these three implementation methods. Alternatively, at least one of the vector projection module 510, vector clustering module 520, clustering update module 530, and result determination module 540 may be implemented at least partially as a computer program module, which can perform corresponding functions when the computer program module is run.

[0131] Figure 6 A block diagram schematically illustrates an electronic device suitable for implementing a train communication quality detection method according to an embodiment of the present disclosure.

[0132] like Figure 6 As shown, an electronic device 600 according to an embodiment of this disclosure includes a processor 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage portion 608 into a random access memory (RAM) 603. The processor 601 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 601 may also include onboard memory for caching purposes. The processor 601 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of this disclosure.

[0133] RAM 603 stores various programs and data required for the operation of electronic device 600. Processor 601, ROM 602, and RAM 603 are interconnected via bus 604. Processor 601 performs various operations of the method flow according to embodiments of the present disclosure by executing programs in ROM 602 and / or RAM 603. It should be noted that the programs may also be stored in one or more memories other than ROM 602 and RAM 603. Processor 601 may also perform various operations of the method flow according to embodiments of the present disclosure by executing programs stored in said one or more memories.

[0134] According to embodiments of this disclosure, the electronic device 600 may further include an input / output (I / O) interface 605, which is also connected to a bus 604. The electronic device 600 may also include one or more of the following components connected to the I / O interface 605: an input section 606 including a keyboard, mouse, etc.; an output section 607 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN card, modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the I / O interface 605 as needed. A removable medium 611, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 610 as needed so that computer programs read from it can be installed into the storage section 608 as needed.

[0135] This disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The computer-readable storage medium carries one or more programs that, when executed, implement the method according to the embodiments of this disclosure.

[0136] According to embodiments of this disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium, such as including, but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this disclosure, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to embodiments of this disclosure, the computer-readable storage medium may include ROM 602 and / or RAM 603 and / or one or more memories other than ROM 602 and RAM 603 described above.

[0137] Embodiments of this disclosure also include a computer program product comprising a computer program containing program code for performing the methods shown in the flowchart. When the computer program product is run on a computer system, the program code is used to cause the computer system to implement the item recommendation method provided in the embodiments of this disclosure.

[0138] When the computer program is executed by the processor 601, it performs the functions defined in the system / apparatus of this disclosure embodiments. According to embodiments of this disclosure, the systems, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0139] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of signals over a network medium, and downloaded and installed via the communication section 609, and / or installed from the removable medium 611. The program code contained in the computer program can be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.

[0140] In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 609, and / or installed from the removable medium 611. When the computer program is executed by the processor 601, it performs the functions defined in the system of this disclosure embodiment. According to embodiments of this disclosure, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0141] According to embodiments of this disclosure, program code for executing the computer programs provided in embodiments of this disclosure can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, Python, "C", or similar programming languages. The program code can execute entirely on a user's computing device, partially on a user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0142] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0143] Those skilled in the art will understand that the features described in the various embodiments and / or claims of this disclosure can be combined or combined in various ways, even if such combinations or combinations are not explicitly described in this disclosure. In particular, the features described in the various embodiments and / or claims of this disclosure can be combined or combined in various ways without departing from the spirit and teachings of this disclosure. All such combinations and / or combinations fall within the scope of this disclosure.

[0144] The embodiments of this disclosure have been described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of this disclosure. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. The scope of this disclosure is defined by the appended claims and their equivalents. Various substitutions and modifications can be made by those skilled in the art without departing from the scope of this disclosure, and all such substitutions and modifications should fall within the scope of this disclosure.

Claims

1. A method for detecting train communication quality, characterized in that, The method includes: Based on the acquired operating parameter sets of the target network devices in the train at multiple times within a first preset time period, the multiple operating parameter sets are projected into a multi-dimensional vector space to obtain the operating vectors at multiple times. The operating parameter sets include the operating parameters of multiple ports of the target network devices, and the dimension of the multi-dimensional vector space is determined according to the number of ports of the target network devices. Clustering is performed on multiple running vectors to obtain clustering results, wherein the clustering results include at least one cluster and the cluster type of the cluster, and the cluster type indicates whether the port of the target network device belonging to the cluster is in a normal or abnormal state. In response to obtaining the update operation parameter groups for multiple times within a second preset time period, at least one of the clustering clusters is updated according to the multiple update operation vectors determined by the update operation parameter groups to obtain a clustering update result; and Based on the clustering update result and the clustering result, the communication quality detection result for each port in the target network device is determined.

2. The method according to claim 1, characterized in that, Each dimension in the multidimensional vector space corresponds to a port of the target network device; the running vector at each time step corresponds to a set of running parameters. The step of projecting the multiple sets of operating parameters into a multidimensional vector space to obtain operating vectors at multiple time points includes: For each of the running parameter groups, based on the mapping relationship between each port in the running parameter group and the dimension of the running vector, and the running parameters of each port, the respective label values ​​of multiple dimensions in the running vector are determined.

3. The method according to claim 2, characterized in that, The step of determining the communication quality detection result for each port in the target network device based on the clustering update result and the clustering result includes: When the number of update running vectors representing the updated clusters is greater than a threshold, the communication quality detection results of each port in the target network device are determined based on the cluster type of the updated clusters. The update running vectors of the updated clusters include the running vectors added by the updated clusters relative to the unupdated clusters. Wherein, the cluster type of the updated cluster is the same as the cluster type of the unupdated cluster; the cluster type of the unupdated cluster is determined based on the running parameter group corresponding to the running vector and the reference running parameter group.

4. The method according to claim 3, characterized in that, The step of determining the communication quality detection results of each port in the target network device based on the cluster type of the updated cluster includes: If the cluster type of the updated cluster indicates that the port of the target network device belonging to the cluster is in a normal state, a communication quality detection result indicating that all ports in the target network device are normal is generated. When the cluster type of the updated cluster indicates that the port of the target network device belonging to the cluster is in an abnormal state, the abnormal port is determined based on the abnormal dimension of the cluster and the correspondence between the dimension and the port; and Generate a communication quality detection result indicating that the abnormal port of the target network device is in an abnormal state; The outlier dimension of the updated cluster is determined in the following way: For the multiple dimensions of the multidimensional vector space, the difference between the label values ​​of the updated cluster and the normal cluster in the dimension is determined respectively, and the operating parameter offset of the updated cluster and the normal cluster in the multiple dimensions is obtained. The cluster type of the normal cluster indicates that the port of the target network device belonging to the normal cluster is in a normal state. If there is an offset in the running parameter that reaches the offset threshold, the dimension of the running parameter offset is determined as the abnormal dimension of the updated cluster.

5. The method according to claim 1, characterized in that, The clustering of multiple running vectors to obtain clustering results includes: Repeat the following steps until all of the stated running vectors have been labeled: From the multiple running vectors, obtain one unlabeled running vector; The distance between the cluster center of the labeled cluster in the multidimensional vector space and the running vector is determined as the labeling distance; Based on the labeled distance and the preset distance, the labeling result for the running vector is determined; and Based on the labeling results of each of the multiple running vectors, at least one cluster is determined, and the clustering result is determined.

6. The method according to claim 5, characterized in that, The step of determining the labeling result for the running vector based on the labeled distance and the preset distance includes: If there is a marker distance less than a preset distance, the running vector is marked to the cluster to which the cluster center corresponding to the marker distance less than the preset distance belongs; In the absence of a marker distance smaller than the preset distance, the neighborhood range of the running vector is determined with the running vector as the center and the preset distance as the radius. If the number of marked running vectors within the neighborhood reaches the preset number, then the running vectors are used as cluster centers to determine a cluster; and If there is no marker distance less than the preset distance, and the number of marked running vectors in the neighborhood range does not reach the preset number, the running vector is marked as noise.

7. The method according to claim 6, characterized in that, The method further includes: If it is determined that the running vector has a label distance of less than the preset distance with multiple labeled clusters, the multiple labeled clusters are merged to obtain an updated cluster, and the running vector is labeled into the updated cluster.

8. The method according to claim 1, characterized in that, The step of updating at least one cluster based on multiple update operation vectors determined by the update operation parameter group to obtain cluster update results includes: Determine the cluster center of each of the multiple clusters; For each of the update running vectors, determine the distance between the update running vector and at least one of the cluster centers to obtain at least one update distance; and Based on at least one update distance, the update running vector is updated to the nearest cluster to obtain the clustering update result.

9. The method according to claim 8, characterized in that, The step of updating the update vector to the nearest cluster based on at least one update distance to obtain the clustering update result includes: The updated running vector is added to the cluster, and the average value of multiple running vectors is recalculated based on the updated running vector and multiple running vectors in the cluster to update the cluster center, thus obtaining the cluster update result.

10. A train, characterized in that, The train includes: A target network device is installed inside the train, and the target network device includes multiple ports; A memory, used to store the operating parameters of each port of the target network device; A controller is configured to obtain the operating parameters of each port of the target network device from the memory, and perform the steps of the method as described in any one of claims 1 to 9 based on the operating parameters of each port of the target network device to obtain the communication quality prediction results of each port in the target network device.

Images