CN107770108A - A kind of combined modulation recognition methods of K mean clusters and classification training SVM classifier - Google Patents
A kind of combined modulation recognition methods of K mean clusters and classification training SVM classifier Download PDFInfo
- Publication number
- CN107770108A CN107770108A CN201710993729.7A CN201710993729A CN107770108A CN 107770108 A CN107770108 A CN 107770108A CN 201710993729 A CN201710993729 A CN 201710993729A CN 107770108 A CN107770108 A CN 107770108A
- Authority
- CN
- China
- Prior art keywords
- svm classifier
- recognition methods
- classification
- mean clusters
- matrix
- 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
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L27/00—Modulated-carrier systems
- H04L27/0012—Modulated-carrier systems arrangements for identifying the type of modulation
-
- 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
- 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
- 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
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Signal Processing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Probability & Statistics with Applications (AREA)
- Digital Transmission Methods That Use Modulated Carrier Waves (AREA)
Abstract
The invention discloses the combined modulation recognition methods of a kind of K mean clusters and classification training SVM classifier, it is characterised in that comprises the following steps：Step S1：Unknown signaling is pre-processed, obtains the data set X of the in-phase component comprising signal and quadrature component；Step S2：Cluster computing is carried out to the sample point of the data set X, obtains subordinated-degree matrix of each sample point to cluster centre；Step S3：The subordinated-degree matrix is handled with Validity Function, obtains some characteristic parameters for distinguishing different modulating mode；Step S4：Using the combination of some characteristic parameters as the input of SVM classifier, the SVM classifier is trained；Step S5：The unknown signaling is identified using the SVM classifier after training, the modulation type of the unknown signaling is obtained in the output of SVM classifier.The problem of computation complexity is high when identifying multi-class problem is the method overcome, convergence rate significantly improves, and improves the efficiency of Modulation Signals Recognition.
Description
Technical field
The invention belongs to signal modulate technical field, and in particular to SVM points of a kind of K- mean clusters and classification training
The combined modulation recognition methods of class device.
Background technology
It is military with civilian wireless communication field that Modulation Identification is widely used in signal monitoring, inquiry, disturbance ecology etc., is to recognize
Know the basis of the area researches such as radio, frequency spectrum perception.Therefore it is how continuous in terms of the extraction of characteristic value and classifier algorithm
Innovation, turn into one of important method of research Modulation Identification.However, existing identifying multi-class problem based on SVMs
When, its computation complexity is high, and when snr of received signal is relatively low, signal modulate rate is low, therefore, how further to carry
The efficiency of high RST Modulation Identification turns into the problem of urgent need to resolve.
The content of the invention
In order to solve the above problems, it is an object of the invention to provide a kind of K- mean clusters and classification to train SVM classifier
Combined modulation recognition methods, this method have higher recognition efficiency.
To achieve the above object, the present invention is achieved by following technical scheme：
The combined modulation recognition methods of a kind of K- mean clusters of the present invention and classification training SVM classifier, including
Following steps：
Step S1：Unknown signaling is pre-processed, obtains the data set of the in-phase component comprising signal and quadrature component
X；
Step S2：Cluster computing is carried out to the sample point of the data set X, obtains each sample point to cluster centre
Subordinated-degree matrix；
Step S3：The subordinated-degree matrix is handled with Validity Function, obtains being used to distinguish different modulating mode
Some characteristic parameters；
Step S4：Using the combination of some characteristic parameters as the input of SVM classifier, the SVM classifier is entered
Row training；
Step S5：The unknown signaling is identified using the SVM classifier after training, in exporting for SVM classifier
To the modulation type of the unknown signaling.
Further, in step sl, unknown signaling is pre-processed specially：Down coversion, low pass are carried out to carrier wave
Filtering and sample process.
Further, the data set X={ x1,x2,…xN, wherein, N is the number of element in data set, i.e., unknown letter
Number sample point symbolic number.
Further, in step s 2, clustering algorithm is specifically calculated using K- mean clusters, and its classification matrix is V(0), by its
Table
Mathematical programming problem is shown as, object function is：
Constraints is：K is cluster centre number, 1 ＜ K ＜ N.
Further, in step s 2, modulated signal corresponding to different cluster centre numbers is iterated, detailed process
Including：
Step S21：Provide iteration criterion epsilon>0, and initialize classification matrix V(0), n=0；
Step S22：Calculate renewal Iterative Matrix U(n),
Step S23：Calculate cluster centre matrix V(n+1),
Step S24：With matrix norm | | | | compare V(n+1)With V(n)If | | V (n+1)-V(n) | |≤ε is set up, then is stopped
Iteration, otherwise, make K=K+1, repeat step step S22 to step S24；In iterative process, optimization object functionObtain subordinated-degree matrix.
Further, step S3 detailed process is：Have to the cluster result obtained during different cluster centre number K values
The analysis of effect property, whether reasonable, Validity Function value is obtained, as distinguishing difference if judging received signal points being divided into K classes
Some characteristic parameters of modulation system.
Further, in step s 4, SVM classifier is trained using the algorithm of classification.
Further, unknown modulated signal is：2PSK, 4PSK, 8PSK, 16QAM, 32QAM and 64QAM；Corresponding modulation
Exponent number is respectively：1,2,3,4,5 and 6；Take cluster centre number K={ 2,4,8,16,32,64 }
Compared with prior art, the beneficial effects of the invention are as follows：
The present invention optimizes processing to the characteristic value of unknown signaling extraction due to using clustering algorithm, while with K averages
The combined modulation recognizer of cluster and classification training SVM, overcomes the problem of computation complexity is high when identifying multi-class problem,
Convergence rate significantly improves, and improves the efficiency of Modulation Signals Recognition.
Brief description of the drawings
The embodiment of the present invention is described in further detail below in conjunction with the accompanying drawings, wherein：
Fig. 1 is the combined modulation recognition methods of a kind of K- mean clusters of the present invention and classification training SVM classifier
Flow chart；
Fig. 2 is the characteristic ginseng value of six kinds of modulation systems of the present invention under different cluster centre numbers；
Fig. 3 is the classification training schematic diagram of SVM classifier of the present invention；
Fig. 4 is the computational complexity emulation comparison schematic diagram of the present invention；
Fig. 5 is the Modulation Identification rate analogous diagram of the present invention when individually using K- means clustering algorithms；
Fig. 6 be it is of the present invention based on conventional BP algorithm SVM classifier Modulation Identification method when Modulation Identification rate emulation
Figure；
Fig. 7 is that the Modulation Identification rate of the present invention using when K mean cluster and SVM combined modulation recognizers emulates
Figure.
Embodiment
The preferred embodiments of the present invention are illustrated below in conjunction with accompanying drawing, it will be appreciated that described herein preferred real
Apply example to be merely to illustrate and explain the present invention, be not intended to limit the present invention.
The present invention provides the combined modulation recognition methods of a kind of K- mean clusters and classification training SVM classifier, for reality
The Modulation Identification of existing signal, the SVM classifier in text refer to supporting vector grader.The present embodiment is with based on planisphere modulation methods
Exemplified by the MPSK/MQAM modulated signals of formula, the course of work of this method is introduced, as shown in figure 1, comprising the following steps：
Step S1：Unknown signaling is pre-processed, obtains the data set of the in-phase component comprising signal and quadrature component
X.Unknown signaling is pre-processed mainly carrier wave is carried out the Signal Pretreatment such as down coversion, LPF and sampling so as to
The in-phase component of reception signal and the value of quadrature component are obtained, is set to data set X={ x1,x2,…xN, wherein, N is data set
The number of middle element, that is, the unknown signaling sample point symbolic number received, sample point xj=[xjI,xjQ]T, i.e., by reception symbol
In-phase component xjIWith quadrature component xjQThe bivector of composition, the reception planisphere of signal can be formed.
Step S2：Cluster computing is carried out to the sample point of the data set X, obtains each sample point to cluster centre
Subordinated-degree matrix.Unknown signaling is pre-processed after obtaining data set X, and the sample point that can be concentrated to data carries out cluster computing, this
The clustering algorithm of embodiment is specifically calculated using K- mean clusters, and its classification matrix is V(0).K- means clustering algorithms can automatic logarithm
Classified according to object, the degree of membership of cluster centre and each sample point to class center is obtained by optimization object function, so as to
Determine the ownership of sample point.
K- mean cluster problems are represented by mathematical programming problem,
Its object function is：
Constraints is：K is cluster centre number, 1 ＜ K ＜ N.
Modulated signal corresponding to different cluster centre numbers is iterated, detailed process includes：
Step S21：Provide iteration criterion epsilon>0, and initialize classification matrix V(0), n=0；
Step S22：Calculate renewal Iterative Matrix U(n),
Step S23：Calculate cluster centre matrix V(n+1),
Step S24：With matrix norm | | | | compare V(n+1)With V(n)If | | V(n+1)-V(n) | |≤ε is set up, then stops changing
In generation, otherwise, make K=K+1, repeat step step S22 to step S24；In iterative process, optimization object functionObtain subordinated-degree matrix.
Step S3：The subordinated-degree matrix is handled with Validity Function, obtains being used to distinguish different modulating mode
Some characteristic parameters.Due to the signal with different modulating exponent number, its optimal cluster centre number is different, thus to be obtained
To the characteristic parameter that can distinguish different modulating mode, the cluster result obtained in different cluster centre number K values can be entered
Row efficiency analysis, whether reasonable, obtain Validity Function value if judging received signal points being divided into K classes, different as distinguishing
The characteristic parameter of modulation type, so as to distinguish different modulated signals.In the present embodiment, unknown modulated signal is：2PSK,
4PSK, 8PSK, 16QAM, 32QAM and 64QAM；Corresponding order of modulation is respectively：1,2,3,4,5 and 6；Take cluster centre number K
={ 2,4,8,16,32,64 }, the signaling point received is carried out more than in the case of six kinds of different cluster centre numbers respectively
Computing is clustered, and calculates the Validity Function value T in different K values respectivelyk, as the feature for distinguishing different modulating type
Parameter.In different signal to noise ratio, different cluster centre numbers, characteristic parameter used by six kinds of modulation systems of differentiation is calculated respectively
T2、T4、T8、T16、T32。
Step S3 result is as shown in Fig. 2 under different signal to noise ratio, modulated signal 2PSK characteristic parameter T2Value it is substantially high
In other five kinds of modulation systems, therefore, pass through characteristic parameter T2Modulation system 2PSK can be distinguished with other modulation types
Come；Similarly under different signal to noise ratio, 4PSK characteristic parameter T4It is worth apparently higher than remaining four kinds of modulation system；8PSK feature ginseng
Number T8It is worth apparently higher than its excess-three kind modulation system；16QAM characteristic parameter T16It is worth apparently higher than remaining two kinds of modulation system；
32QAM characteristic parameter T32Value is also slightly above 64QAM modulation systems；Therefore, T is passed through respectively4、T8、T16、T32Can be modulation methods
Formula 2PSK, 8PSK, 16QAM, 32QAM, which are sorted out, to be come, and realizes the Modulation Identification to six kinds of signals.
Step S4：Using the combination of some characteristic parameters as the input of SVM classifier, the SVM classifier is entered
Row training.The present embodiment is from the Modulation Identification performance of the system of raising, the six kinds of features extracted using above clustering algorithm
Input of the Combination nova of parameter as SVM, is trained to SVM classifier, then using the SVM classifier after training to the above
Six kinds of modulation systems based on planisphere are identified.Run into simultaneously in order to overcome during conventional SVMs identification multi-class problem
The high shortcoming of computation complexity, and reach the required precision of setting；As shown in figure 3, herein using classification algorithm to support to
Amount machine is trained.
Step S5：The unknown signaling is identified using the SVM classifier after training, in exporting for SVM classifier
To the modulation type of the unknown signaling.
The advantages of in order to verify the SVM classifier using classification training algorithm training, the present embodiment provides one to having a competition
Test, the SVM classifier using the training of classification training algorithm and conventional training algorithm is compared, computational complexity emulation ratio
Compared with as shown in figure 4, SVM classifications training algorithm computational complexity compared with conventional SVM training algorithms of the present embodiment use is big
To lower, the time as used in emulation data understand SVM classification training algorithms is 338.57 seconds, conventional SVM training algorithms institute's used time
Between 589.65 seconds, SVM classification training algorithm computational complexity greatly lower.
The present embodiment individually using clustering method, conventional SVM classifier Modulation Identification method and set forth herein joint tune
Recognition methods processed is emulated and is compared.The present embodiment is to common using 2PSK, 4PSK, 8PSK, 16QAM, 32QAM, 64QAM
Six kinds of modulated signals, its carrier signal are sine wave, frequency fc=4MHZ, using frequency fs=120MHZ, chip rate fd=
1MHZ, single analysis is using points (data length) Ns=2048.When signal to noise ratio is respectively -2dB, 0dB, 4dB, 8dB, 10dB
To individually using clustering method, conventional SVM classifier Modulation Identification method and set forth herein combined modulation recognition methods progress
Emulation, each case are respectively tested 1000 times, obtain the statistics of correct recognition rata.And assume the probability that every kind of modulation system occurs
It is identical, so as to calculate under different signal to noise ratio system to the average recognition rate of various modulation systems.Fig. 5, Fig. 6, Fig. 7 distinguish
For individually using K- means clustering algorithms, based on conventional SVM classifier Modulation Identification method and using cluster and SVM combined modulations
The Modulation Identification rate analogous diagram of recognizer constantly.
In Figure 5, six kinds of signals to be identified, with the difference of order of modulation, K- means Clusterings value is different, six kinds
Signal modulation correct recognition rata difference is larger, especially when signal to noise ratio is relatively low, due to the influence of noise, receives planisphere and compares point
Dissipate, now directly use cluster result and unreliable.
In figure 6, six kinds of signals to be identified, when being modulated identification using conventional BP algorithm SVM classifier, its discrimination
It is more or less the same, and in the case of low signal-to-noise ratio, correct recognition rata is not also high.
Fig. 7 is understood compared with Fig. 5, Fig. 6 respectively, under different signal to noise ratio, with individually using clustering algorithm or based on normal
Rule BP algorithm SVM classifier is compared, and the discrimination of algorithm presented here is significantly improved.When signal to noise ratio is 4dB, four
Kind modulation system 2PSK, 4PSK, 8PSK, 16QAM discrimination up to more than 90%, 32QAM and 64QAM Modulation Identification rate also have
Significantly improve, 32QAM has reached 98%, 64QAM in SNR=6dB, and in SNR=2dB, Modulation Identification is up to 70%, far above list
Solely using discrimination during cluster.The main reason for system identification performance improves：One is due to using clustering algorithm, to be identified
The characteristic value of signal extraction optimizes processing；Second, herein using K mean cluster and classification training SVM combined modulation identification
Algorithm, overcome the problem of computation complexity is high when identifying multi-class problem, convergence rate significantly improves so that in reception signal
When signal to noise ratio is relatively low, Modulation Identification rate is also greatly improved.
The above described is only a preferred embodiment of the present invention, any formal limitation not is made to the present invention, therefore
Every any modification that without departing from technical solution of the present invention content, the technical spirit according to the present invention is made to above example,
Equivalent variations and modification, in the range of still falling within technical solution of the present invention.
Claims (8)
1. the combined modulation recognition methods of a kind of K- mean clusters and classification training SVM classifier, it is characterised in that including as follows
Step：
Step S1：Unknown signaling is pre-processed, obtains the data set X of the in-phase component comprising signal and quadrature component；
Step S2：Cluster computing is carried out to the sample point of the data set X, obtains person in servitude of each sample point to cluster centre
Category degree matrix；
Step S3：The subordinated-degree matrix is handled with Validity Function, if obtaining for distinguishing different modulating mode
Dry characteristic parameter；
Step S4：Using the combination of some characteristic parameters as the input of SVM classifier, the SVM classifier is instructed
Practice；
Step S5：The unknown signaling is identified using the SVM classifier after training, institute is obtained in the output of SVM classifier
State the modulation type of unknown signaling.
2. the combined modulation recognition methods of a kind of K- mean clusters according to claim 1 and classification training SVM classifier,
Characterized in that,
In step sl, unknown signaling is pre-processed specially：Carrier wave is carried out at down coversion, LPF and sampling
Reason.
3. the combined modulation recognition methods of a kind of K- mean clusters according to claim 1 and classification training SVM classifier,
Characterized in that,
Data set X={ the x1,x2,…xN, wherein, N is the number of element in data set, i.e. the sample glyph of unknown signaling
Number.
4. the combined modulation recognition methods of a kind of K- mean clusters according to claim 3 and classification training SVM classifier,
Characterized in that,
In step s 2, clustering algorithm is specifically calculated using K- mean clusters, and its classification matrix is V(0), it is denoted as mathematics rule
The problem of drawing, object function are：
Constraints is：K is cluster centre number, 1 ＜ K ＜ N.
5. the combined modulation recognition methods of a kind of K- mean clusters according to claim 4 and classification training SVM classifier,
Characterized in that,
In step s 2, modulated signal corresponding to different cluster centre numbers is iterated, detailed process includes：
Step S21：Provide iteration criterion epsilon>0, and initialize classification matrix V(0), n=0；
Step S22：Calculate renewal Iterative Matrix U(n),
Step S23：Calculate cluster centre matrix V(n+1),
Step S24：With matrix norm | | | | compare V(n+1)With V(n)If | | V(n+1)-V(n)| |≤ε is set up, then stops iteration, no
Then, K=K+1, repeat step step S22 to step S24 are made；In iterative process, optimization object functionObtain subordinated-degree matrix.
6. the combined modulation recognition methods of a kind of K- mean clusters according to claim 5 and classification training SVM classifier,
Characterized in that,
Step S3 detailed process is：Efficiency analysis is carried out to the cluster result obtained during different cluster centre number K values, judged
It is whether reasonable that received signal points are divided into K classes, Validity Function value are obtained, as distinguishing some of different modulating mode
Characteristic parameter.
7. the combined modulation recognition methods of a kind of K- mean clusters according to claim 6 and classification training SVM classifier,
Characterized in that,
In step s 4, SVM classifier is trained using the algorithm of classification.
A kind of 8. combined modulation of the K- mean clusters and classification training SVM classifier according to claim any one of 3-7
Recognition methods, it is characterised in that unknown modulated signal is：2PSK, 4PSK, 8PSK, 16QAM, 32QAM and 64QAM；Corresponding tune
Exponent number processed is respectively：1,2,3,4,5 and 6；Take cluster centre number K={ 2,4,8,16,32,64 }.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710993729.7A CN107770108A (en) | 2017-10-23 | 2017-10-23 | A kind of combined modulation recognition methods of K mean clusters and classification training SVM classifier |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710993729.7A CN107770108A (en) | 2017-10-23 | 2017-10-23 | A kind of combined modulation recognition methods of K mean clusters and classification training SVM classifier |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107770108A true CN107770108A (en) | 2018-03-06 |
Family
ID=61269171
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710993729.7A Pending CN107770108A (en) | 2017-10-23 | 2017-10-23 | A kind of combined modulation recognition methods of K mean clusters and classification training SVM classifier |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107770108A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108600135A (en) * | 2018-04-27 | 2018-09-28 | 中国科学院计算技术研究所 | A kind of recognition methods of signal modulation mode |
| CN109286458A (en) * | 2018-10-31 | 2019-01-29 | 天津大学 | Cooperation frequency spectrum sensing method based on fuzzy support vector machine |
| CN109347583A (en) * | 2018-10-31 | 2019-02-15 | 天津大学 | Cooperation frequency spectrum sensing method based on nuclear space optimization under noise uncertain condition |
| CN109617843A (en) * | 2018-12-28 | 2019-04-12 | 上海铿诚智能科技有限公司 | A kind of elastic optical network modulation format recognition methods based on KNN |
| CN109787928A (en) * | 2019-01-25 | 2019-05-21 | 中国人民解放军国防科技大学 | Decision tree TWSVM-based high-speed optical signal modulation format recognition method |
| CN110320033A (en) * | 2019-08-08 | 2019-10-11 | 广东省智能机器人研究院 | A kind of rolling bearing health degenerate state discrimination method |
| CN110381002A (en) * | 2019-07-18 | 2019-10-25 | 黄山市上光科技有限公司 | A kind of QAM-OFDM equalization algorithm based on SVM |
| CN111666247A (en) * | 2020-05-26 | 2020-09-15 | 中国人民解放军国防科技大学 | Hierarchical training method and device supporting link physical layer serial signals |
| CN111814917A (en) * | 2020-08-28 | 2020-10-23 | 成都千嘉科技有限公司 | Character wheel image digital identification method with fuzzy state |
| CN113188570A (en) * | 2021-04-27 | 2021-07-30 | 西南石油大学 | Attitude error calibration method of inclinometer while drilling based on support vector classifier and K-proximity method |
| WO2022062533A1 (en) * | 2020-09-24 | 2022-03-31 | 长鑫存储技术有限公司 | Semiconductor product grading method and system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101917369A (en) * | 2010-07-30 | 2010-12-15 | 中国人民解放军信息工程大学 | Method for identifying modulation mode of communication signal |
| CN102279390A (en) * | 2011-05-06 | 2011-12-14 | 西南交通大学 | Intra-pulse modulation and recognition method of low signal-to-noise radar radiation source signal |
| CN102497343A (en) * | 2011-11-25 | 2012-06-13 | 南京邮电大学 | Combined modulation recognition method based on clustering and support vector machine |
| KR101427149B1 (en) * | 2014-03-06 | 2014-08-07 | 국방과학연구소 | A method of recognizing complex PRI modulation type of radar signals based on 2-layer SVM using exponential moving average |
-
2017
- 2017-10-23 CN CN201710993729.7A patent/CN107770108A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101917369A (en) * | 2010-07-30 | 2010-12-15 | 中国人民解放军信息工程大学 | Method for identifying modulation mode of communication signal |
| CN102279390A (en) * | 2011-05-06 | 2011-12-14 | 西南交通大学 | Intra-pulse modulation and recognition method of low signal-to-noise radar radiation source signal |
| CN102497343A (en) * | 2011-11-25 | 2012-06-13 | 南京邮电大学 | Combined modulation recognition method based on clustering and support vector machine |
| KR101427149B1 (en) * | 2014-03-06 | 2014-08-07 | 국방과학연구소 | A method of recognizing complex PRI modulation type of radar signals based on 2-layer SVM using exponential moving average |
Non-Patent Citations (2)
| Title |
|---|
| LIU AISHENG ET.AL: "Automatic modulation classification based on the combination of clustering and neural network", 《THE JOURNAL OF CHINA UNIVERSITIES OF POSTS AND TELECOMMUNICATIONS》 * |
| 杨发权: "D1基于聚类与神经网络的无线通信联合调制识别新方法", 《中山大学学报(自然科学版)》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108600135A (en) * | 2018-04-27 | 2018-09-28 | 中国科学院计算技术研究所 | A kind of recognition methods of signal modulation mode |
| CN108600135B (en) * | 2018-04-27 | 2020-07-31 | 中国科学院计算技术研究所 | Method for identifying signal modulation mode |
| CN109286458A (en) * | 2018-10-31 | 2019-01-29 | 天津大学 | Cooperation frequency spectrum sensing method based on fuzzy support vector machine |
| CN109347583A (en) * | 2018-10-31 | 2019-02-15 | 天津大学 | Cooperation frequency spectrum sensing method based on nuclear space optimization under noise uncertain condition |
| CN109617843A (en) * | 2018-12-28 | 2019-04-12 | 上海铿诚智能科技有限公司 | A kind of elastic optical network modulation format recognition methods based on KNN |
| CN109617843B (en) * | 2018-12-28 | 2021-08-10 | 上海铿诚智能科技有限公司 | KNN-based elastic optical network modulation format identification method |
| CN109787928B (en) * | 2019-01-25 | 2021-07-09 | 中国人民解放军国防科技大学 | Decision tree TWSVM-based high-speed optical signal modulation format recognition method |
| CN109787928A (en) * | 2019-01-25 | 2019-05-21 | 中国人民解放军国防科技大学 | Decision tree TWSVM-based high-speed optical signal modulation format recognition method |
| CN110381002A (en) * | 2019-07-18 | 2019-10-25 | 黄山市上光科技有限公司 | A kind of QAM-OFDM equalization algorithm based on SVM |
| CN110320033A (en) * | 2019-08-08 | 2019-10-11 | 广东省智能机器人研究院 | A kind of rolling bearing health degenerate state discrimination method |
| CN111666247A (en) * | 2020-05-26 | 2020-09-15 | 中国人民解放军国防科技大学 | Hierarchical training method and device supporting link physical layer serial signals |
| CN111814917A (en) * | 2020-08-28 | 2020-10-23 | 成都千嘉科技有限公司 | Character wheel image digital identification method with fuzzy state |
| CN111814917B (en) * | 2020-08-28 | 2020-11-24 | 成都千嘉科技有限公司 | Character wheel image digital identification method with fuzzy state |
| WO2022062533A1 (en) * | 2020-09-24 | 2022-03-31 | 长鑫存储技术有限公司 | Semiconductor product grading method and system |
| CN113188570A (en) * | 2021-04-27 | 2021-07-30 | 西南石油大学 | Attitude error calibration method of inclinometer while drilling based on support vector classifier and K-proximity method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107770108A (en) | A kind of combined modulation recognition methods of K mean clusters and classification training SVM classifier | |
| CN109802905B (en) | CNN convolutional neural network-based digital signal automatic modulation identification method | |
| CN107124381B (en) | Automatic identification method for digital communication signal modulation mode | |
| CN110855591B (en) | QAM and PSK signal intra-class modulation classification method based on convolutional neural network structure | |
| Wu et al. | Robust automatic modulation classification under varying noise conditions | |
| CN108234370B (en) | Communication signal modulation mode identification method based on convolutional neural network | |
| CN104994045B (en) | A kind of digital modulation mode automatic identification platform and method based on USRP platforms | |
| CN102497343A (en) | Combined modulation recognition method based on clustering and support vector machine | |
| CN108052956A (en) | Wireless light communication subcarrier modulation constellation recognition methods under a kind of atmospheric turbulance | |
| CN107135017A (en) | Backscatter communication system signal method of sending and receiving | |
| CN102647391A (en) | Cooperative modulation signal identifying method based on data fusion of decision layer | |
| CN101895494B (en) | Stochastic resonance preprocessing-based digital modulation mode automatic identification method | |
| CN107612867A (en) | A kind of order of modulation recognition methods of MQAM signals | |
| CN111832462A (en) | Frequency hopping signal detection and parameter estimation method based on deep neural network | |
| CN100521670C (en) | Detecting and analyzing method for multi system frequency shift key control signal | |
| CN102263716A (en) | Modulation type identifying method and system | |
| CN110601764A (en) | Radio frequency modulation format identification method based on optical assistance | |
| Almohamad et al. | Dual-determination of modulation types and signal-to-noise ratios using 2D-ASIQH features for next generation of wireless communication systems | |
| Subbarao et al. | Automatic modulation recognition in cognitive radio receivers using multi-order cumulants and decision trees | |
| CN110826425B (en) | VHF/UHF frequency band radio signal modulation mode identification method based on deep neural network | |
| CN110166389B (en) | Modulation identification method based on least square support vector machine | |
| Zhang et al. | High-order convolutional attention networks for automatic modulation classification in communication | |
| CN108809874B (en) | Radar and communication multi-signal classification method based on circulation support vector machine | |
| CN109145889A (en) | A kind of bright ciphertext Modulation recognition detection method carrying out blind estimate for wireless signal | |
| CN113869156A (en) | Radio frequency fingerprint identification method |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180306 |
|
| RJ01 | Rejection of invention patent application after publication |