CN117216573A - Method and device for detecting interference - Google Patents
Method and device for detecting interference Download PDFInfo
- Publication number
- CN117216573A CN117216573A CN202311283734.0A CN202311283734A CN117216573A CN 117216573 A CN117216573 A CN 117216573A CN 202311283734 A CN202311283734 A CN 202311283734A CN 117216573 A CN117216573 A CN 117216573A
- Authority
- CN
- China
- Prior art keywords
- interference
- data
- sample
- interference data
- frequency domain
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 45
- 238000012549 training Methods 0.000 claims abstract description 41
- 238000013507 mapping Methods 0.000 claims abstract description 19
- 238000012545 processing Methods 0.000 claims abstract description 19
- 238000001228 spectrum Methods 0.000 claims abstract description 15
- 238000002372 labelling Methods 0.000 claims description 4
- 238000003062 neural network model Methods 0.000 claims 2
- 238000000638 solvent extraction Methods 0.000 claims 1
- 238000013527 convolutional neural network Methods 0.000 description 13
- 238000010586 diagram Methods 0.000 description 5
- 238000012935 Averaging Methods 0.000 description 4
- 238000001514 detection method Methods 0.000 description 4
- 238000010801 machine learning Methods 0.000 description 4
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000004140 cleaning Methods 0.000 description 2
- 238000011478 gradient descent method Methods 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000003909 pattern recognition Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000003066 decision tree Methods 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24133—Distances to prototypes
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
- Noise Elimination (AREA)
Abstract
The embodiment of the invention discloses a method and a device for detecting interference, wherein the method comprises the following steps: performing interference type marking processing on the collected sample interference data by utilizing a plurality of classifier voting modes to obtain interference data marked with the interference types; obtaining a time-frequency domain mapped interference map by performing time-frequency domain mapping processing on the interference data marked with the interference types; training the time-frequency domain mapped interference spectrum to obtain a trained interference model; and classifying the interference type of the interference data to be verified in the external field by using the trained interference model to obtain the interference type of the interference data to be verified in the external field.
Description
Description of the division
The scheme is a divisional application of an invention patent application with the application number of 201810693247.4 and the application date of 29 days of 2018 6 month, named as a method and a device for detecting interference.
Technical Field
The invention relates to the technical field of LTE (Long Term Evoltion, long term evolution) wireless performance, in particular to a method and a device for detecting interference.
Background
In the past, the interference is checked on site by a spectrometer or the interference type is judged by drawing the waveform of the interference, but if the interference is more or the sites are more, the interference checking method is troublesome and consumes labor. Because of the large number of interference types of the external field, the interference form is complex, so that the manual detection efficiency is low, and the integral concept is difficult to form;
currently existing automated analysis requires the human to identify each interfering feature and to identify it in a programmed manner. However, although some interference features can be identified, it is sometimes difficult to identify such features by encoding.
Disclosure of Invention
The technical problem solved by the scheme provided by the embodiment of the invention is that the interference discrimination is very complicated, and the interference characteristics are required to be extracted manually.
The method for detecting interference provided by the embodiment of the invention comprises the following steps:
performing interference type marking processing on the collected sample interference data by utilizing a plurality of classifier voting modes to obtain interference data marked with the interference types;
obtaining a time-frequency domain mapped interference map by performing time-frequency domain mapping processing on the interference data marked with the interference types;
training the time-frequency domain mapped interference spectrum to obtain a trained interference model;
and classifying the interference type of the interference data to be verified in the external field by using the trained interference model to obtain the interference type of the interference data to be verified in the external field.
Preferably, the marking of the interference type is performed on the collected sample interference data by using multiple classifier voting modes, and obtaining the interference data marked with the interference type includes:
respectively carrying out initial marking on the collected sample interference data of various interference types to obtain sample interference data with initial marking;
marking the sample interference data as positive interference marked sample interference data, other interference marked sample interference data and interference-free marked sample interference data according to the interference energy intensity of the sample interference data with the same initial mark;
and finally marking the sample interference data of the positive interference mark, the sample interference data of other interference marks and the sample interference data without interference marks in a classifier voting mode to obtain interference data with marked interference types.
Preferably, the final marking of the sample interference data with the positive interference mark, the sample interference data with other interference marks and the sample interference data without interference marks by means of voting by a classifier includes:
randomly dividing the sample interference data with the same initial mark to obtain a plurality of interference subgroups containing the sample interference data with the positive interference mark, the sample interference data with other interference marks and the sample interference data without interference marks;
generating a classifier for voting on the sample interference data by training each interference group;
and finally marking the sample interference data of the positive interference mark, the sample interference data of other interference marks and the sample interference data without interference marks by using the classifier to obtain interference data with marked interference types.
Preferably, the method further comprises:
and when the interference data to be verified in the external field does not accord with the interference type in the trained interference model, carrying out recursive training on the interference model again according to the interference type of the interference data to be verified in the external field to obtain and store a new interference model.
According to an embodiment of the present invention, an apparatus for detecting interference includes:
the marking module is used for marking the collected sample interference data in an interference type by utilizing a plurality of classifier voting modes to obtain interference data marked with the interference type;
the time-frequency domain mapping processing module is used for obtaining a time-frequency domain mapped interference map by performing time-frequency domain mapping processing on the interference data marked with the interference types;
the training module is used for training the time-frequency domain mapped interference spectrum to obtain a trained interference model;
and the interference classification module is used for classifying the interference type of the external field interference data to be verified by utilizing the trained interference model to obtain the interference type of the external field interference data to be verified.
Preferably, the marking module includes:
the initial marking unit is used for respectively carrying out initial marking on the collected sample interference data of various interference types to obtain sample interference data with initial marking, and marking the sample interference data into sample interference data with positive interference marking, sample interference data with other interference marking and sample interference data without interference marking according to the interference energy intensity of the sample interference data with the same initial marking;
and the final marking unit is used for finally marking the sample interference data of the positive interference mark, the sample interference data of other interference marks and the sample interference data without interference marks in a classifier voting mode to obtain interference data with marked interference types.
Preferably, the final marking unit includes:
the dividing subunit is used for randomly dividing the sample interference data with the same initial mark to obtain a plurality of interference subgroups containing the sample interference data with the positive interference mark, the sample interference data with other interference marks and the sample interference data without interference marks;
a training subunit, configured to generate a classifier for voting on the sample interference data by training each interference group;
and the final marking subunit is used for finally marking the sample interference data of the positive interference mark, the sample interference data of other interference marks and the sample interference data without interference marks by using the classifier to obtain interference data with marked interference types.
Preferably, the training module is further configured to, when the interference data to be verified in the external field does not conform to the interference type in the trained interference model, re-perform recursive training on the interference model according to the interference type of the interference data to be verified in the external field, so as to obtain and store a new interference model.
According to an embodiment of the present invention, there is provided an apparatus for detecting interference, including: a processor, and a memory coupled to the processor; the memory stores a program for detecting interference, which can be run on the processor, and the program for detecting interference, when executed by the processor, realizes the steps of the method for detecting interference provided by the embodiment of the invention.
According to the embodiment of the invention, a computer storage medium is provided, and stores a program for detecting interference, and when the program for detecting interference is executed by a processor, the steps of the method for detecting interference provided according to the embodiment of the invention are realized.
According to the scheme provided by the embodiment of the invention, whether interference exists or not can be detected through the spectrum identification of NI (Noise Interference ), so that the interference of a network is checked, the network performance is improved, the system performance such as LTE is better optimized, the user experience of UE is improved, and the user perception is increased.
Drawings
FIG. 1 is a flow chart of a method for detecting interference according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of an apparatus for detecting interference according to an embodiment of the present invention;
FIG. 3 is a flowchart of a method for detecting interference in accordance with an embodiment of the present invention;
FIG. 4 is a schematic diagram of a time-frequency domain provided by an embodiment of the present invention;
fig. 5 is a schematic diagram of radio and television interference provided by an embodiment of the present invention;
fig. 6 is a schematic diagram of atmospheric waveguide interference according to an embodiment of the present invention.
Detailed Description
The following detailed description of the preferred embodiments of the present invention is provided in conjunction with the accompanying drawings, and it is to be understood that the preferred embodiments described below are merely illustrative and explanatory of the invention, and are not restrictive of the invention.
Fig. 1 is a flowchart of a method for detecting interference according to an embodiment of the present invention, as shown in fig. 1, including:
step S101: performing interference type marking processing on the collected sample interference data by utilizing a plurality of classifier voting modes to obtain interference data marked with the interference types;
step S102: obtaining a time-frequency domain mapped interference map by performing time-frequency domain mapping processing on the interference data marked with the interference types;
step S103: training the time-frequency domain mapped interference spectrum to obtain a trained interference model;
step S104: and classifying the interference type of the interference data to be verified in the external field by using the trained interference model to obtain the interference type of the interference data to be verified in the external field.
Wherein, the step S101 includes: respectively carrying out initial marking on the collected sample interference data of various interference types to obtain sample interference data with initial marking; marking the sample interference data as positive interference marked sample interference data, other interference marked sample interference data and interference-free marked sample interference data according to the interference energy intensity of the sample interference data with the same initial mark; and finally marking the sample interference data of the positive interference mark, the sample interference data of other interference marks and the sample interference data without interference marks in a classifier voting mode to obtain interference data with marked interference types.
Specifically, the final marking of the sample interference data of the positive interference mark, the sample interference data of other interference marks and the sample interference data without interference marks by means of voting through a classifier includes: randomly dividing the sample interference data with the same initial mark to obtain a plurality of interference subgroups containing the sample interference data with the positive interference mark, the sample interference data with other interference marks and the sample interference data without interference marks; generating a classifier for voting on the sample interference data by training each interference group; and finally marking the sample interference data of the positive interference mark, the sample interference data of other interference marks and the sample interference data without interference marks by using the classifier to obtain interference data with marked interference types.
The embodiment of the invention also comprises the following steps: and when the interference data to be verified in the external field does not accord with the interference type in the trained interference model, carrying out recursive training on the interference model again according to the interference type of the interference data to be verified in the external field to obtain and store a new interference model.
Fig. 2 is a schematic diagram of an apparatus for detecting interference according to an embodiment of the present invention, as shown in fig. 2, including: a marking module 201, a time-frequency domain mapping processing module 202, a training module 203 and an interference classification module 204.
The marking module 201 is configured to perform marking processing of an interference type on the collected sample interference data by using multiple classifier voting modes, so as to obtain interference data marked with the interference type;
the time-frequency domain mapping processing module 202 is configured to obtain a time-frequency domain mapped interference spectrum by performing time-frequency domain mapping processing on the interference data with the marked interference type;
the training module 203 is configured to obtain a trained interference model by training the time-frequency domain mapped interference spectrum;
the interference classification module 204 is configured to classify an interference type of the outfield to-be-verified interference data by using the trained interference model, so as to obtain an interference type of the outfield to-be-verified interference data.
Wherein the marking module 201 comprises: the initial marking unit is used for respectively carrying out initial marking on the collected sample interference data of various interference types to obtain sample interference data with initial marking, and marking the sample interference data into sample interference data with positive interference marking, sample interference data with other interference marking and sample interference data without interference marking according to the interference energy intensity of the sample interference data with the same initial marking; and the final marking unit is used for finally marking the sample interference data of the positive interference mark, the sample interference data of other interference marks and the sample interference data without interference marks in a classifier voting mode to obtain interference data with marked interference types.
Specifically, the final marking unit includes: the dividing subunit is used for randomly dividing the sample interference data with the same initial mark to obtain a plurality of interference subgroups containing the sample interference data with the positive interference mark, the sample interference data with other interference marks and the sample interference data without interference marks; a training subunit, configured to generate a classifier for voting on the sample interference data by training each interference group; and the final marking subunit is used for finally marking the sample interference data of the positive interference mark, the sample interference data of other interference marks and the sample interference data without interference marks by using the classifier to obtain interference data with marked interference types.
The training module 203 is further configured to, when the interference data to be verified in the external field does not conform to the interference type in the trained interference model, re-perform recursive training on the interference model according to the interference type of the interference data to be verified in the external field, so as to obtain and store a new interference model.
The device for detecting interference provided by the embodiment of the invention is characterized by comprising the following components: a processor, and a memory coupled to the processor; the memory stores a program for detecting interference, which can be run on the processor, and the program for detecting interference, when executed by the processor, realizes the steps of the method for detecting interference provided by the embodiment of the invention.
The embodiment of the invention provides a computer storage medium, which is characterized in that the computer storage medium stores a program for detecting interference, and the program for detecting interference realizes the steps of the method for detecting interference provided by the embodiment of the invention when being executed by a processor.
Fig. 3 is a flowchart of a method for specifically detecting interference according to an embodiment of the present invention, as shown in fig. 3, including:
step 1: various outfield NI uplink interference data are collected.
Step 2: the different types of interference data are marked, a few samples in some categories can be marked in a few-sample mode, and then all samples are marked in a multi-classifier voting mode.
For multi-classifier voting, it should be noted that:
firstly, because of a large number of unlabeled samples of the external field, if the numerous unlabeled samples harvested for each specific scene are labeled by people, the labels which need to be labeled are too many, are difficult to be manually completed, and can lead to fatigue judgment and human error.
Therefore, a mode of voting by using multiple classifiers is adopted;
firstly, a large number of samples are averaged on a cell level by adopting a cell merging mode from a plurality of unlabeled samples, and the samples are labeled in a manual observation mode after being averaged so as to be divided into a plurality of categories, namely A, B, C and the like;
two, in either category, a is assumed. The class a is a sub-sample space containing a plurality of unlabeled samples, and then the samples with the strongest energy (the threshold needs to be set according to the specific situation) are labeled as positive labels (interference labels) or no interference labels in the class respectively.
And thirdly, randomly dividing the class A subspace into a plurality of small spaces, namely a1, a 2..an (which can be overlapped or not and have the same size as possible), wherein n is determined according to the situation, generally 5, each newly divided space comprises a certain positive sample and a certain unmarked sample, one classifier is trained by using the marks, and n subspaces generate n classifiers.
Fourth, for each large class A, B, C, etc., voting is scored by using the obtained multi-classifier respectively,
the total division is n, n-2/3 n is positive sample mark, 2/3 n-1/3 n is other category mark, 1/3 n-0 is no mark category.
By means of multi-classifier voting, some new classes other than the current sample labels can be identified here and labeled as others. The method effectively solves the problem of how to find new class marks in the classifier.
This other category requires manual intervention if it occurs frequently in the testing of subsequent new samples, to see if it is a new interference category, and if so, to add a new category label.
Through the multi-classifier voting mode, a large number of unlabeled samples can be labeled. The method effectively improves the marking accuracy and the marking speed, and creates conditions for subsequent identification.
Step 3: according to the labeling result, the interference is subjected to pattern recognition in a mode of combining a time domain and a frequency domain, the specific recognition is that CNN (Convolutional Neural Network ) is adopted for machine learning, and the accuracy of the machine learning is improved by adjusting proper parameters.
For how to form a sample image by adopting a mode of combining a time domain and a frequency domain, it needs to be explained that:
1) The interference itself is not a pattern, but the base station reception signals distributed over 100RB are used as a data input, which is a frequency domain input, and the 100RB can be arranged from low to high. This input in the frequency domain reflects this variation in the frequency domain for each different disturbance that is different.
2) The frequency domain characteristics of each dimension also vary from the time domain, so that the NI variation on each RB in the time domain alone can also be used as an input to one dimension. This time domain variation is also a response to the interference signature.
3) Combining the two characteristics of interference in the frequency domain and the time domain, comprehensively considering that the frequency domain 100RB and the time domain intercept 1 hour granularity (which can be changed specifically as appropriate) are adopted, and the generated spectrum pattern of the time domain is used as the input of the convolutional neural network CNN. The interference is characterized by utilizing the efficient spectrum recognition capability of CNN, as shown in figure 4.
Meanwhile, by utilizing the characteristic of uplink and downlink symmetry, whether downlink interference exists can be judged by detecting uplink interference.
Step 4: and checking the interference model which is successfully trained.
And aiming at a network with successful training, comparing and analyzing the outfield data, checking the model, and recording an incorrect sample. And analyze whether there is a new type.
Step 5: the external field is deployed and the model is recursively analyzed for improvement according to the actual situation.
And carrying out recursive training on the model according to the collected new sample and the new labeling condition, and obtaining a training model again.
Step 6: and (5) redeploying the new training model to the external field, and identifying and judging the external field interference again.
The reported NI data is utilized, the pattern recognition interference detection of machine learning is adopted according to a joint analysis method of a frequency domain and a time domain, and meanwhile, the uplink and downlink symmetry is utilized, so that the interference type can be effectively judged.
Embodiments of the present invention will be described in detail with reference to fig. 5 and 6
Example 1
For example, the LTE field has radio interference in a certain area 1, and the distribution of waveforms of the radio interference on 100 RBs is generally shown in fig. 5. When such interference occurs, it is necessary to manually identify the presence of such interference to take effective measures.
The uplink NI data collected by the system is extracted by a background tool or a network management tool, and uplink NI data with 15 minutes granularity is taken according to the characteristics of a time-frequency domain (the specific granularity can be set to be 5 minutes, 10 minutes and the like, for example). Through necessary data cleaning processes, such as removing null values, removing abnormal values and the like; then obtaining sample data to be processed; since the data of the extracted samples of various disturbances (including no disturbances) are unlabeled, the collected samples need to be labeled according to a set of process flows.
The method comprises the following steps:
301. and (3) averaging a large number of unlabeled samples at a cell level by adopting a cell merging mode, and marking the samples by a manual observation mode after averaging.
302. The samples of radio and television interference marked according to the cell level are classified into a type, and are marked as A, and then the samples marked by the type A are actually a set consisting of a large number of samples, and part of the samples in the set theoretically should be the samples attributed to the radio and television interference, but part of the samples still belong to no interference or other interference.
303. The samples in the A-type space are ordered according to the energy level and are divided into positive samples and negative samples according to the threshold, a specific threshold method can be adopted, for example, the samples below-113 db are negative samples, the samples above-70 db are positive samples, and the specific threshold can be adjusted according to actual conditions.
304. The space A is randomly divided into a plurality of subspaces, which are denoted as a1, a2 and aN; for example, it may be divided into 5 subspaces. N takes 5, and needs to ensure that there should be some positive and negative marked samples in each space
305. The positive and negative samples in each subspace are trained into classifiers, and the classifier algorithm may employ a random forest or some other classifier (as well as other methods set forth in the examples below).
306. Voting and scoring are carried out on the whole space sample by using the plurality of classifiers, the total classification is n, n-2/3 n is positive sample marks, 2/3 n-1/3 n is other category marks, and 1/3 n-0 is no mark category.
A large number of unlabeled exemplars are labeled by the above steps.
Next, time-frequency domain mapping is required for the labeled data.
The operation is as follows
307. The obtained data are ordered according to time dimension, samples can be segmented again through the time dimension ordering, and samples of 1 hour (specific parameters can be set) can be combined into one sample according to the ordered samples assuming 15 minutes granularity of the collected samples.
308. The newly obtained samples contain information in two dimensions, one being the time dimension and one being the frequency dimension, the frequency dimension being 1-100 RB, the time dimension being 1 hour, where the time span is settable, alternatively 2 hours or more.
309. During this merging process, some samples cannot be merged, specifically, if merging is performed in one hour, then there are 4 samples with granularity of 15 minutes to be merged, and if merging is performed in 8 samples with granularity of 2 hours, but if the total number of samples is not 4 or a multiple of 8, then there is a phenomenon that some more samples cannot be merged. When these unwanted samples occur, the embodiments herein use direct discard, or other methods (as will be described in the examples below).
And training the data subjected to time-frequency domain mapping through a CNN convolutional neural network.
The CNN convolutional neural network trained here uses 3 convolutional sampling layers, two fully connected layers, and the last fully connected outputs class probabilities with softmax.
The model is generated by using training data by a gradient descent method to the network, an interference library is formed, and model errors are continuously corrected by external field iteration, so that the recognition accuracy is continuously improved.
Example 2
For example, the LTE external field causes interference of the atmospheric waveguide in a certain area 2, and the interference waveform of the atmospheric waveguide is generally the same, as shown in fig. 6.
When such interference occurs, it is necessary to manually identify the presence of such interference to take effective measures.
The uplink NI data collected by the system is extracted by a background tool or a network management tool, and uplink NI data with 15 minutes granularity is taken according to the characteristics of a time-frequency domain (the specific granularity can be set to be 5 minutes, 10 minutes and the like, for example). Through necessary data cleaning processes, such as removing null values, removing abnormal values and the like; then obtaining sample data to be processed; since the data of the extracted samples of various disturbances (including no disturbances) are unlabeled, the collected samples need to be labeled according to a set of process flows.
The method comprises the following steps:
401. and (3) averaging a large number of unlabeled samples at a cell level by adopting a cell merging mode, and marking the samples by a manual observation mode after averaging.
402. The samples of radio and television interference marked according to the cell level are classified into a type, and are marked as A, and then the samples marked by the type A are actually a set consisting of a large number of samples, and part of the samples in the set theoretically should be the samples attributed to the radio and television interference, but part of the samples still belong to no interference or other interference.
403. The samples in the A-type space are ordered according to the energy level and are divided into positive samples and negative samples according to the threshold, a relative threshold method can be adopted, for example, less than 20% of the samples are negative samples, more than 80% of the samples are positive samples, and the specific threshold can be adjusted according to actual conditions.
404. The space A is randomly divided into a plurality of subspaces, which are denoted as a1, a2 and aN; for example, it may be divided into 5 subspaces. N is taken to be 5, which is required to ensure that there should be some positive and negative marked samples in each space.
405. In contrast to the embodiment, here, the positive and negative samples in each subspace may be trained into different classifiers by using different kinds of classifier algorithms, for example, a bayesian classifier, a decision tree classifier, a svm classifier, and other classifiers may be used.
406. Voting and scoring are carried out on the whole space sample by using the plurality of classifiers, the total classification is n, n-2/3 n is positive sample marks, 2/3 n-1/3 n is other category marks, and 1/3 n-0 is no mark category.
A large number of unlabeled exemplars are labeled by the above steps.
Next, time-frequency domain mapping is required for the labeled data.
The operation is as follows
407. The obtained data are ordered according to time dimension, samples can be segmented again through the time dimension ordering, and samples of 1 hour (specific parameters can be set) can be combined into one sample according to the ordered samples assuming 15 minutes granularity of the collected samples.
408. The newly obtained samples contain information in two dimensions, one being the time dimension and one being the frequency dimension, the frequency dimension being 1-100 RB, the time dimension being 1 hour, where the time span is settable, alternatively 2 hours or more.
409. During this merging process, some samples cannot be merged, specifically, if merging is performed in one hour, then there are 4 samples with granularity of 15 minutes to be merged, and if merging is performed in 8 samples with granularity of 2 hours, but if the total number of samples is not 4 or a multiple of 8, then there is a phenomenon that some more samples cannot be merged. When these extra samples occur, the difference with one embodiment is that other samples can be duplicated here to be combined into a new sample after completion.
And training the data subjected to time-frequency domain mapping through a CNN convolutional neural network.
The CNN convolutional neural network trained here uses 3 convolutional sampling layers, two fully connected layers, and the last fully connected outputs class probabilities with softmax.
The model is generated by using training data by a gradient descent method to the network, an interference library is formed, and model errors are continuously corrected by external field iteration, so that the recognition accuracy is continuously improved.
In summary, the embodiment of the invention has a continuous updating process, and can discover new interference types and learn storage. The cost of interference detection is reduced, the efficiency of interference judgment in the outfield can be improved by the interference detection mode, and the network quality is improved. The embodiment of the invention can be applied not only in LTE but also in FDD and 5G, and has universality
According to the scheme provided by the embodiment of the invention, based on the external field or the on-site sample, the time-frequency domain data of NI is changed into the identifiable spectrum, and the interference model is formed by a training iteration method and a machine learning method, so that the interference type can be automatically analyzed, and the interference characteristic can be actively learned, so that the efficiency can be improved, and the external field can be conveniently examined for interference.
Although the present invention has been described in detail hereinabove, the present invention is not limited thereto and various modifications may be made by those skilled in the art in accordance with the principles of the present invention. Therefore, all modifications made in accordance with the principles of the present invention should be understood as falling within the scope of the present invention.
Claims (18)
1. A method of detecting interference, comprising:
acquiring interference data classified according to interference types;
obtaining a time-frequency domain mapped interference map by performing time-frequency domain mapping processing on the interference data classified according to the interference types;
training the time-frequency domain mapped interference spectrum to obtain a trained interference model;
and classifying the interference type of the interference data to be verified in the external field by using the trained interference model to obtain the interference type of the interference data to be verified in the external field.
2. The method of claim 1, wherein the obtaining the time-frequency domain mapped interference pattern by performing time-frequency domain mapping processing on the interference data classified according to the interference type comprises:
and dividing the interference data classified according to the interference types according to the frequency domain granularity and the time domain granularity to obtain the time-frequency domain mapped interference spectrum.
3. The method of claim 2, wherein the partitioning the interference data classified by the interference type according to the frequency domain granularity and the time domain granularity to obtain the time-frequency domain mapped interference map comprises:
sorting the interference data which are distributed on the frequency domain granularity and are classified according to the interference types according to the time dimension to obtain sorted interference data;
and merging the ordered interference data according to the time domain granularity to obtain the time-frequency domain mapped interference map.
4. A method according to any one of claims 1 to 3, wherein said obtaining a trained interference model by training said time-frequency domain mapped interference spectrum comprises:
and inputting the interference spectrum subjected to time-frequency domain mapping into a neural network model for training to obtain the trained interference model.
5. A method according to any one of claims 1 to 3, wherein said obtaining interference data that has been classified by interference type comprises:
and performing interference type marking processing on the collected sample interference data by utilizing a plurality of classifier voting modes to obtain the interference data classified according to the interference types.
6. The method of claim 5, wherein the performing the interference type labeling process on the collected sample interference data by using a plurality of classifier voting modes, and obtaining the interference data classified according to the interference type comprises:
respectively carrying out initial marking on the collected sample interference data of various interference types to obtain sample interference data with initial marking;
marking the sample interference data as positive interference marked sample interference data, other interference marked sample interference data and interference-free marked sample interference data according to the interference energy intensity of the sample interference data with the same initial mark;
and finally marking the sample interference data of the positive interference mark, the sample interference data of other interference marks and the sample interference data without interference marks in a classifier voting mode to obtain the interference data classified according to the interference types.
7. The method of claim 6, wherein final labeling of the positively-labeled sample interference data, the other interference-labeled sample interference data, and the non-interference-labeled sample interference data by means of classifier voting to obtain the interference data classified by interference type comprises:
randomly dividing the sample interference data with the same initial mark to obtain a plurality of interference subgroups containing the sample interference data with the positive interference mark, the sample interference data with other interference marks and the sample interference data without interference marks;
generating a classifier for voting on the sample interference data by training each interference group;
and finally marking the sample interference data of the positive interference mark, the sample interference data of other interference marks and the sample interference data without interference marks by using the classifier to obtain the interference data classified according to the interference types.
8. A method according to any one of claims 1 to 3, further comprising:
and when the interference data to be verified in the external field does not accord with the interference type in the trained interference model, carrying out recursive training on the interference model again according to the interference type of the interference data to be verified in the external field to obtain and store a new interference model.
9. An apparatus for detecting interference, comprising:
the marking module is used for acquiring interference data which are classified according to the interference types;
the time-frequency domain mapping processing module is used for obtaining a time-frequency domain mapped interference map by performing time-frequency domain mapping processing on the interference data classified according to the interference types;
the training module is used for training the time-frequency domain mapped interference spectrum to obtain a trained interference model;
and the interference classification module is used for classifying the interference type of the external field interference data to be verified by utilizing the trained interference model to obtain the interference type of the external field interference data to be verified.
10. The apparatus of claim 9, wherein the time-frequency domain mapping processing module is configured to segment the interference data classified according to the interference type according to a frequency domain granularity and a time domain granularity, to obtain the time-frequency domain mapped interference map.
11. The apparatus of claim 10, wherein the time-frequency domain mapping processing module is configured to sort the interference data distributed on the frequency domain granularity and classified according to interference types according to a time dimension, to obtain sorted interference data; and merging the ordered interference data according to a preset granularity to obtain the time-frequency domain mapped interference map.
12. The apparatus according to any one of claims 9 to 11, wherein the training module is configured to input the time-frequency domain mapped interference pattern into a neural network model for training, to obtain the trained interference model.
13. The apparatus according to any one of claims 9 to 11, wherein the marking module is configured to perform an interference type marking process on the collected sample interference data by using a plurality of classifier voting modes, so as to obtain the interference data classified according to the interference type.
14. The apparatus of claim 13, wherein the marking module comprises:
the initial marking unit is used for respectively carrying out initial marking on the collected sample interference data of various interference types to obtain sample interference data with initial marking, and marking the sample interference data into sample interference data with positive interference marking, sample interference data with other interference marking and sample interference data without interference marking according to the interference energy intensity of the sample interference data with the same initial marking;
and the final marking unit is used for finally marking the sample interference data of the positive interference marks, the sample interference data of other interference marks and the sample interference data without interference marks in a classifier voting mode to obtain the interference data classified according to the interference types.
15. The apparatus of claim 14, wherein the final marking unit comprises:
dividing the sample interference data with the same initial mark into a plurality of interference subgroups containing the sample interference data with the positive interference mark, the sample interference data with other interference marks and the sample interference data without interference marks;
a training subunit for generating a classifier for voting on the sample interference data by training each of the interference subgroups;
and the final marking subunit is used for finally marking the sample interference data of the positive interference mark, the sample interference data of other interference marks and the sample interference data without interference marks by using the classifier to obtain the interference data classified according to the interference type.
16. The apparatus according to any one of claims 10 to 11, wherein the training module is further configured to, when the outfield to-be-verified interference data does not conform to the interference type in the trained interference model, re-recursively train the interference model according to the interference type of the outfield to-be-verified interference data, to obtain and store a new interference model.
17. An apparatus for detecting interference, the apparatus comprising: a processor, and a memory coupled to the processor; stored on the memory is a program for detecting disturbances executable on the processor, which when executed by the processor implements the steps of the method for detecting disturbances according to any one of claims 1 to 8.
18. A computer storage medium storing a program for detecting interference, which when executed by a processor carries out the steps of the method for detecting interference according to any one of claims 1 to 8.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311283734.0A CN117216573A (en) | 2018-06-29 | 2018-06-29 | Method and device for detecting interference |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311283734.0A CN117216573A (en) | 2018-06-29 | 2018-06-29 | Method and device for detecting interference |
CN201810693247.4A CN110659656A (en) | 2018-06-29 | 2018-06-29 | Method and device for detecting interference |
Related Parent Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810693247.4A Division CN110659656A (en) | 2018-06-29 | 2018-06-29 | Method and device for detecting interference |
Publications (1)
Publication Number | Publication Date |
---|---|
CN117216573A true CN117216573A (en) | 2023-12-12 |
Family
ID=68986142
Family Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810693247.4A Pending CN110659656A (en) | 2018-06-29 | 2018-06-29 | Method and device for detecting interference |
CN202311283734.0A Pending CN117216573A (en) | 2018-06-29 | 2018-06-29 | Method and device for detecting interference |
Family Applications Before (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810693247.4A Pending CN110659656A (en) | 2018-06-29 | 2018-06-29 | Method and device for detecting interference |
Country Status (2)
Country | Link |
---|---|
CN (2) | CN110659656A (en) |
WO (1) | WO2020001311A1 (en) |
Families Citing this family (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111835443B (en) * | 2020-07-23 | 2023-03-24 | 中国联合网络通信集团有限公司 | Method and device for inhibiting atmospheric waveguide interference |
CN112083393B (en) * | 2020-10-27 | 2023-06-30 | 西安电子科技大学 | Intermittent sampling forwarding interference identification method based on spectrogram average time characteristics |
CN115052309A (en) * | 2021-03-09 | 2022-09-13 | 中兴通讯股份有限公司 | Interference detection method, device, equipment and storage medium |
CN114581252B (en) * | 2022-03-03 | 2024-04-05 | 平安科技(深圳)有限公司 | Target case prediction method and device, electronic equipment and storage medium |
CN115379476B (en) * | 2022-08-10 | 2024-03-19 | 中国联合网络通信集团有限公司 | Method, device, equipment and storage medium for determining cell interference type |
CN115379477B (en) * | 2022-08-10 | 2024-03-15 | 中国联合网络通信集团有限公司 | Method, device, equipment and storage medium for determining cell interference type |
Family Cites Families (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103632168B (en) * | 2013-12-09 | 2017-01-18 | 天津工业大学 | Classifier integration method for machine learning |
DE102015207415A1 (en) * | 2015-04-23 | 2016-10-27 | Adidas Ag | Method and apparatus for associating images in a video of a person's activity with an event |
CN107734507A (en) * | 2016-08-12 | 2018-02-23 | 索尼公司 | Wireless scene identification device and method and Wireless Telecom Equipment and system |
CN106330385B (en) * | 2016-08-29 | 2018-11-09 | 电子科技大学 | A kind of interference type recognition methods |
CN107563303B (en) * | 2017-08-09 | 2020-06-09 | 中国科学院大学 | Robust ship target detection method based on deep learning |
CN108135003A (en) * | 2017-12-25 | 2018-06-08 | 广东海格怡创科技有限公司 | The construction method and system of interference type identification model |
CN108197545A (en) * | 2017-12-25 | 2018-06-22 | 广东海格怡创科技有限公司 | The recognition methods of interference type and system |
-
2018
- 2018-06-29 CN CN201810693247.4A patent/CN110659656A/en active Pending
- 2018-06-29 CN CN202311283734.0A patent/CN117216573A/en active Pending
-
2019
- 2019-06-17 WO PCT/CN2019/091548 patent/WO2020001311A1/en active Application Filing
Also Published As
Publication number | Publication date |
---|---|
CN110659656A (en) | 2020-01-07 |
WO2020001311A1 (en) | 2020-01-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN117216573A (en) | Method and device for detecting interference | |
US9923912B2 (en) | Learning detector of malicious network traffic from weak labels | |
CN106357618B (en) | Web anomaly detection method and device | |
US9715723B2 (en) | Optimization of unknown defect rejection for automatic defect classification | |
US10043264B2 (en) | Integration of automatic and manual defect classification | |
US20230300159A1 (en) | Network traffic anomaly detection method and apparatus, and electronic apparatus and storage medium | |
US20150135318A1 (en) | Method of detecting intrusion based on improved support vector machine | |
CN111148142B (en) | Dormant cell detection method based on anomaly detection and integrated learning | |
WO2021139249A1 (en) | Data anomaly detection method, apparatus and device, and storage medium | |
CN112188532A (en) | Training method of network anomaly detection model, network detection method and device | |
US20230385699A1 (en) | Data boundary deriving system and method | |
CN115913691A (en) | Network flow abnormity detection method and system | |
CN111401420A (en) | Abnormal data clustering method and device for wafer test, electronic equipment and medium | |
Moulay et al. | Ttrees: Automated classification of causes of network anomalies with little data | |
CN117172093A (en) | Method and device for optimizing strategy of Linux system kernel configuration based on machine learning | |
CN114970694B (en) | Network security situation assessment method and model training method thereof | |
US20220083810A1 (en) | Convex optimized stochastic vector sampling based representation of ground truth | |
CN113468555A (en) | Method, system and device for identifying client access behavior | |
CN117711608B (en) | Comprehensive analysis system for integrating nerve injury data | |
Dong et al. | A region selection model to identify unknown unknowns in image datasets | |
CN109271855B (en) | Method for extracting characteristics of industrial control signals | |
CN115511106B (en) | Method, device and readable storage medium for generating training data based on time sequence data | |
CN111651433B (en) | Sample data cleaning method and system | |
CN116193470A (en) | LTE network cell problem positioning and optimizing method | |
CN117235516A (en) | Data processing method, device, computer equipment and storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |