CN107729916A - A kind of interference source classification and identification algorithm and device based on ISODATA - Google Patents
A kind of interference source classification and identification algorithm and device based on ISODATA Download PDFInfo
- Publication number
- CN107729916A CN107729916A CN201710812785.6A CN201710812785A CN107729916A CN 107729916 A CN107729916 A CN 107729916A CN 201710812785 A CN201710812785 A CN 201710812785A CN 107729916 A CN107729916 A CN 107729916A
- Authority
- CN
- China
- Prior art keywords
- monitoring
- cluster
- isodata
- finding
- interference source
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S3/00—Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received
- G01S3/02—Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received using radio waves
- G01S3/14—Systems for determining direction or deviation from predetermined direction
- G01S3/46—Systems for determining direction or deviation from predetermined direction using antennas spaced apart and measuring phase or time difference between signals therefrom, i.e. path-difference systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Mobile Radio Communication Systems (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The present invention relates to a kind of interference source classification and identification algorithm and device based on ISODATA, belong to electronic information field, Classification and Identification is carried out for interference source unknown in space：The PDW of each signal in the region is obtained using interference source monitoring location equipment, each category feature value of interference source monitoring location equipment output is subjected to one-dimensional ISODATA resolvings, from characteristic value data storehouse primitive character value in a period of time is taken out by queue to be retrieved, cluster analysis is carried out to it using multidimensional ISODATA, clusters number and feature clustering central value are exported to user in real time, in this, as the validity feature of most probable interference signal；The Unsupervised clustering carried out quasi real time to interference source monitoring location equipment raw monitored result is analyzed, and so as to which the characteristic parameter of key signal in its monitoring result is clearly presented into user, flexibility is stronger, more rationally.
Description
Technical Field
The invention relates to an ISODATA-based interference source classification and identification algorithm, and belongs to the field of electronic information.
Background
With the rapid development and widespread use of information technology, electromagnetic space has been brought into the field of national interest. The complex electromagnetic environment is an important characteristic in the information era, the radiation power of electronic equipment such as radar detection, communication and identification, navigation and the like in an electronic information system is increased, the frequency spectrum is widened, the number of the equipment is multiplied, the working frequency is seriously overlapped, the surrounding electromagnetic environment is complex day by day, and the electromagnetic environment in a limited space is very severe. In view of this, the industry market has urgent needs for performing electromagnetic monitoring on complex electromagnetic environments in a local space, acquiring frequency spectrums and signal characteristics in a monitoring area, and identifying, classifying, managing, and positioning and orienting interference sources in the monitoring area.
In the existing mature products, the interference source monitoring and positioning equipment adopting a relevant interferometer direction-finding system occupies a mainstream position relatively, and has the advantages that the direction-finding sensitivity, the direction-finding accuracy and the minimum measurement time have certain advantages compared with other methods, the cost is relatively controllable, and mature application cases are more. However, in practical application, various interferences exist in a spatial environment, the interference forms are complex, and the multipath effect of signals in an urban environment cannot be avoided, so that many inconveniences, such as more original observation data, mixing of real signals and multipath reflected signals, difficulty in selecting initial parameters, and the like, exist in using the type of equipment to monitor an interference source, and the development of interference source monitoring and positioning related applications is restricted. Therefore, an adaptive interference source classification and identification algorithm can greatly improve the user experience of the interference source monitoring and positioning equipment.
At present, a classification recognition algorithm special for interference sources is rarely disclosed in published publications, but with reference to the related research in the radar field, the related research in the radar field and the patent in the prior art is mainly developed around the processing of unknown radar signals in an electronic countermeasure scene, and the processing algorithm has high requirements on real-time performance, hardware and parallelization. The patent can be analogized with the following technical means:
the first category is radar signal sorting technology. Currently, most of the conventional radar signal sorting methods perform pre-sorting based on relevant processing of radar pulse arrival angle (AOA), pulse Width (PW), carrier frequency (RF), and the like, and then perform main sorting based on de-interlacing processing of Pulse Repetition Interval (PRI). Representative PRI sorting algorithms include dynamic correlation, histogram, PRI transformation, etc.;
the second category is parametric filtering techniques. The algorithm firstly divides a space omega formed by expanding characteristic parameters of an angle of arrival (AOA), a Pulse Width (PW) and a carrier frequency (RF) into n sorting subspacesThen, collecting pulses belonging to the same subspace and projecting PDW (pulse description word) on the subspace, thereby realizing sorting;the generation principle of the method mainly comprises rectangular uniform division and rectangular non-uniform division, and the specific division needs prior knowledge of a parameter distribution probability density function.
Disclosure of Invention
The invention aims to provide an ISODATA-based interference source classification and identification algorithm for overcoming the defects of the prior art.
The principle of the present invention is an Iterative Self-Organizing Data Analysis Algorithm, ISODATA is an abbreviation of Iterative Self-Organizing Data Analysis Technique, meaning an Iterative Self-Organizing Data Analysis Technique. The algorithm is similar to the K-means algorithm, and also determines the clustering center by means of mean iteration, but adds a man-machine conversation link, can adjust parameters, and introduces a mechanism of merging and splitting, namely when the distance between centers of two classes is smaller than a certain threshold, the centers of the two classes are merged into one class, when the standard deviation of a sample of the class is larger than a certain threshold or the number of samples exceeds a certain threshold, the sample is divided into two classes, and when the number of the classes is smaller than a certain threshold, the splitting is also carried out. In addition, when the number of samples of a certain type is less than a certain threshold, it needs to be eliminated again.
The direction finding principle of the direction finding system of the correlation interferometer is as follows: when radio waves from different directions reach the direction-finding antenna array during traveling, the incoming wave direction can be determined by measuring the incoming wave phase and the phase difference because the phase difference between the received radio waves is different in each direction-finding antenna unit spatially. In the direction finding mode of the correlation interferometer, the phase of the induced voltage of the direction finding antenna is directly measured, and then the phase difference is solved.
The technical scheme of the invention comprises that aiming at unknown interference sources in space, the unknown interference sources are classified and identified through the following steps:
step 1: using interference source monitoring and positioning equipment to obtain PDW of each signal in the area, setting a decision threshold value (such as-80 dBm) according to a conventional signal, deciding the signal with the power intensity larger than the decision threshold value as a suspected interference signal, monitoring an original characteristic value of the interference signal in real time and storing the original characteristic value in a characteristic value database;
step 2: performing one-dimensional ISODATA solution on each class of characteristic values output by the interference source monitoring and positioning equipment, and bringing the characteristic values with the clustering number larger than 3 into a queue to be retrieved;
and step 3: and (3) taking out the original characteristic value (for example, 5 minutes, the user can define) in any time period from the characteristic value database according to the queue to be retrieved, carrying out cluster analysis on the original characteristic value by adopting multi-dimensional ISODATA, and outputting the cluster number and the characteristic cluster center value to the user in real time to serve as the effective characteristic of the most possible interference signal.
The ISODATA algorithm in the step 2 and the step 3 comprises the following steps:
the first step is as follows: for a set of N pattern samples, C initial cluster centers Z are determined 1 ,Z 2 ,……Z C C is not necessarily equal to K, and these cluster centers may be any sample in the pattern set;
the second step is that: determining the six parameters, i.e. K, theta N ,θ S ,θ C ,L,I，
K is the required clustering center number;
θ N a category should have at least the number of samples;
θ S a class sample standard deviation threshold;
θ C a threshold value of the distance between the clustering centers, namely a merging coefficient;
l is the maximum logarithm of the categories which can be merged in one iteration;
i, allowing the maximum number of iterations;
the third step: sorting the N samples by minimum distance, i.e. if
D j ＝min(||x-z i ||,i＝1,2,…C)→x∈f j ；
The fourth step: if for a certain cluster domain f j Number of samples N j ＜θ N The subset f is cancelled i And Z i And C = C-1, i.e. the number of classes is reduced by 1;
the fifth step: correcting the central value of each cluster:
wherein N is j Is f i The number of samples of (a);
and a sixth step: for each cluster field f j Calculating the average value of the distances from all samples to the clustering center;
the seventh step: calculating the average value of the distances from all the samples to the corresponding clustering centers;
the eighth step:
(1) if the iteration number reaches I times, the value is set to theta c =0, shift to the twelfth step, end the operation,
(2) if C is less than or equal to K/2, namely the number of the clustering centers is equal to or less than one half of the specified number, the method is switched to the ninth step to split the existing classes,
(3) if C is more than or equal to 2K, skipping splitting, turning to the twelfth step, otherwise turning down,
(4) if the K/2-woven fabric (C) -woven fabric (2K) is adopted, when the iteration times are odd to ninth step (split), and the iteration times are even, the twelfth step (combination) is carried out;
the ninth step: calculating the standard deviation vector of the distance between the sample and the cluster center in each class:
σ j ＝(σ 1j ，σ 2j ，…σ nj ) t ；
wherein each component isWhere n is the dimension of the pattern sample, x il Is the i component of the L sample, z ij Is z j The ith component of (a);
the tenth step: for each standard deviation vector sigma j J =1,2, \ 8230c, C, in which the maximum component σ is determined jmax ，
WhereinNote that it is the second component by S;
the eleventh step: at the maximum set of components { σ jmax J =1,2, \ 8230; C } if there is a Kmax ＞θ S And one of the following conditions is satisfied:
that is, the average distance between the samples in the class and the cluster center is greater than the total distance and the total number of the samples in the class exceeds twice of the specified data, then z is determined j Split into two new cluster centersAnd the number of classes C = C +1;
the determining method comprises the following steps: for σ jmax Z of (a) j On the s-th componentAdding r to j ＝Mσ jmax ,0＜M≤1，z j S component minus r j Other components are unchanged, constitute
And (5) after the splitting is finished, turning to the third step. If there are no fracturable categories, continue;
a twelfth step: calculating the distance between every two cluster centers:
the thirteenth step: each d is ij Value of (a) and theta c Is compared with the value of d ij <θ c Are all taken out to form a setHere, theAnd t is less than or equal to l;
the fourteenth step is that: fromAt the beginning, merge d one by one ij <θ c Each category i, j can be merged only once, and after merging, the new category takes a new clustering center as a mark:
the fifteenth step: if the last iteration calculation (i.e. I time), the algorithm is ended; otherwise, if the parameters need to be changed, the second step is carried out, and if the parameters do not need to be changed, the third step is carried out.
An interference source classification and identification device based on ISODATA mainly comprises a monitoring direction-finding antenna and a monitoring direction-finding host which are connected with each other through a signal line and a control line;
the monitoring direction-finding antenna consists of a monitoring side antenna array, a switch matrix and an electronic compass, and is supported by the antenna bracket;
the monitoring and direction finding host consists of a radio frequency module, an intermediate frequency module and a power supply module,
the monitoring direction-finding host machine is powered by the power supply unit;
the monitoring direction-finding host is also connected with Beidou communication positioning equipment through a comprehensive cable and is connected with a network switch through network communication;
the Beidou communication positioning equipment is communicated and reported with a satellite through a Beidou link;
the network switch is also connected with the display control terminal and is used for network reporting and network communication with the satellite through the network.
The invention comprises the following steps: monitoring electromagnetic wave signals in a receiving space of the direction-finding antenna, finishing switching of output signals of the antenna by the switch matrix, providing directions by the electronic compass, receiving the signals of the direction-finding antenna by the monitoring direction-finding receiver, and finishing the functions of searching, intercepting, monitoring and direction finding of the signals; the output original characteristic values comprise interference center frequency points, bandwidth, power, standard, direction and duration.
The invention has the following advantages that the specific application scene of the interference source monitoring and positioning equipment concerned by the invention generally has no urgent need on the characteristics, particularly in the civil field, users pay more attention to the simplicity and the universality of the algorithm, and cost factors are also important consideration indexes, so the invention provides the interference source classification and identification algorithm based on ISODATA (iterative self-organizing data analysis) algorithm, and aims to realize the quasi-real-time unsupervised cluster analysis on the original monitoring result of the interference source monitoring and positioning equipment, thereby clearly presenting the characteristic parameters of key signals in the monitoring result to the users, and having stronger flexibility and more reasonable.
Drawings
Fig. 1 is a schematic block diagram of an interference source monitoring and positioning device.
Fig. 2 is a flow chart of the algorithm.
Fig. 3 is a graph of raw monitoring data.
FIG. 4 is a final experimental result chart.
Fig. 5 is a detailed numerical diagram.
Detailed Description
In the following, referring to fig. 1 to 2, a preferred embodiment of the present invention is further described, wherein the interference source monitoring and positioning device mainly includes a monitoring direction-finding antenna 1 and a monitoring direction-finding host 2, which are connected to each other through a signal line and a control line;
the monitoring direction-finding antenna 1 is composed of a monitoring side antenna array 11, a switch matrix 12 and an electronic compass 13, and the monitoring direction-finding antenna 1 is supported by an antenna support 14;
the monitoring direction-finding host 2 consists of a radio frequency module 21, an intermediate frequency module 22 and a power supply module 23,
the monitoring direction-finding host 2 is powered by a power supply unit 24;
the monitoring direction-finding host 2 is also connected with the Beidou communication positioning equipment 3 through a comprehensive cable and is connected with the network switch 4 through network communication;
the Beidou communication positioning equipment 3 communicates and reports with a satellite through a Beidou link;
the network switch 4 is also connected with the display control terminal 5, and performs network report and network communication with the satellite through a network.
The interference source monitoring and positioning equipment comprises the following methods: the monitoring direction-finding antenna 1 receives electromagnetic wave signals in space, the switch matrix 12 completes the switching of output signals of the antenna, the electronic compass 13 provides directions, and the monitoring direction-finding host 2 receives the signals of the monitoring direction-finding antenna 1 and completes the functions of searching, intercepting, monitoring and direction finding of the signals; the output original characteristic values comprise interference center frequency points, bandwidth, power, standard, direction and duration.
For unknown interference sources in space, classifying and identifying the unknown interference sources through the following steps:
step 1: using interference source monitoring and positioning equipment to obtain PDW of each signal in the area, setting a judgment threshold value (such as-80 dBm) according to a conventional signal, judging the signal with the power intensity larger than the judgment threshold value as a suspected interference signal, monitoring an original characteristic value of the interference signal in real time and storing the original characteristic value into a characteristic value database;
and 2, step: performing one-dimensional ISODATA resolving on each type of characteristic values output by the interference source monitoring and positioning equipment, and bringing the characteristic values of which the clustering number is more than 3 into a queue to be retrieved;
and 3, step 3: and (3) taking out the original characteristic value (for example, 5 minutes, which can be defined by a user) in any time period from the characteristic value database according to the queue to be retrieved, carrying out cluster analysis on the original characteristic value by adopting multi-dimensional ISODATA, and outputting the cluster number and the characteristic cluster center value to the user in real time to serve as the effective characteristic of the most possible interference signal.
The ISODATA algorithm in the step 2 and the step 3 comprises the following steps:
the first step is as follows: for a set of N pattern samples, C initial cluster centers Z are determined 1 ,Z 2 ,……Z C C is not necessarily equal to K, and these cluster centers may be any sample in the pattern set;
the second step is that: determining the six parameters, i.e., K, θ N ,θ S ,θ C ,L,I，
K is the required clustering center number;
θ N a category should have at least the number of samples;
θ S a class sample standard deviation threshold;
θ C a threshold value of the distance between the clustering centers, namely a merging coefficient;
l is the maximum logarithm of the categories which can be merged in one iteration;
i, allowing the maximum number of iterations;
the third step: sorting the N samples by minimum distance, i.e. if
D j ＝min(||x-z i ||,i＝1,2,…C)→x∈f j ；
The fourth step: if for a certain cluster domain f j Number of samples N j <θ N The subset f is cancelled i And Z i And C = C-1, i.e. the number of classes is reduced by 1;
the fifth step: correcting the center value of each cluster:
wherein N is j Is f i The number of samples of (a);
and a sixth step: for each cluster field f j Calculating the average value of the distances from all samples to the clustering center;
the seventh step: calculating the average value of the distances from all the samples to the corresponding clustering centers;
eighth step:
(1) if the iteration number reaches I times, put theta c =0, shift to the twelfth step, end the operation,
(2) if C is less than or equal to K/2, namely the number of the clustering centers is equal to or less than one half of the number of the specified number, the method shifts to the ninth step to split the existing classes,
(3) if C is more than or equal to 2K, skipping splitting, turning to the twelfth step, otherwise turning down,
(4) if the K/2-woven fabric (C) -woven fabric (2K) is adopted, when the iteration times are odd to ninth step (split), and the iteration times are even, the twelfth step (combination) is carried out;
the ninth step: calculating the standard deviation vector of the distance between the sample and the cluster center in each class:
σ j ＝(σ 1j ，σ 2j ，…σ nj ) t ；
wherein each component isWhere n is the dimension of the pattern sample, x il Is the i component of the L sample, z ij Is z j The ith component of (2);
the tenth step: for each standard deviation vector sigma j J =1,2, \8230C, where the maximum component σ is found jmax ，
WhereinNote that it is the second component by S;
the eleventh step: at the maximum set of components { σ jmax J =1,2, \ 8230; C } if there is a Kmax ＞θ S And one of the following conditions is satisfied:
i.e. the average distance from the sample to the cluster center in the class is greater than the total distance and the total number of samples in the class exceeds twice the specified data, then z is j Split into two new cluster centersAnd the number of classes C = C +1;
the determination method of (1): for σ jmax Z of (a) j Is added r to the s-th component j ＝Mσ jmax ,0＜M≤1，z j S component minus r j Other components are unchanged, constitute
And after the splitting is finished, the third step is carried out. If there are no fracturable categories, continue;
a twelfth step: calculating the distance between every two cluster centers:
the thirteenth step: each d is ij Value of (a) and theta c Is compared to the value of (d) ij <θ c Are all taken out to form a setHere, theAnd t is less than or equal to l;
the fourteenth step is that: fromAt the beginning, merge d one by one ij <θ c Each category i, j of (1) can be merged only once, and after merging, the new category takes a new clustering center as a mark:
the fifteenth step: if the last iteration calculation (i.e. I time), the algorithm is ended; otherwise, if the parameters need to be changed, the second step is carried out, and if the parameters do not need to be changed, the third step is carried out.
In the experiment, 1250.1MHz is monitored by adopting ZSTX-F21PP type interference source monitoring and positioning equipment of Sen communication Limited in lake south, two groups of characteristic values of interference power and interference direction are selected to be brought into a queue to be retrieved according to the actual situation, the original monitoring data generated in one minute is shown in figure 3, the actual data acquisition of the experiment lasts for 4 minutes, 132 groups of data are generated, and for the original observation data, an ISODATA program written under Matlab is adopted for calculation, and the range of the initial parameters is as follows:
k =2 to 10 (required number of cluster centers);
(the number of samples a class should have at least);
a class sample standard deviation threshold;
a threshold value of the distance between the clustering centers, namely a merging coefficient;
l = 2-5 maximum logarithm of classes that can be merged in one iteration;
i = 5-20 allows the maximum number of iterations.
The final experimental result is shown in fig. 4, in which all the feature values are aggregated into three groups of signals (different colors), and the red asterisk points respectively represent the corresponding cluster centers, that is, the features of the interference signals to be finally presented to the user, and the specific values thereof are shown in fig. 5.
Through the processing, the characteristic value data presented to the user are simplified and abstracted from the original 132 groups of disordered original characteristic values into 3 groups of results, manual intervention is not needed in the whole process, the data processing efficiency is greatly improved, and the user experience is optimized.
Claims (4)
1. An ISODATA-based interference source classification and identification algorithm is characterized in that: for an unknown source of interference in the space,
classifying and identifying the same by the following steps:
step 1: using interference source monitoring and positioning equipment to obtain PDW of each signal in the area, setting a judgment threshold value according to a conventional signal, judging the signal with the power intensity larger than the judgment threshold value as a suspected interference signal, monitoring an original characteristic value of the interference signal in real time and storing the original characteristic value into a characteristic value database;
and 2, step: performing one-dimensional ISODATA resolving on each type of characteristic values output by the interference source monitoring and positioning equipment, and bringing the characteristic values of which the clustering number is more than 3 into a queue to be retrieved;
and 3, step 3: and taking out the original characteristic value in a set time period from the characteristic value database according to the queue to be retrieved, carrying out cluster analysis on the original characteristic value by adopting multi-dimensional ISODATA, and outputting the cluster number and the characteristic cluster center value to a user in real time to serve as the effective characteristic of the most possible interference signal.
2. The ISODATA-based interference source classification and identification algorithm as claimed in claim 1, wherein the steps of ISODATA algorithm in step 2 and step 3 are as follows:
the first step is as follows: for a set of N pattern samples, C initial cluster centers Z are determined 1 ,Z 2 ,……Z C C is not necessarily equal to K, and these cluster centers may be any sample in the pattern set;
the second step: determining the six parameters, i.e., K, θ N ,θ S ,θ C ,L,I，
K is the required clustering center number;
θ N a category should have at least the number of samples;
θ S a class sample standard deviation threshold;
θ C a threshold value of the distance between the clustering centers, namely a merging coefficient;
l is the maximum logarithm of the categories which can be merged in one iteration;
i, allowing the maximum number of iterations;
the third step: sorting the N samples by minimum distance, i.e. if
D j ＝min(||x-z i ||,i＝1,2,…C)→x∈f j ；
The fourth step: if for a certain cluster domain f j Number of samples N thereof j <θ N The subset f is cancelled i And Z i And C = C-1,i.e. the number of classes is reduced by 1;
the fifth step: correcting the central value of each cluster:
wherein N is j Is f i The number of samples of (a);
and a sixth step: for each cluster field f j Calculating the average value of the distances from all samples to the clustering center;
the seventh step: calculating the average value of the distances from all the samples to the corresponding clustering centers;
eighth step:
(1) if the iteration number reaches I times, put theta c =0, shift to the twelfth step, end the operation,
(2) if C is less than or equal to K/2, namely the number of the clustering centers is equal to or less than one half of the number of the specified number, the method shifts to the ninth step to split the existing classes,
(3) if C is more than or equal to 2K, skipping splitting, turning to the twelfth step, otherwise turning down,
(4) if K/2-C-P (C-P) is constructed, when the iteration times are odd to ninth step (split) and even, the twelfth step (combination) is carried out;
the ninth step: calculating the standard deviation vector of the distance between the sample and the cluster center in each class:
σ j ＝(σ 1j ，σ 2j ，…σ nj ) t ；
wherein each component isWhere n is the dimension of the pattern sample, x il Is the ith component of the Lth sample, z ij Is z j The ith component of (a);
the tenth step: for each standard deviation vector sigma j J =1,2, \8230C, where the maximum component σ is found jmax ，
Whereinj =1,2, \8230C; k = dimension, which is the fractional component noted by S;
the eleventh step: at the maximum set of components { σ jmax J =1,2, \ 8230; C } if there is a Kmax ＞θ S And one of the following conditions is satisfied:
that is, the average distance between the samples in the class and the cluster center is greater than the total distance and the total number of the samples in the class exceeds twice of the specified data, then z is determined j Split into two new cluster centersAnd the number of classes C = C +1;
the determining method comprises the following steps: for σ jmax Z of (a) j Is added r to the s-th component j ＝Mσ jmax ,0＜M≤1，z j S component minus r j Other components are unchanged, constitute
After splitting, turning to the third step; if there are no fracturable categories, continue;
a twelfth step: calculating the distance between every two cluster centers:
the thirteenth step: each d is ij Value of (a) and theta c Is compared with the value of d ij <θ c Are all taken out to form a setHere, theAnd t is less than or equal to l;
the fourteenth step is that: fromAt the beginning, merge d one by one ij <θ c Each category i, j can be merged only once, and after merging, the new category takes a new clustering center as a mark:
the fifteenth step: if the iteration is the last iteration calculation (i.e. I), the algorithm is ended; otherwise, if the parameters need to be changed, the second step is carried out, and if the parameters do not need to be changed, the third step is carried out.
3. An interference source classification and identification device based on ISODATA is characterized in that: the device mainly comprises a monitoring direction-finding antenna (1) and a monitoring direction-finding host (2), wherein the monitoring direction-finding antenna and the monitoring direction-finding host are mutually connected through a signal wire and a control wire;
the monitoring direction-finding antenna (1) is composed of a monitoring side antenna array (11), a switch matrix (12) and an electronic compass (13), and the monitoring direction-finding antenna (1) is supported by an antenna support (14); the monitoring direction-finding host (2) is composed of a radio frequency module (21), an intermediate frequency module (22) and a power supply module (23), and the monitoring direction-finding host (2) supplies power through a power supply unit (24);
the monitoring direction-finding host (2) is also connected with the Beidou communication positioning equipment (3) through a comprehensive cable and is connected with the network switch (4) through network communication;
the Beidou communication positioning equipment (3) is communicated and reported with a satellite through a Beidou link;
the network switch (4) is also connected with the display and control terminal (5) and is used for network reporting and network communication with the satellite through a network.
4. The apparatus of claim 3, comprising: the monitoring direction-finding antenna (1) receives electromagnetic wave signals in a space, the switch matrix (12) completes switching of output signals of the antenna, the electronic compass (13) provides directions, and the monitoring direction-finding host (2) receives the signals of the monitoring direction-finding antenna (1) and completes searching, intercepting, monitoring and direction-finding functions of the signals; the output original characteristic values comprise interference center frequency points, bandwidth, power, standard, direction and duration.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710812785.6A CN107729916B (en) | 2017-09-11 | 2017-09-11 | ISODATA-based interference source classification and identification algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710812785.6A CN107729916B (en) | 2017-09-11 | 2017-09-11 | ISODATA-based interference source classification and identification algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107729916A true CN107729916A (en) | 2018-02-23 |
| CN107729916B CN107729916B (en) | 2021-11-19 |
Family
ID=61205981
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710812785.6A Active CN107729916B (en) | 2017-09-11 | 2017-09-11 | ISODATA-based interference source classification and identification algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107729916B (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108964830A (en) * | 2018-05-18 | 2018-12-07 | 中通服建设有限公司 | A kind of unmanned plane interference positioning system |
| CN109557562A (en) * | 2018-12-03 | 2019-04-02 | 中国电波传播研究所（中国电子科技集团公司第二十二研究所） | A kind of GNSS interference source localization method based on radio waves propagation model |
| CN109557560A (en) * | 2018-12-17 | 2019-04-02 | 北斗航天卫星应用科技集团有限公司 | A kind of Beidou Navigation System fault detection method and detection system based on ISODATA clustering algorithm |
| CN109765555A (en) * | 2018-12-28 | 2019-05-17 | 通鼎互联信息股份有限公司 | The method and device of WSN node locating is realized by constructing virtual SAR image |
| CN110390028A (en) * | 2019-04-16 | 2019-10-29 | 杭州电子科技大学 | A kind of method for building up in plant spectral library |
| CN111836189A (en) * | 2020-06-28 | 2020-10-27 | 四川省大见通信技术有限公司 | Device and method for positioning interference source |
| CN112231392A (en) * | 2020-10-29 | 2021-01-15 | 广东机场白云信息科技有限公司 | Civil aviation customer source data analysis method, electronic equipment and computer readable storage medium |
| CN112543411A (en) * | 2019-09-20 | 2021-03-23 | 中兴通讯股份有限公司 | Interference positioning method, device and system of wireless communication system |
| WO2023045926A1 (en) * | 2021-09-23 | 2023-03-30 | 中兴通讯股份有限公司 | Interference signal avoidance method and apparatus, and base station and storage medium |
| CN116628481A (en) * | 2023-07-21 | 2023-08-22 | 江西红声技术有限公司 | Electronic countermeasure information source identification method, system, computer and readable storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH06215196A (en) * | 1992-11-27 | 1994-08-05 | Ind Technol Res Inst | Method for establishment of group database in character recognition |
| CN202587395U (en) * | 2012-04-16 | 2012-12-05 | 成都光大灵曦科技发展有限公司 | GSM-R vehicle-mounted wireless interference testing system |
| CN102945238A (en) * | 2012-09-05 | 2013-02-27 | 南京航空航天大学 | Fuzzy ISODATA (interactive self-organizing data) based feature selection method |
| CN103139816A (en) * | 2011-12-01 | 2013-06-05 | 鼎桥通信技术有限公司 | Positioning method and positioning device of interference sources |
| CN105182282A (en) * | 2015-10-23 | 2015-12-23 | 成都九华圆通科技发展有限公司 | Monitoring direction finding system based on aircraft-mounted lift-off interferometer |
-
2017
- 2017-09-11 CN CN201710812785.6A patent/CN107729916B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH06215196A (en) * | 1992-11-27 | 1994-08-05 | Ind Technol Res Inst | Method for establishment of group database in character recognition |
| CN103139816A (en) * | 2011-12-01 | 2013-06-05 | 鼎桥通信技术有限公司 | Positioning method and positioning device of interference sources |
| CN202587395U (en) * | 2012-04-16 | 2012-12-05 | 成都光大灵曦科技发展有限公司 | GSM-R vehicle-mounted wireless interference testing system |
| CN102945238A (en) * | 2012-09-05 | 2013-02-27 | 南京航空航天大学 | Fuzzy ISODATA (interactive self-organizing data) based feature selection method |
| CN105182282A (en) * | 2015-10-23 | 2015-12-23 | 成都九华圆通科技发展有限公司 | Monitoring direction finding system based on aircraft-mounted lift-off interferometer |
Non-Patent Citations (4)
| Title |
|---|
| MINGCHAOLI 等: "An enhanced ISODATA algorithm for recognizing multiple electric appliances from the aggregated power consumption dataset", 《ENERGY AND BUILDINGS》 * |
| 周文辉: "未知雷达辐射源信号分选方法研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 张洪亮 等: "基于ISODATA聚类的复杂体制雷达信号分选", 《电子信息对抗技术》 * |
| 王文松 等: "基于抗ISODATA聚类分析的辐射源反侦察研究", 《电子信息对抗技术》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108964830B (en) * | 2018-05-18 | 2020-11-24 | 中通服建设有限公司 | Unmanned aerial vehicle interference positioning system |
| CN108964830A (en) * | 2018-05-18 | 2018-12-07 | 中通服建设有限公司 | A kind of unmanned plane interference positioning system |
| CN109557562A (en) * | 2018-12-03 | 2019-04-02 | 中国电波传播研究所（中国电子科技集团公司第二十二研究所） | A kind of GNSS interference source localization method based on radio waves propagation model |
| CN109557562B (en) * | 2018-12-03 | 2023-03-31 | 中国电波传播研究所（中国电子科技集团公司第二十二研究所） | GNSS interference source positioning method based on radio wave propagation model |
| CN109557560A (en) * | 2018-12-17 | 2019-04-02 | 北斗航天卫星应用科技集团有限公司 | A kind of Beidou Navigation System fault detection method and detection system based on ISODATA clustering algorithm |
| CN109765555A (en) * | 2018-12-28 | 2019-05-17 | 通鼎互联信息股份有限公司 | The method and device of WSN node locating is realized by constructing virtual SAR image |
| CN110390028B (en) * | 2019-04-16 | 2021-08-10 | 杭州电子科技大学 | Method for establishing plant spectrum library |
| CN110390028A (en) * | 2019-04-16 | 2019-10-29 | 杭州电子科技大学 | A kind of method for building up in plant spectral library |
| CN112543411A (en) * | 2019-09-20 | 2021-03-23 | 中兴通讯股份有限公司 | Interference positioning method, device and system of wireless communication system |
| WO2021052336A1 (en) * | 2019-09-20 | 2021-03-25 | 中兴通讯股份有限公司 | Interference positioning method, apparatus and system for wireless communication system |
| CN112543411B (en) * | 2019-09-20 | 2023-07-11 | 中兴通讯股份有限公司 | Interference positioning method, device and system of wireless communication system |
| CN111836189A (en) * | 2020-06-28 | 2020-10-27 | 四川省大见通信技术有限公司 | Device and method for positioning interference source |
| CN112231392A (en) * | 2020-10-29 | 2021-01-15 | 广东机场白云信息科技有限公司 | Civil aviation customer source data analysis method, electronic equipment and computer readable storage medium |
| WO2023045926A1 (en) * | 2021-09-23 | 2023-03-30 | 中兴通讯股份有限公司 | Interference signal avoidance method and apparatus, and base station and storage medium |
| CN116628481A (en) * | 2023-07-21 | 2023-08-22 | 江西红声技术有限公司 | Electronic countermeasure information source identification method, system, computer and readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107729916B (en) | 2021-11-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107729916B (en) | ISODATA-based interference source classification and identification algorithm | |
| CN108520023B (en) | Thunderstorm kernel identification and tracking method based on hybrid clustering algorithm | |
| CN111585671B (en) | Electric power LTE wireless private network electromagnetic interference monitoring and identifying method | |
| Ghadimi et al. | Deep learning-based approach for low probability of intercept radar signal detection and classification | |
| CN110427878A (en) | A kind of sudden and violent signal recognition method of Rapid Radio and system | |
| CN111339986B (en) | Device frequency law mining method and system based on time domain/frequency domain analysis | |
| CN117031509A (en) | Soil humidity inversion method and device integrating isolated forest and deep learning | |
| Khudhair et al. | Estimation of direction of arrival for antenna array based on esprit and multiple signal classification algorithms | |
| CN103076595B (en) | Abnormal type identifying method for multivariate discrete radar emitter signal | |
| US10151823B2 (en) | Method for passive approximate localization using frequency modulation and software defined radio | |
| Man et al. | On the technology of frequency hopping communication network-station selection | |
| CN116033345A (en) | Indoor abnormal signal high-precision positioning method, system and device | |
| CN111901026B (en) | Arrival angle estimation method in communication | |
| CN112130119B (en) | Distributed multi-station electronic reconnaissance cooperative signal processing system and method | |
| US20190339357A1 (en) | Built-in sensor for intercepting radioelectric com/rad emissions | |
| CN114401055A (en) | Intelligent frequency spectrum detection system | |
| Aydin et al. | Faster Wi-Fi fingerprinting using feature selection | |
| CN117851837B (en) | Radiation source target identification method and system based on fine histogram features | |
| Žák et al. | A histogram based radar signal identification process | |
| CN115078845B (en) | Method and device for acquiring pulse repetition period and computer storage medium | |
| Chen et al. | Data Driven GNSS Based Interference Signal Detection Technology | |
| CN113037408B (en) | Signal sensing method and device combining space arrival angle and frequency spectrum two-dimensional | |
| Guo et al. | An SEI-Based Identification Scheme for Illegal FM Broadcast | |
| Mottier et al. | Deinterleaving RADAR emitters with optimal transport distances | |
| Ying et al. | Channel Attention Mechanism-based Multi-Feature Fusion Network for Specific Emitter Identification |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20211222 Address after: 610000 c129 and C130, floor 1, Zone C, No. 4, Xixin Avenue, high tech Zone, Chengdu, Sichuan Patentee after: Chengdu Zhongsen Communication Technology Co.,Ltd. Address before: 410006 11th floor, headquarters building, Changsha Zhongdian Software Park, 39 Jianshan Road, Yuelu Avenue, Yuelu District, Changsha City, Hunan Province Patentee before: HUNAN ZHONGSEN COMMUNICATION TECHNOLOGY Co.,Ltd. |