Base station detection method and apparatus

Abstract

The application provides a base station detection method and device, the method comprises the following steps: obtaining a first strong correlation index pair of a to-be-detected base station; inputting the first strong correlation index pair into at least two supervised learning models respectively to obtain a first detection result output by each supervised learning model; and determining whether the to-be-detected base station has a hidden problem according to all the first detection results. The base station detection method and device provided by the application first obtain a strong correlation index pair of a to-be-detected base station, then input the strong correlation index pair into at least two supervised learning models respectively to obtain a first detection result output by each supervised learning model, and finally determine whether the to-be-detected base station has a hidden problem according to all the first detection results, which does not depend on expert experience and improves the accuracy of detection of hidden problems of base stations.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and specifically to a base station detection method and apparatus. Background Technology

[0002] Hidden base station problems refer to instances where users experience abnormal wireless network usage without any warning signals. Identifying hidden base station problems before user complaints arise is a key daily task for wireless network optimization.

[0003] In the relevant plan, based on historical complaint data, relevant experts will analyze the data and set thresholds for various performance statistical indicators based on experience. Poor base stations that are below the threshold will be investigated on-site.

[0004] However, the accuracy of existing solutions in identifying hidden problems in base stations is closely related to the experience of experts. For new networks such as the 5th generation mobile communication (5G), it is difficult to select appropriate indicators and thresholds as the basis for judging hidden problems in base stations when relevant experience is insufficient. Summary of the Invention

[0005] This invention provides a base station detection method and apparatus to solve the technical problem of inaccurate detection of hidden problems in base stations.

[0006] In a first aspect, the present invention provides a base station detection method, comprising:

[0007] Obtain the first strongly correlated index pair of the base station under test;

[0008] The first strongly correlated index is input into at least two supervised learning models respectively to obtain the first detection result output by each supervised learning model;

[0009] Based on all the initial test results, determine whether the base station under test has any hidden problems.

[0010] In one embodiment, the method further includes a training step for the supervised learning model:

[0011] Obtain the second strongly correlated indicator pair of the sample base stations;

[0012] Cluster analysis was performed on the sample base stations based on the second strong correlation index to determine the target base station;

[0013] Based on the results of on-site testing of the target base station, the label corresponding to the target base station is determined, and training samples are obtained;

[0014] The supervised learning model is trained using the training samples.

[0015] In one embodiment, the step of performing cluster analysis on the sample base stations based on the second strong correlation index to determine the target base station includes:

[0016] Based on the second strongly correlated index pair, the K-Means algorithm is used to perform cluster analysis on the sample base stations to determine the number of sample base stations in each cluster.

[0017] The sample base stations in the target cluster are used as the target base stations, and the number of sample base stations in the target cluster is less than a first preset threshold.

[0018] In one embodiment, it also includes:

[0019] The third strongly correlated index pair was obtained from the results of on-site testing of the target base station;

[0020] The third strongly correlated index pair is added to the training samples to update the training samples.

[0021] In one embodiment, determining whether the base station under test has a hidden problem based on all the first detection results includes:

[0022] Determine the proportion of different results in the first detection result;

[0023] The presence of hidden problems in the base station under test is determined based on the result that the ratio value is greater than the second preset threshold.

[0024] In one embodiment, the first strongly correlated indicator includes at least one of the following:

[0025] The proportion of uplink quadrature phase shift keying (QPSK) modulation and the proportion of downlink QPSK modulation;

[0026] The proportion of uplink 256 quadrature amplitude modulation (QAM) and the proportion of downlink 256 QAM;

[0027] Uplink single-stream ratio and downlink single-stream ratio;

[0028] The proportion of upstream dual flows and the proportion of downstream four flows;

[0029] Uplink physical resource block (PRB) utilization and downlink PRB utilization;

[0030] Uplink average modulation and coding strategy (MCS) and downlink average MCS;

[0031] Media Access Control (MAC) layer uplink block error rate and MAC layer downlink block error rate;

[0032] Measurement reports show MR coverage and power margin as a percentage of PHR.

[0033] In a second aspect, the present invention provides a base station detection device, comprising:

[0034] The acquisition module is used to acquire the first strongly correlated indicator pair of the base station under test;

[0035] The processing module is used to input the first strongly correlated index pair into at least two supervised learning models respectively, and obtain the first detection result output by each supervised learning model;

[0036] The determination module is used to determine whether the base station under test has any hidden problems based on all the first detection results.

[0037] Thirdly, the present invention provides an electronic device, including a memory and a memory storing a computer program, wherein the processor executes the program to implement the steps of the base station detection method described in the first aspect.

[0038] Fourthly, the present invention provides a processor-readable storage medium storing a computer program for causing the processor to perform the steps of the base station detection method described in the first aspect.

[0039] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of any of the base station detection methods described above.

[0040] The base station detection method and apparatus provided by the present invention first obtains the strongly correlated index pairs of the base station to be tested, then inputs the strongly correlated index pairs into at least two supervised learning models respectively to obtain the first detection result output by each supervised learning model, and finally determines whether the base station to be tested has a hidden problem based on all the first detection results. This method does not rely on expert experience and improves the accuracy of detecting hidden problems in base stations. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0042] Figure 1 This is a flowchart illustrating the base station detection method provided by the present invention;

[0043] Figure 2 This is a flowchart of the base station detection logic provided by the present invention;

[0044] Figure 3 This is a schematic diagram of the base station detection device provided by the present invention;

[0045] Figure 4 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0046] Hidden base station problems refer to instances where users experience abnormal wireless network usage without any warning signals. Identifying hidden base station problems before user complaints arise is a key daily task for wireless network optimization.

[0047] In the relevant plan, based on historical complaint data, relevant experts analyze the data and set thresholds for various performance statistical indicators based on experience. Base stations that fall below these thresholds are then investigated on-site. The 5G base station indicator degradation thresholds are shown in Table 1.

[0048] Table 1. 5G Base Station Indicator Degradation Thresholds

[0049]

[0050]

[0051] However, the above solution has the following main drawbacks:

[0052] 1. The accuracy of identifying hidden problems in base stations is closely related to the experience of experts. For new networks such as 5G, it is difficult to select appropriate indicators and thresholds as the basis for judging hidden problems in base stations when relevant experience is insufficient.

[0053] 2. 5G technology is still evolving, and upgrades to base station versions and changes in business models can render previously established indicator thresholds inaccurate. Furthermore, different hidden problems affect different indicators; for new hidden problems, old indicators may not be applicable. Therefore, continuous summarization and improvement are necessary after handling user complaints.

[0054] 3. Due to the high speed and large capacity of 5G networks, even if performance degradation occurs, related indicators may not easily reflect abnormalities, making threshold selection crucial. For example, the proportion of Quadrature Phase Shift Keying (QPSK) in 5G base stations is related to both equipment malfunctions and user distribution. Some 5G base stations with equipment malfunctions may have similar QPSK proportions to those with poor coverage. If the threshold is defined too high, it is easy to miss detections and fail to identify problematic base stations in a timely manner; if the threshold is defined too low, it is easy to make false judgments and waste investigation resources.

[0055] The technical problems to be solved by this invention include: proposing a method and apparatus for investigating hidden problems in 5G base stations, integrating clustering and supervised learning algorithms and the selection and optimization of related algorithm parameters, so as to achieve the identification of hidden problems in 5G base stations without relying on manually defined indicator thresholds.

[0056] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0057] The terminology used in this embodiment of the invention is explained as follows:

[0058] Base station cell: also called base station cellular cell or cell, refers to the area covered by one base station or part of a base station (fan antenna) in a cellular mobile communication system, within which mobile stations can reliably communicate with the base station through a wireless channel.

[0059] Quadrature Phase Shift Keying (QPSK) ratio: QPSK is a digital modulation method. The ratio of QPSK modulation can reflect the coding efficiency when using radio resources, and indirectly reflect the quality of the radio channel.

[0060] Measurement Report (MR) Coverage: MR coverage reflects signal coverage capability.

[0061] Power Headroom Report (PHR) percentage: PHR represents how much transmit power the terminal / user equipment (UE) has left, that is, the difference between the current transmit power and the maximum allowed transmit power. The larger the margin, the lower the transmit power and the better the uplink quality.

[0062] Radio Remote Unit (RRU) Type: The function of the RRU is to convert the intermediate frequency (IF) signal to a radio frequency (RF) signal using the transceiver module, and then transmit the RF signal through the antenna port after passing through the power amplifier and filter module. The RRU consists of four main modules: IF module, transceiver module, power amplifier, and filter module.

[0063] 256 Quadrature Amplitude Modulation (QAM) Ratio: The 256QAM coding ratio reflects the coding efficiency when using radio resources, and indirectly reflects the quality of the radio channel.

[0064] Physical Resource Block (PRB) Utilization: The average utilization of all shared physical resource blocks (PDSCHPRB).

[0065] Average Modulation and Coding Scheme (MCS) Level: MCS determines the corresponding modulation method and transmission rate. The higher the MCS, the better the wireless environment, the higher the transmission rate, and the higher the coding efficiency.

[0066] Media Access Control (MAC) Layer: MAC, or Media Access Control layer, is a sublayer below the Data Link Layer in the OSI model. It defines how data frames are transmitted over the medium.

[0067] Figure 1 This is a flowchart illustrating the base station detection method provided by the present invention, as shown below. Figure 1 As shown, the present invention provides a base station detection method, the method comprising:

[0068] Step 101: Obtain the first strongly correlated index pair of the base station to be tested.

[0069] First, obtain the first strongly correlated indicator pair for the base station under test. A strongly correlated indicator pair consists of a pair of performance indicators with a strong correlation. When the base station has hidden problems, the correlation of the strongly correlated indicator pair is easily disrupted.

[0070] In some embodiments, the first strongly correlated indicator includes at least one of the following:

[0071] The proportion of uplink quadrature phase shift keying (QPSK) modulation and the proportion of downlink QPSK modulation;

[0072] The proportion of uplink 256 quadrature amplitude modulation (QAM) and the proportion of downlink 256 QAM;

[0073] Uplink single-stream ratio and downlink single-stream ratio;

[0074] The proportion of upstream dual flows and the proportion of downstream four flows;

[0075] Uplink physical resource block (PRB) utilization and downlink PRB utilization;

[0076] Uplink average modulation and coding strategy (MCS) and downlink average MCS;

[0077] Media Access Control (MAC) layer uplink block error rate and MAC layer downlink block error rate;

[0078] Measurement reports show MR coverage and power margin as a percentage of PHR.

[0079] Step 102: Input the first strongly correlated index pair into at least two supervised learning models respectively, and obtain the first detection result output by each supervised learning model.

[0080] Specifically, after obtaining the first strongly correlated index pair of the base station to be tested, the strongly correlated index pair is input into at least two supervised learning models to obtain the first detection result output by each supervised learning model.

[0081] This supervised learning model is a pre-trained supervised learning model.

[0082] The supervised learning algorithms in the embodiments of this invention may include, but are not limited to, decision trees, Bayesian algorithms, and neural networks.

[0083] Decision tree algorithms can be tuned using the minimum description length (MDL) pruning method.

[0084] The Bayesian algorithm can be optimized using parameters such as default probability, minimum standard deviation, and threshold standard deviation.

[0085] The neural network algorithm can use a fully connected neural network, and be optimized with 3 hidden layers and 30 hidden neurons per layer.

[0086] The first test result includes two scenarios: the presence of hidden problems and the absence of hidden problems.

[0087] Step 103: Determine whether the base station under test has any hidden problems based on all the first detection results.

[0088] In some embodiments, determining whether the base station under test has a hidden problem based on all the first detection results includes:

[0089] Determine the proportion of different results in the first detection result;

[0090] The presence of hidden problems in the base station under test is determined based on the result that the ratio value is greater than the second preset threshold.

[0091] Specifically, in this embodiment, multiple supervised learning models are used for prediction. After determining the prediction results (first detection results) of each supervised learning model, the proportion of different results in the prediction results is analyzed. Based on the proportion of results greater than a second preset threshold, it is determined whether the base station under test has a hidden problem. The result with a proportion greater than the second preset threshold is taken as the final detection result of the base station under test. The second preset threshold can be configured according to the actual application.

[0092] For example, if more than 2 / 3 of the supervised learning models predict the true value, and the rest predict the false value, then the final detection result of the base station under test is the true value. In this case, select the base station in which more than 2 / 3 of the supervised learning models predict the true value to carry out field testing, so as to improve the accuracy and the accuracy of the true value and avoid wasting test resources.

[0093] For example, Table 2 shows the comparison between the prediction results of a multi-supervised learning model for a certain category and the results of on-site testing and investigation. If only a single supervised learning algorithm is used, the highest accuracy and true value accuracy are 88.9% and 75.0%, respectively. When using a multi-supervised learning algorithm (more than 2 / 3 of the predictions are true), the accuracy and true value accuracy reach 94.4% and 80.0%, respectively.

[0094] Table 2 Comparison of On-site Testing and Supervised Learning Predictions

[0095]

[0096]

[0097] Figure 2 This is a flowchart of the base station detection logic provided by the present invention, such as... Figure 2 As shown, after determining whether the target base station has hidden problems, on-site testing is conducted, and true / false value labels are fed back. The above process is continuously iterated to increase the training samples for supervised learning, thereby improving the effectiveness of supervised learning prediction, realizing the iterative optimization of the base station hidden problem detection device, and continuously improving the prediction accuracy and true value accuracy.

[0098] The base station detection method provided by this invention first obtains the strongly correlated index pairs of the base station to be tested, then inputs the strongly correlated index pairs into at least two supervised learning models respectively to obtain the first detection result output by each supervised learning model, and finally determines whether the base station to be tested has a hidden problem based on all the first detection results. This method does not rely on expert experience and improves the accuracy of detecting hidden problems in base stations.

[0099] In some embodiments, the method further includes a training step for the supervised learning model:

[0100] Obtain the second strongly correlated indicator pair of the sample base stations;

[0101] Cluster analysis was performed on the sample base stations based on the second strong correlation index to determine the target base station;

[0102] Based on the results of on-site testing of the target base station, the label corresponding to the target base station is determined, and training samples are obtained;

[0103] The supervised learning model is trained using the training samples.

[0104] Specifically, at least one pair of strongly correlated indicators for each sample base station is first obtained, namely the second strongly correlated indicator pair.

[0105] Under normal circumstances, when the distribution of users within the coverage area of ​​a base station does not change significantly, certain indicators will show a strong correlation.

[0106] For example, changes in traffic volume will be correlated with changes in utilization; under conditions of good coverage, the proportion of uplink and downlink QPSK will be low.

[0107] For example, changes in base station utilization and the increase in base station traffic, the proportion of uplink QPSK and the proportion of downlink QPSK, etc.

[0108] Figure 2 This is a flowchart of the base station detection logic provided by the present invention, such as... Figure 2 As shown, when a base station has hidden problems, its correlation is easily disrupted. Clustering strongly correlated indicator pairs can identify abnormal base stations. Therefore, at least one strongly correlated indicator pair should be selected first.

[0109] The second strongly correlated indicator mentioned above may include, but is not limited to, the ratio of the normalized peak hour utilization rate of the base station to the peak hour utilization rate of this week, the ratio of the normalized peak hour traffic of the base station to the peak hour traffic of this week, the proportion of peak hour uplink QPSK and peak hour downlink QPSK, MR coverage and PHR, etc.

[0110] After obtaining at least one pair of strongly correlated indicators for each base station, these strongly correlated indicator pairs can be normalized.

[0111] For example, Z-Score standardization can be used for normalization, i.e., new data = (original data - mean) / standard deviation.

[0112] Normalization of strongly correlated indicators further improves the accuracy of base station detection.

[0113] After obtaining at least one pair of strongly correlated indicators for the sample base stations, cluster analysis is performed on all sample base stations to identify the target base station; the target base station is the base station that may have hidden problems.

[0114] In some embodiments, the step of performing cluster analysis on the sample base stations based on the second strong correlation index to determine the target base station includes:

[0115] Based on the second strongly correlated index pair, the K-Means algorithm is used to perform cluster analysis on the sample base stations to determine the number of sample base stations in each cluster.

[0116] The sample base stations in the target cluster are used as the target base stations, and the number of sample base stations in the target cluster is less than a first preset threshold.

[0117] Specifically, based on the second strong correlation index pair, the K-Means algorithm can be used to perform cluster analysis on base stations without alarms across the entire network. When the number of clusters, i.e., the K value, is chosen appropriately, base stations with the same latent problems are likely to be clustered into one category. Furthermore, since base stations with latent problems account for a small proportion of the entire network, the category with fewer samples has a higher probability of being abnormal.

[0118] In this embodiment, base stations in clusters with fewer than a first preset threshold are considered as potentially problematic base stations. The first preset threshold can be configured according to the actual application.

[0119] For example, base stations in the cluster with the fewest base stations can be selected as potentially problematic. Alternatively, base stations in clusters with fewer than n samples (n = total number of samples / 2K, where K is the number of clusters) can be selected as potentially problematic. Another approach is to first sort the base stations in each cluster in ascending order of the number of base stations, and then select the base stations from the top-ranked clusters as potentially problematic. The specific number of clusters to select can be determined based on the capabilities of the on-site testing; examples are not provided here.

[0120] Additionally, it should be noted that the K-Means algorithm is used as an example for cluster analysis in this embodiment of the invention. The cluster analysis algorithm is not limited to the K-Means algorithm, and other classification algorithms can also be used for example analysis, which will not be given here.

[0121] The base stations are clustered using K-Means. The algorithm implementation steps are as follows:

[0122] (1) First, select the number of clusters, i.e., the K value. Since the difference between the number of base stations with hidden problems and the total number of base stations in the network is more than two orders of magnitude, if the K value is too small, it will be difficult to cluster abnormal base stations. Assuming that the total number of base stations in the network is N, it is recommended that the K value be between N / 10 and N / 50. After selecting the K value, randomly initialize the center points of each class.

[0123] (2) Calculate the distance from each data point to the center point. Assign the data point to the category closest to the center point, resulting in K clusters {S1, S2, S3, ..., S...}. K}

[0124] (3) After the first clustering, the centroids of each cluster are selected as the new centroids. The centroids are calculated as follows:

[0125]

[0126] Among them, C t Let |S| represent the center of the t-th cluster, 1≤t≤K, |S| t| represents the number of objects in the t-th cluster, X i This represents the i-th object in the t-th cluster, where 1 ≤ i ≤ |S|. t |

[0127] (4) Repeat the above steps until the change in the cluster center point is very small, or the specified number of iterations M is reached. In this embodiment of the invention, the calculation ends when the two classification results are the same or M exceeds the specified number of iterations.

[0128] After identifying target base stations that may have hidden problems, it is necessary to conduct on-site tests on these base stations to determine their true and false value labels and generate training samples.

[0129] In some embodiments, it also includes:

[0130] The third strongly correlated index pair was obtained from the results of on-site testing of the target base station;

[0131] The third strongly correlated index pair is added to the training samples to update the training samples.

[0132] Specifically, in this embodiment, by adding other strongly correlated indicator pairs of the base station (the third strongly correlated indicator pair), the feature values ​​corresponding to the label are enriched, which is more conducive to supervised learning training and prediction. The base station feature data applied to supervised learning in this embodiment includes RRU type, indoor distribution system coverage mode, and the following performance indicators (all taken as base station busy time):

[0133] Uplink QPSK percentage, uplink 256QAM percentage, downlink QPSK percentage, downlink 256QAM percentage, uplink single-stream percentage, uplink dual-stream percentage, downlink single-stream percentage, downlink four-stream percentage, uplink PRB utilization, downlink PRB utilization, uplink average MCS, downlink average MCS, MAC layer uplink block error rate, MAC layer downlink block error rate, MR coverage, PHR percentage, etc.

[0134] The base station detection device provided by the present invention is described below. The base station detection device described below can be referred to in correspondence with the base station detection method described above.

[0135] Figure 3 This is a schematic diagram of the base station detection device provided by the present invention, as shown below. Figure 3 As shown, the present invention provides a base station detection device, including an acquisition module 301, a processing module 302, and a determination module 303, wherein:

[0136] The acquisition module 301 is used to acquire the first strongly correlated index pair of the base station under test; the processing module 302 is used to input the first strongly correlated index pair into at least two supervised learning models respectively to obtain the first detection result output by each supervised learning model; the determination module 303 is used to determine whether the base station under test has a hidden problem based on all the first detection results.

[0137] It should be noted that the base station detection device provided in this embodiment of the invention can implement all the method steps implemented in the above method embodiment and can achieve the same technical effect. Therefore, the parts and beneficial effects that are the same as those in the method embodiment will not be described in detail here.

[0138] Figure 4 This is a schematic diagram of the structure of the electronic device provided by the present invention, such as... Figure 4 As shown, the electronic device may include: a processor 410, a communication interface 420, a memory 430, and a communication bus 440, wherein the processor 410, the communication interface 420, and the memory 430 communicate with each other via the communication bus 440. The processor 410 can call a computer program stored in the memory 430 to execute the steps of the base station detection method, such as including:

[0139] Obtain the first strongly correlated index pair of the base station under test; input the first strongly correlated index pair into at least two supervised learning models respectively, and obtain the first detection result output by each supervised learning model; determine whether the base station under test has a hidden problem based on all the first detection results.

[0140] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0141] On the other hand, the present invention also provides a computer program product, the computer program product comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions, which, when executed by a computer, enable the computer to perform the steps of the base station detection methods provided by the above methods, for example including:

[0142] Obtain the first strongly correlated index pair of the base station under test; input the first strongly correlated index pair into at least two supervised learning models respectively, and obtain the first detection result output by each supervised learning model; determine whether the base station under test has a hidden problem based on all the first detection results.

[0143] On the other hand, embodiments of this application also provide a processor-readable storage medium storing a computer program for causing the processor to perform the steps of the methods provided in the above embodiments, such as including:

[0144] Obtain the first strongly correlated index pair of the base station under test; input the first strongly correlated index pair into at least two supervised learning models respectively, and obtain the first detection result output by each supervised learning model; determine whether the base station under test has a hidden problem based on all the first detection results.

[0145] The processor-readable storage medium can be any available medium or data storage device that the processor can access, including but not limited to magnetic memory (e.g., floppy disk, hard disk, magnetic tape, magneto-optical disk (MO)), optical memory (e.g., CD, DVD, BD, HVD), and semiconductor memory (e.g., ROM, EPROM, EEPROM, non-volatile memory (NAND FLASH), solid-state drive (SSD)).

[0146] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0147] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0148] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A base station detection method, characterized in that, include: Obtain the first strongly correlated index pair of the base station under test; the first strongly correlated index pair consists of a pair of performance indicators that have a strong correlation relationship. The first strongly correlated index is input into at least two supervised learning models respectively to obtain the first detection result output by each supervised learning model; Based on all the initial detection results, determine whether the base station under test has any hidden problems; The first strongly correlated indicator pair includes: The proportion of uplink quadrature phase shift keying (QPSK) modulation and the proportion of downlink QPSK modulation; The proportion of uplink 256 quadrature amplitude modulation (QAM) and the proportion of downlink 256 QAM; Uplink single-stream ratio and downlink single-stream ratio; The proportion of upstream dual flows and the proportion of downstream four flows; Uplink physical resource block (PRB) utilization and downlink PRB utilization; Uplink average modulation and coding strategy (MCS) and downlink average MCS; Media Access Control (MAC) layer uplink block error rate and MAC layer downlink block error rate; Measurement reports show MR coverage and power margin as a percentage of PHR.

2. The base station detection method according to claim 1, characterized in that, It also includes the training steps for the supervised learning model: Obtain the second strongly correlated indicator pair of the sample base stations; Cluster analysis was performed on the sample base stations based on the second strong correlation index to determine the target base station; Based on the results of on-site testing of the target base station, the label corresponding to the target base station is determined, and training samples are obtained; The supervised learning model is trained using the training samples.

3. The base station detection method according to claim 2, characterized in that, The step of performing cluster analysis on the sample base stations based on the second strong correlation index to determine the target base station includes: Based on the second strongly correlated index pair, the K-Means algorithm is used to perform cluster analysis on the sample base stations to determine the number of sample base stations in each cluster. The sample base stations in the target cluster are used as the target base stations, and the number of sample base stations in the target cluster is less than a first preset threshold.

4. The base station detection method according to claim 2, characterized in that, Also includes: The third strongly correlated index pair was obtained from the results of on-site testing of the target base station; The third strongly correlated index pair is added to the training samples to update the training samples.

5. The base station detection method according to claim 1, characterized in that, The step of determining whether the base station under test has a hidden problem based on all the first detection results includes: Determine the proportion of different results in the first detection result; The presence of hidden problems in the base station under test is determined based on the result that the ratio value is greater than the second preset threshold.

6. A base station detection device, characterized in that, include: The acquisition module is used to acquire the first strongly correlated indicator pair of the base station under test; the first strongly correlated indicator pair consists of a pair of performance indicators that have a strong correlation relationship. The processing module is used to input the first strongly correlated index pair into at least two supervised learning models respectively, and obtain the first detection result output by each supervised learning model; The determination module is used to determine whether the base station under test has any hidden problems based on all the first detection results; The first strongly correlated indicator pair includes: uplink quadrature phase shift keying (QPSK) ratio and downlink QPSK ratio; uplink 256 quadrature amplitude modulation (QAM) ratio and downlink 256 QAM ratio; uplink single-stream ratio and downlink single-stream ratio; uplink dual-stream ratio and downlink quad-stream ratio; uplink physical resource block (PRB) utilization rate and downlink PRB utilization rate; uplink average modulation and coding scheme (MCS) and downlink average MCS; uplink block error rate and downlink block error rate at the Media Access Control (MAC) layer; and measurement report (MR) coverage rate and power headroom (PHR) reporting ratio.

7. An electronic device comprising a processor and a memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the base station detection method according to any one of claims 1 to 5.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the base station detection method as described in any one of claims 1 to 5.

9. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the base station detection method as described in any one of claims 1 to 5.

Images