CN106093899A - A kind of radar operation mode recognition methods - Google Patents
A kind of radar operation mode recognition methods Download PDFInfo
- Publication number
- CN106093899A CN106093899A CN201610729293.6A CN201610729293A CN106093899A CN 106093899 A CN106093899 A CN 106093899A CN 201610729293 A CN201610729293 A CN 201610729293A CN 106093899 A CN106093899 A CN 106093899A
- Authority
- CN
- China
- Prior art keywords
- support vector
- vector machine
- radar
- machine classifier
- operation mode
- 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
-
- 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
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/40—Means for monitoring or calibrating
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention discloses a kind of radar operation mode recognition methods, the method includes following key step: (1) sets up set of factors;(2) structure passes judgment on collection;(3) determination of support vector machine classifier parameter.
Description
Technical field
The present invention is a kind of radar operation mode recognition methods, for improving the recognition accuracy to radar operation mode.
Background technology
Radar operation mode identification is an important content in recognizing radar radiation source, and it belongs to after classifying with stream of pulses
The train of pulse of same portion radar, for processing object, is used for the currently employed mode of operation of Discrimination Radar radiation source, technique and tactics
The contents such as way.
Identification to radar operation mode at present is typically based on conventional five big parametric evolving, i.e. arrives angle (DOA), carrier frequency
(RF), the time of advent (TOA), pulsewidth (PW) and pulse amplitude (PA).On this basis, quadratic parameter pulse is repeated frequency by the present invention
Rate (PRF) and data transfer rate (DR) include identification in, and increase the judge to support vector machine classifier parameter and select link,
To obtain preferable recognition accuracy.
Summary of the invention
(1) set of factors is set up;
(2) structure passes judgment on collection;
(3) determination of support vector machine classifier parameter.
Accompanying drawing explanation
Accompanying drawing 1 is the flow chart of the present invention.Referring to the drawings 1, the flow process of the present invention by setting up set of factors, structure passes judgment on collection and
The determination of support vector machine classifier parameter 3 part composition.Wherein 1 for establishing the parameter of Discrimination Radar mode of operation;2 use
In providing the mode of operation classification that signal to be identified is possible;3 determination being used for penalty coefficient and the selections of kernel function.
Detailed description of the invention
The principle implementing the present invention is as follows: for the radar signal after sorting, utilizes the data time of advent in PDW to calculate
Quadratic parameter, i.e. pulse repetition period and data transfer rate;Carrier frequency in PDW, pulsewidth are carried out group with pulse repetition period, data transfer rate
Close, obtain parameter set;Select suitable kernel function, model parameter and many classification schemes that SVM is trained afterwards, utilize training
After SVM radar emitter signal is identified.
(1) set of factors is set up
Set of factors is the input of support vector machine classifier, mode of operation identification the characteristic parameter utilized is constituted, mainly
Including four parameters, i.e. carrier frequency (RF), pulsewidth (PW), pulse recurrence frequency (PDF) and data transfer rate (DR).
(2) structure passes judgment on collection
Passing judgment on the output that collection is support vector machine classifier, reflection is final recognition result, and corresponding is to be identified
The mode of operation classification that signal is possible.
(3) determination of support vector machine classifier parameter
1) determination of penalty coefficient
Penalty factor > 0 conduct balanceWithWeight, decide the punishment degree that mistake is divided sample, it determines
Most important, there is document, around penalty coefficient, the impact of SVM recognition effect is expanded in-depth study, the present invention grinds with reference to it
Studying carefully achievement, when utilizing SVM to be operated pattern recognition as grader, taking penalty coefficient C is 2.
2) selection of kernel function
Non-linear sample can be classified by SVM, and principle is that it utilizes kernel function by sample from inseparable low-dimensional
Space is mapped to the higher dimensional space that can divide, and nonlinear non-separable problem is thus converted to linear separability problem.
Conventional kernel function has:
A. linear kernel function (Linear)
K (x, x')=x x'(1) this kernel function is a kind of special case of polynomial kernel.
B. Polynomial kernel function (Polynomial is called for short Poly)
K (x, x')=(v (x x')+c)p (2)
This kernel functional parameter is many, calculates complexity, and in formula (2), v typically takes 1, according to the difference of c and p value, can obtain
Remaining specific form of this kernel function in addition to linear kernel, as
Quantic kernel function: K (x, x')=(x x')p, i.e. c=0, p ∈ R+;
Inhomogeneous polynomial kernel function: K (x, x')=((x x')+c)p, i.e. c, p ∈ R+
C. gaussian kernel function (Gauss)
K (x, x')=exp (-x-x'| |/2 δ2) (3)
This kernel function need not priori, and kernel functional parameter δ controls the performance of kernel function, and this function is also radially base
(RBF) kernel function.
D. multilayer perceptron kernel function (Sigmoid)
K (x, x')=tanh [v (x x')+c] (4)
In formula (4), v, c > 0.When using this kernel function, SVM is equivalent to comprise a multilayer perceptron.
Advantage below in conjunction with the most whole invention of example.
If through long-term detecting, the radar model grasped in prior data bank has A, B, C, D.Simulation produces thunder after sorting
Reach signal overall pulse data sequence, the different working modes of different model radar is produced respectively 1000 groups of samples, wherein 500 groups
Sample is used for training, and 500 groups of samples are used for testing, and signal parameter scope is as shown in table 1.
Table 1 radar parameter scope
The overall pulse data produced for emulation, first carry out quadratic parameter calculating, extract the pulse recurrence frequency of signal
And data transfer rate, then carrier frequency, pulsewidth, repetition, four characteristic parameters of data transfer rate being carried out independent assortment, composition characteristic is vectorial, and
It is utilized respectively BP neutral net, PNN neutral net and three kinds of graders of support vector machine and is operated pattern recognition experiment.Its
In, the maximum iteration time of BP neutral net is set to 500, training objective error is set to 0.0001.
Carrier frequency, pulsewidth, repetition, four characteristic parameters of data transfer rate are carried out independent assortment, separately constitute one-dimensional, two-dimentional, three
Dimension, four dimensional feature vectors, for identifying multiple radar radar signal under different working modes, select BP neutral net, PNN
9 mode of operations of 4 shown in table a kind radar, as the grader of pattern recognition, are comprehensively known by neutral net and SVM
, the recognition result not obtained is as shown in table 2.
The comprehensive recognition result of mode of operation (%) under table 2 different characteristic Parameter Conditions
Experimental result has absolutely proved the advantage of the present invention, selects four dimensional feature vectors can obtain preferably based on SVM
Radar operation mode recognition accuracy.
Claims (1)
1. a radar operation mode recognition methods, it is characterised in that include following technical measures:
(1) set of factors is set up
Set of factors is the input of support vector machine classifier, mode of operation identification the characteristic parameter utilized is constituted, and mainly includes
Four parameters, i.e. carrier frequency (RF), pulsewidth (PW), pulse recurrence frequency (PDF) and data transfer rate (DR)
(2) structure passes judgment on collection
Passing judgment on the output that collection is support vector machine classifier, reflection is final recognition result, and corresponding is signal to be identified
Possible mode of operation classification
(3) determination of support vector machine classifier parameter
Specifically include the determination of penalty coefficient and the selection of kernel function.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610729293.6A CN106093899A (en) | 2016-08-25 | 2016-08-25 | A kind of radar operation mode recognition methods |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610729293.6A CN106093899A (en) | 2016-08-25 | 2016-08-25 | A kind of radar operation mode recognition methods |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106093899A true CN106093899A (en) | 2016-11-09 |
Family
ID=57226030
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610729293.6A Pending CN106093899A (en) | 2016-08-25 | 2016-08-25 | A kind of radar operation mode recognition methods |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106093899A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107561530A (en) * | 2017-07-27 | 2018-01-09 | 中国船舶重工集团公司第七二四研究所 | A kind of target identification method based on collaboration passive detection multidimensional information |
CN108197146A (en) * | 2017-11-29 | 2018-06-22 | 山东航天电子技术研究所 | The essence classification in-orbit generation system of Radar recognition parameter based on pulse flow data |
CN108983167A (en) * | 2018-05-22 | 2018-12-11 | 中国电子科技集团公司第三十八研究所 | Radar universal description modeling method and device |
CN109272040A (en) * | 2018-09-20 | 2019-01-25 | 中国科学院电子学研究所苏州研究院 | A kind of radar operation mode generation method |
CN112884059A (en) * | 2021-03-09 | 2021-06-01 | 电子科技大学 | Small sample radar working mode classification method fusing priori knowledge |
-
2016
- 2016-08-25 CN CN201610729293.6A patent/CN106093899A/en active Pending
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107561530A (en) * | 2017-07-27 | 2018-01-09 | 中国船舶重工集团公司第七二四研究所 | A kind of target identification method based on collaboration passive detection multidimensional information |
CN108197146A (en) * | 2017-11-29 | 2018-06-22 | 山东航天电子技术研究所 | The essence classification in-orbit generation system of Radar recognition parameter based on pulse flow data |
CN108983167A (en) * | 2018-05-22 | 2018-12-11 | 中国电子科技集团公司第三十八研究所 | Radar universal description modeling method and device |
CN109272040A (en) * | 2018-09-20 | 2019-01-25 | 中国科学院电子学研究所苏州研究院 | A kind of radar operation mode generation method |
CN112884059A (en) * | 2021-03-09 | 2021-06-01 | 电子科技大学 | Small sample radar working mode classification method fusing priori knowledge |
CN112884059B (en) * | 2021-03-09 | 2022-07-05 | 电子科技大学 | Small sample radar working mode classification method fusing priori knowledge |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106093899A (en) | A kind of radar operation mode recognition methods | |
CN106855941B (en) | Gesture identification method and system based on the sparse optimization of radar micro-doppler signal | |
CN111160176B (en) | Fusion feature-based ground radar target classification method for one-dimensional convolutional neural network | |
CN104077787B (en) | A kind of Aircraft Targets sorting technique based on time domain and Doppler domain | |
CN107830996B (en) | Fault diagnosis method for aircraft control surface system | |
CN102721952B (en) | Method for steadily classifying ground moving target based on super-resolution Doppler spectrum | |
Boddy et al. | Identification of 72 phytoplankton species by radial basis function neural network analysis of flow cytometric data | |
CN108182450A (en) | A kind of airborne Ground Penetrating Radar target identification method based on depth convolutional network | |
CN102043148B (en) | Ground moving target classification method based on general matched filtering | |
CN108051781A (en) | A kind of radar signal operating mode recognition methods based on DBN model | |
CN104809358A (en) | Radar radiation source identification method based on phase noise unintentional modulation characteristic | |
CN103824093B (en) | It is a kind of based on KFDA and SVM SAR image target's feature-extraction and recognition methods | |
CN111175718A (en) | Time-frequency domain combined ground radar automatic target identification method and system | |
Cain et al. | Convolutional neural networks for radar emitter classification | |
CN107818298A (en) | General Raman spectral characteristics extracting method for machine learning material recognition | |
CN103886342A (en) | Hyperspectral image classification method based on spectrums and neighbourhood information dictionary learning | |
CN103885043B (en) | The miscellaneous Robust classification method of making an uproar of Aircraft Targets based on broad match filtering | |
CN107330457B (en) | A kind of Classification of Polarimetric SAR Image method based on multi-feature fusion | |
CN107169469A (en) | A kind of material identification method of the MIMO radar based on machine learning | |
CN102184382A (en) | Empirical mode decomposition based moving vehicle target classification method | |
CN104680184A (en) | Polarization SAR terrain classification method based on deep RPCA | |
CN102982339A (en) | Hyperspectral characteristic variable selection method | |
CN104199007A (en) | Radar distributed ground target discrimination method based on neighbor one-class classifiers | |
CN107682109A (en) | A kind of interference signal classifying identification method suitable for UAV Communication system | |
CN104850844A (en) | Pedestrian detection method based on rapid construction of image characteristic pyramid |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20161109 |