CN105721371A - Method for identifying common digital modulation signal based on cyclic spectrum correlation - Google Patents

Abstract

The invention discloses a method for identifying a common digital modulation signal based on cyclic spectrum correlation. The reliability of signal analysis is improved by utilizing the noise-proof feature of a signal cyclic spectrum; the steps of alpha section wavelet de-noising and averaging through superposition are introduced into a calculation process of a signal spectral correlation function, so that the random fluctuation caused by the limited sampling number and the external disturbance in an original spectrum correlation estimation algorithm result is effectively weakened to facilitate identification and extraction of modulation features; and meanwhile, an alpha section and an f section of an obtained spectral correlation diagram are computed by utilizing signal spectral correlation, and appropriate features and parameters (such as a ratio of maximum absolute values of spectral correlation functions, namely the alpha section and the f section, the number of intense lines of the alpha section, a coefficient of fluctuation of the alpha section, the normalized area of the f section, a predominance ratio of spectral lines of the alpha section and the like) are selected to construct a classification method to identify the modulation mode of the communication signal.

Description

A kind of commonly used digital Modulation Signals Recognition method relevant based on Cyclic Spectrum

Technical field

The invention belongs to Modulation Recognition of Communication Signal field, be specifically related to a kind of commonly used digital Modulation Signals Recognition method relevant based on Cyclic Spectrum.

Technical background

The Modulation Mode Recognition of signal of communication is an important step between signal acquisition and demodulation, is all play key player in military or commercial communication field, particularly in the management of dynamic radio frequency spectrum and the discriminating work of not clear interference.It addition, the Modulation Mode Recognition of signal of communication is also the basis building software radio or cognitive radio application, also provide technical support for many systems communication interconnecting application.

Signal Power Spectrum Analysis based on Fourier transformation is a kind of conventional comparatively classical analysis method for stationary signal.By the power spectrumanalysis of signal, it is possible to obtain the basic feature of analyzed signal preferably, including the signal parameter such as carrier frequency, bandwidth.But in communication process, signal is converted to non-stationary signal due to manual intervention (such as modulation, coding, scanning etc.) greatly, power spectrumanalysis can not disclose the modulation signature of signal comparatively all sidedly, therefore generally will by more Time-domain Statistics feature based on the Modulation Identification method of power spectrumanalysis, limitation is bigger.The Cyclic Spectrum correlation theory that Gardner proposes is that people study an important tool of non-stationary signal in recent years, this theory has taken into full account the critical nature cyclo-stationary that signal has after periodicity manual intervention, can disclosing the modulation signature of signal of communication comparatively all sidedly, this is that classical power spectrumanalysis is incomparable.

Although Cyclic Spectrum correlation theory obtains relatively broad application in the Non-stationary Signal Analysis in each field, but does not popularize in Modulation Recognition of Communication Signal field at present.

Summary of the invention

The technical problem to be solved in the present invention is: overcome the deficiencies in the prior art, it is provided that a kind of commonly used digital Modulation Signals Recognition method relevant based on Cyclic Spectrum, utilizes the noiseproof feature that signal cycle spectrum has to improve the reliability of signal analysis；And average link is asked in introducing alpha cross section Wavelet Denoising Method and superposition in the calculating process of signal spectrum correlation function, effectively reduce in former spectrum correlation estimation algorithm result by the limited random fluctuation caused with external interference of sampling number, be beneficial to identification and the extraction of modulation signature；Simultaneously, utilize alpha cross section and the f cross section of the relevant figure of the acquired spectrum of signal spectrum correlation computations, choose suitable feature and parameter (as Spectral correlation function alpha cross section and f cross section maximum value ratio, alpha cross section intense line number, alpha cross section coefficient of variation, f cross section normalized area, alpha cross section spectral line significance than etc.) build sorting technique the modulation system of signal of communication be identified.

The technical solution adopted for the present invention to solve the technical problems is: a kind of commonly used digital Modulation Signals Recognition method relevant based on Cyclic Spectrum, including following step:

(1) after sampled for the signal to noise ratio modulation signal of intercepting and capturing, it is divided into n equal portions, as the input signal carrying out spectrum related operation；

(2) choose suitable Fourier transform to count and smoothing windows width, respectively this n part signal is made by the spectrum correlation estimation algorithm based on frequency domain smoothing spectrum related operation；

The described spectrum correlation estimation algorithm based on frequency domain smoothing is as follows:

$S_{x\mathrm{\Δ}t}^{\mathrm{\α}}{(t,f)}_{\mathrm{\Δ}f}=\frac{1}{M}\underset{v=-(M-1)/2}{\overset{(M-1)/2}{\Σ}}\frac{1}{\mathrm{\Δ}t}{X}_{\mathrm{\Δ}t}(t,f+\frac{\mathrm{\α}}{2}+{\mathrm{vF}}_S)X_{\mathrm{\Δ}t}^{*}(t,f-\frac{\mathrm{\α}}{2}+{\mathrm{vF}}_S)$

$X_{\mathrm{\Δ}t}(t,f)=\underset{K=0}{\overset{N-1}{\Σ}}{a}_{\mathrm{\Δ}t}\left({\mathrm{KT}}_S\right)x_k(t-{\mathrm{KT}}_S)\exp (-j2\mathrm{\π}f(t-{\mathrm{KT}}_S\left)\right)$

Wherein,For the result after spectrum related operation；X_Δt(t, f) for signal x_kThe result of (t) short time discrete Fourier transform；Signal x_kOne of t n equal portions that () is institute's intercepted signal；Δ t is x_k(t) persistent period；A_ΔtIt it is window function；Δ f is frequency domain smoothing interval；F_SFor frequency domain smallest incremental unit；T_SIt it is time-domain sampling interval；N is signal x_k(t) sample number；M is smoothing windows spread factor；α is cycle frequency；F is frequency；T is the time；

(3) operation result in step (2) is added after alpha cross section Wavelet Denoising Method and averages, the output result of the spectrum correlation estimation algorithm being improved；

The spectrum correlation estimation algorithm of described improvement is as follows:

$S_{x\mathrm{\Δ}t}^{\mathrm{\α}}\left(f\right)=\frac{1}{N}\underset{k=1}{\overset{N}{\Σ}}{\mathrm{WDN}}_{\mathrm{\α}}\[S_{x\mathrm{\Δ}t}^{\mathrm{\α}}(t_k,f)\]$

Wherein, WDN_α() represents rightIn cycle frequency α carry out Wavelet Denoising Method；

(4) extract 5 kinds of Modulation Identification features according to the alpha cross section of step (3) gained Spectral correlation function and f cross section, and in conjunction with a kind of signal characteristic structure sorting technique based on Time-domain Statistics, several digital modulation signals are identified；Described 5 kinds of Modulation Identification are characterized as: Spectral correlation function alpha cross section and f cross section maximum value ratio (are designated as R₁), alpha cross section intense line number (be designated as R₂), alpha cross section coefficient of variation (be designated as R₃), f cross section normalized area (be designated as R₄), the significance ratio of alpha cross section spectral line (be designated as R₅)；Described a kind of statistical nature based on signal time domain is: the standard deviation of zero center normalization instantaneous amplitude absolute value (is designated as R₆)；Described several digital modulation signals includes 2ASK, 4ASK, BPSK, QPSK, 8PSK, MSK, 2FSK, 4FSK, 2FSK*, 4FSK*, and wherein FSK* represents the incoherent frequency shift keying of code element initial phase；

The described identification process to several digital modulation signals is:

IfThen judge that the modulation system of this signal is as MSK；

If $\{{\hat{R}}_1>r_{12},{\hat{R}}_2>r_{21}\}$ Or $\{{\hat{R}}_1<r_{12},{\hat{R}}_3<r_{31}\},$ Then judge that the modulation methods of this signal is as 4FSK；

IfThen judge that the modulation system of this signal is as 2FSK；

IfThen judge that the modulation system of this signal is as BPSK；

If $\{{\hat{R}}_1>r_{11},{\hat{R}}_2<r_{22},{\hat{R}}_5<r_{51},{\hat{R}}_6<r_{61}\},$ Then judge that the modulation system of this signal is as 2ASK；

If $\{{\hat{R}}_1>r_{11},{\hat{R}}_2<r_{22},{\hat{R}}_5<r_{51},{\hat{R}}_6>r_{61}\},$ Then judge that the modulation system of this signal is as 4ASK；

IfThen judge the modulation system 2FSK* of this signal；

IfThen judge that the modulation system of this signal is as 4FSK*；

If $\{{\hat{R}}_1<r_{12},{\hat{R}}_3>r_{31},{\hat{R}}_4<r_{42},{\hat{R}}_5>r_{52}\},$ Then judge the modulation system QPSK of this signal；

If $\{{\hat{R}}_1<r_{12},{\hat{R}}_3>r_{31},{\hat{R}}_4<r_{42},{\hat{R}}_5<r_{52}\},$ Then judge the modulation system 8PSK of this signal；

WhereinIt is feature R₁Estimated value, r₁₁And r₁₂For character pair R₁Threshold value；All the other are similar.

In described step (1), the sample frequency of signal meets

In described step (2), the sampling number of every part of signal is taken as N=1024, and smoothing windows spread factor is M=63.

In described step (3), the principle of Wavelet Denoising Method is: use sym8 small echo that noisy sequence is decomposed, on the layer 5 decomposed, use soft sure thresholding system of selection that sequence carries out denoising, and this thresholding adjusts with the noise variance of ground floor wavelet decomposition.

Described 5 kinds in step (4) are as follows based on the definition of the Modulation Identification feature of modulation signal spectrum correlation function:

1. Spectral correlation function alpha cross section and f cross section maximum value compare R₁It is defined as:

$R_1=\underset{N\&RightArrow;\mathrm{\&infin;}}{\lim }\frac{1}{N}\underset{n=1}{\overset{N}{\&Sigma;}}\frac{\max \left(\right|S_{xn}^{\mathrm{\&alpha;}}\left(0\right)\left|\right)}{\max \left(\right|S_{xn}^0\left(f\right)\left|\right)}$

WhereinWithRespectively Spectral correlation functionAlpha cross section and f cross-section function；

2. alpha cross section intense line number R₂It is defined as:

$R_2=\underset{N\&RightArrow;\mathrm{\&infin;}}{\lim }\frac{1}{N}\underset{n=1}{\overset{N}{\&Sigma;}}count({\mathrm{\&rho;}}_n\&GreaterEqual;{\mathrm{\&rho;}}_{th}*{\mathrm{\&rho;}}_{max})$

Wherein count () represents the notable peak value seeking the relevant figure alpha cross section of spectrum, ρ_nFor the significance of cross section crest, ρ_maxFor the maximum significance of cross section crest, ρ_thFor significance threshold value, take definite value 0.27；

Its medium wave peak significance ρ is defined as:

$\mathrm{\&rho;}=\frac{h^2}{l*max\left(h\right)}$

Wherein h is the relevant figure alpha cross section medium wave peak amplitude of spectrum and the difference of the greater in adjacent two trough amplitudes, and l is the width value of crest, and max (h) is the maximum h value in alpha cross section；

3. alpha cross section coefficient of variation R₃It is defined as:

$R_3=\underset{N\&RightArrow;\mathrm{\&infin;}}{\lim }\frac{1}{N}\underset{n=1}{\overset{N}{\&Sigma;}}count({\mathrm{\&beta;}}_n\&GreaterEqual;{\mathrm{\&beta;}}_{th}*{\mathrm{\&beta;}}_{max})$

Wherein count () represents that statistics Spectral correlation function alpha cross section waviness is more than β_th*β_maxCrest number, β_nFor the waviness of cross section crest, β_maxFor the maximum waviness of cross section crest, β_thFor waviness threshold value, take definite value 0.1；

Its medium wave peak waviness β is defined as:

$\mathrm{\&beta;}=\frac{h}{l}$

Wherein the implication of h and l is with described in 2.；

4. f cross section normalized area R₄It is defined as:

$R_4=\underset{N\&RightArrow;\mathrm{\&infin;}}{\lim }\frac{1}{N}\underset{n=1}{\overset{N}{\&Sigma;}}\left\{\frac{1}{max\left(S_{xn}^0\right(f\left)\right)}{\&Integral;}_{-f_0}^{f_0}\right|S_{xn}^0\left(f\right)\left|df\right\}$

WhereinRepresent the maximum asking for Spectral correlation function f cross-section function,Represent the area asking for f cross section；

5. the significance of alpha cross section spectral line compares R₅It is defined as:

$R_5=\underset{N\&RightArrow;\mathrm{\&infin;}}{\lim }\frac{1}{N}\underset{n=1}{\overset{N}{\&Sigma;}}\frac{\sec \left({\mathrm{\&rho;}}_n\right)}{max\left({\mathrm{\&rho;}}_n\right)}$

Wherein max (ρ_n) for the maximum notable angle value of alpha cross section spectral line, sec (ρ_n) for time big notable angle value of alpha cross section spectral line；

The standard deviation R of the statistical nature zero center normalization instantaneous amplitude absolute value based on signal time domain in described step (4)₆It is defined as:

$\mathrm{\&delta;}=\sqrt{\frac{1}{N}\&lsqb;\underset{i=1}{\overset{N}{\&Sigma;}}{A}_{cn}^2\left(i\right)\&rsqb;-{\&lsqb;\frac{1}{N}\underset{i=1}{\overset{N}{\&Sigma;}}\left|A_{cn}\left(i\right)\right|\&rsqb;}^2}$

Wherein N is observation sample number, A_cnI () is zero center normalization instantaneous amplitude, be defined as A_cn(i)=A_n(i)-1；

WhereinDivisorAnd A (i) is sample point instantaneous amplitude；

Each debugging in described step (4) identifies that the threshold value value of feature is as follows:

Feature R₁Threshold value: r₁₁=0.67, r₁₂=0.36；Feature R₂Threshold value: r₂₁=4.8, r₂₂=3.2；Feature R₃Threshold value: r₃₁=60.46；；Feature R₄Threshold value: r₄₁=903.51, r₄₂=498.54；Feature R₅Threshold value: r₅₁=0.073, r₅₂=7.95 × 10^-4；Feature R₆Threshold value: r₆₁=0.29.

Principles of the invention is: Cyclic Spectrum correlation theory is a kind of Non-stationary Signal Analysis method, it is adaptable to have the analysis of the communication modulation signal of cyclo-stationary.Export in result by the limited random fluctuation phenomenon caused of external interference and sample points for former spectrum correlation estimation algorithm, introduce Wavelet Denoising Method and superposition is asked for average link and spectrum related operation process is optimized, effectively reduce the random fluctuation of output result, and make spectrum signature significantly strengthen.On this basis, it is possible to obtain 5 kinds of identification features based on Cyclic Spectrum of modulation signal, by conjunction with a kind of statistical nature based on signal time domain, multiple digital modulation signals being identified.

The present invention is in that with the advantage of prior art: compared with traditional power spectrumanalysis, the advantage that the present invention inherits cyclic-spectral Analysis, can fully excavate the modulation signature of signal of communication, Modulation Identification process is made to depend on the signal Time-domain Statistics feature that noise susceptibility is stronger not too much, thus accuracy of identification when improving low signal-to-noise ratio；Compared with former spectrum correlation estimation algorithm, present invention introduces Wavelet Denoising Method and superposition is asked for average link and spectrum related operation process is optimized, thus weakening in output result by the limited random fluctuation caused of external interference and sample points, make spectrum signature be enhanced, be beneficial to the extraction of Modulation Identification feature；5 Modulation Identification features based on Spectral correlation function that the present invention proposes are simple and clear compared with the identification characteristic formp based on signal Time-domain Statistics, and noise susceptibility is more weak.

Accompanying drawing explanation

Fig. 1 is the Modulation Identification method flow diagram relevant based on Cyclic Spectrum of the present invention；

Fig. 2 is the Modulation Identification flow chart after spectrum related operation.

Detailed description of the invention

The present invention is discussed in detail below in conjunction with the drawings and the specific embodiments.

As it is shown in figure 1, the commonly used digital Modulation Signals Recognition method relevant based on Cyclic Spectrum of the present invention to be embodied as step as follows:

(1) asking for average link owing to the present invention introduces to add up in former spectrum correlation estimation algorithm, need to carry out equal portions after sampled for primary signal, sample frequency is set toT_SBeing time-domain sampling interval, every equal portions signal generally takes 1024 sample points；

(2) Cyclic Spectrum correlation theory is a kind of theoretical tool for Non-stationary Signal Analysis proposed by Gardner, and it is defined as based on the algorithm for estimating of frequency domain smoothing:

$S_{x\mathrm{\&Delta;}t}^{\mathrm{\&alpha;}}{(t,f)}_{\mathrm{\&Delta;}f}=\frac{1}{M}\underset{v=-(M-1)/2}{\overset{(M-1)/2}{\&Sigma;}}\frac{1}{\mathrm{\&Delta;}t}{X}_{\mathrm{\&Delta;}t}(t,f+\frac{\mathrm{\&alpha;}}{2}+{\mathrm{vF}}_S)X_{\mathrm{\&Delta;}t}^{*}(t,f-\frac{\mathrm{\&alpha;}}{2}+{\mathrm{vF}}_S)---\left(1\right)$

X in above formula_kOne of t n equal portions that () is institute's intercepted signal；Δ t is x_k(t) persistent period；Δ f is frequency domain smoothing interval；F_SFor frequency domain smallest incremental unit；T_SIt it is time-domain sampling interval；M is smoothing windows spread factor, and taking definite value is 63；α is cycle frequency；F is frequency；T is the time；X_Δt(t, f) for equal portions signal X in step (1)_ΔtT the Short Time Fourier Transform of (), is defined as:

$X_{\mathrm{\&Delta;}t}(t,f)=\underset{K=0}{\overset{N-1}{\&Sigma;}}{a}_{\mathrm{\&Delta;}t}\left({\mathrm{KT}}_S\right)x_k(t-{\mathrm{KT}}_S)\exp (-j2\mathrm{\&pi;}f(t-{\mathrm{KT}}_S\left)\right)---\left(2\right)$

A in above formula_ΔtIt is window function, is taken as the rectangular window function isometric with equal portions signal herein；Counting of Short Time Fourier Transform is taken as N=1024；

(3) due to external interference and the limited impact of sample points, spectrum correlation calculation result fluctuation in step (2) is bigger, it is unfavorable for observation and the extraction of weak feature, therefore first the cycle frequency α of the Spectral correlation function of above-mentioned n part equal portions signal is carried out Wavelet Denoising Method, it is added again and asks on average, be shown below:

$S_{x\mathrm{\&Delta;}t}^{\mathrm{\&alpha;}}\left(f\right)=\frac{1}{N}\underset{k=1}{\overset{N}{\&Sigma;}}{\mathrm{WDN}}_{\mathrm{\&alpha;}}\&lsqb;S_{x\mathrm{\&Delta;}t}^{\mathrm{\&alpha;}}(t_k,f)\&rsqb;$

(3)

WDN in formula_α() represents rightIn cycle frequency α carry out Wavelet Denoising Method, as follows:

Y=wden (x, ' heursure', ' s', ' sln', 5, ' sym8') (4)

Wherein x and y is the discrete series before and after denoising, whole formula represents that signals and associated noises sequence is decomposed by use sym8 small echo, on the layer 5 decomposed, use soft sure thresholding system of selection that sequence carries out denoising, and this thresholding adjusts with the noise variance of ground floor wavelet decomposition.

(4) by analyze modulation signal Spectral correlation function it appeared that, spectrum signature is concentrated mainly on alpha cross section and the f cross section of the relevant figure of spectrum, therefore can being combined into debugging and identify that modulation signal is identified by feature by extracting the spectrum signature in the two cross section, proposed feature is Spectral correlation function alpha cross section and f cross section maximum value compares R₁, alpha cross section intense line number R₂, alpha cross section coefficient of variation R₃, f cross section normalized area R₄, alpha cross section spectral line significance compare R₅；It addition, this Modulation Identification process further relates to the standard deviation R of a kind of signal Time-domain Statistics feature zero center normalization instantaneous amplitude absolute value₆；For accurate description Modulation Identification feature, introducing the definition of cross section crest waviness β and significance ρ, the two definition is respectively used to describe background shake and the intensity of spectral line of alpha cross section,

Its medium wave peak waviness β is defined as:

$\mathrm{\&beta;}=\frac{h}{l}---\left(5\right)$

Wherein h is the relevant figure alpha cross section medium wave peak amplitude of spectrum and the difference of the greater in adjacent two trough amplitudes, and l is the width value of crest；

Crest significance ρ is defined as:

$\mathrm{\&rho;}=\frac{h^2}{l*max\left(h\right)}---\left(6\right)$

Wherein max (h) is the maximum h value in alpha cross section；On this basis each Modulation Identification feature is described as follows:

1. Spectral correlation function alpha cross section and f cross section maximum value compare R₁It is defined as:

$R_1=\underset{N\&RightArrow;\mathrm{\&infin;}}{\lim }\frac{1}{N}\underset{n=1}{\overset{N}{\&Sigma;}}\frac{\max \left(\right|S_{xn}^{\mathrm{\&alpha;}}\left(0\right)\left|\right)}{\max \left(\right|S_{xn}^0\left(f\right)\left|\right)}---\left(7\right)$

WhereinWithRespectively Spectral correlation functionAlpha cross section and f cross-section function；The threshold value of this feature is taken as r₁₁=0.67, r₁₂=0.36；

2. alpha cross section intense line number R₂It is defined as:

$R_2=\underset{N\&RightArrow;\mathrm{\&infin;}}{\lim }\frac{1}{N}\underset{n=1}{\overset{N}{\&Sigma;}}count({\mathrm{\&rho;}}_n\&GreaterEqual;{\mathrm{\&rho;}}_{th}*{\mathrm{\&rho;}}_{max})$

Wherein count () represents the notable peak value seeking the relevant figure alpha cross section of spectrum, ρ_nFor the significance of cross section crest, ρ_maxFor the maximum significance of cross section crest, ρ_thFor significance threshold value, take definite value 0.27；The threshold value of this feature is taken as r₂₁=4.8, r₂₂=3.2；

3. alpha cross section coefficient of variation R₃It is defined as:

$R_3=\underset{N\&RightArrow;\mathrm{\&infin;}}{\lim }\frac{1}{N}\underset{n=1}{\overset{N}{\&Sigma;}}count({\mathrm{\&beta;}}_n\&GreaterEqual;{\mathrm{\&beta;}}_{th}*{\mathrm{\&beta;}}_{max})$

Wherein count () represents that statistics Spectral correlation function alpha cross section waviness is more than β_th*β_maxCrest number, β_nFor the waviness of cross section crest, β_maxFor the maximum waviness of cross section crest, β_thFor waviness threshold value, take definite value 0.1；The threshold value of this feature is taken as r₃₁=60.46；

4. f cross section normalized area R₄It is defined as:

$R_4=\underset{N\&RightArrow;\mathrm{\&infin;}}{\lim }\frac{1}{N}\underset{n=1}{\overset{N}{\&Sigma;}}\left\{\frac{1}{max\left(S_{xn}^0\right(f\left)\right)}{\&Integral;}_{-f_0}^{f_0}\right|S_{xn}^0\left(f\right)\left|df\right\}$

WhereinRepresent the maximum asking for Spectral correlation function f cross-section function,Represent the area asking for f cross section；The threshold value of this feature is taken as r₄₁=903.51, r₄₂=498.54；

5. the significance of alpha cross section spectral line compares R₅It is defined as:

$R_5=\underset{N\&RightArrow;\mathrm{\&infin;}}{\lim }\frac{1}{N}\underset{n=1}{\overset{N}{\&Sigma;}}\frac{\sec \left({\mathrm{\&rho;}}_n\right)}{max\left({\mathrm{\&rho;}}_n\right)}$

Wherein max (ρ_n) for the maximum notable angle value of alpha cross section spectral line, sec (ρ_n) for time big notable angle value of alpha cross section spectral line；The threshold value of this feature is taken as r₅₁=0.073, r₅₂=7.95 × 10^-4

6. the standard deviation R of zero center normalization instantaneous amplitude absolute value₆It is defined as:

$\mathrm{\&delta;}=\sqrt{\frac{1}{N}\&lsqb;\underset{i=1}{\overset{N}{\&Sigma;}}{A}_{cn}^2\left(i\right)\&rsqb;-{\&lsqb;\frac{1}{N}\underset{i=1}{\overset{N}{\&Sigma;}}\left|A_{cn}\left(i\right)\right|\&rsqb;}^2}$

Wherein N is observation sample number, A_cnI () is zero center normalization instantaneous amplitude, be defined as A_cn(i)=A_n(i)-1；WhereinDivisorAnd A (i) is sample point instantaneous amplitude；The threshold value of this feature is taken as r₆₁=0.29.

In conjunction with above-mentioned 6 kinds of Modulation Identification features, several digital modulation signals are identified, concrete Modulation Identification flow process as in figure 2 it is shown,

IfThen judge that the modulation system of this signal is as MSK；

If $\{{\hat{R}}_1>r_{12},{\hat{R}}_2>r_{21}\}$ Or $\{{\hat{R}}_1<r_{12},{\hat{R}}_3<r_{31}\},$ Then judge that the modulation system of this signal is as 4FSK；

IfThen judge that the modulation system of this signal is as 2FSK；

IfThen judge that the modulation system of this signal is as BPSK；

If $\{{\hat{R}}_1>r_{11},{\hat{R}}_2<r_{22},{\hat{R}}_5<r_{51},{\hat{R}}_6<r_{61}\},$ Then judge that the modulation system of this signal is as 2ASK；

If $\{{\hat{R}}_1>r_{11},{\hat{R}}_2<r_{22},{\hat{R}}_5<r_{51},{\hat{R}}_6>r_{61}\},$ Then judge that the modulation system of this signal is as 4ASK；

IfThen judge that the modulation system of this signal is as 2FSK*；

IfThen judge that the modulation system of this signal is as 4FSK*；

If $\{{\hat{R}}_1<r_{12},{\hat{R}}_3>r_{31},{\hat{R}}_4<r_{42},{\hat{R}}_5>r_{52}\},$ Then judge the modulation system QPSK of this signal；

If $\{{\hat{R}}_1<r_{12},{\hat{R}}_3>r_{31},{\hat{R}}_4<r_{42},{\hat{R}}_5<r_{52}\},$ Then judge the modulation system 8PSK of this signal；

WhereinFor the estimated value of feature R in Modulation Identification process.

According to such criterion of identification, when signal to noise ratio is 10dB, the discrimination of each digital modulation mode is all more than 88.8%, recognition effect is more satisfactory, identified modulation system includes 2ASK, 4ASK, BPSK, QPSK, 8PSK, MSK, 2FSK, 4FSK, 2FSK*, 4FSK*, and wherein FSK* represents the incoherent frequency shift keying of code element initial phase.

The content not being described in detail in description of the present invention belongs to the known prior art of professional and technical personnel in the field.

Although disclosing preferred example and the accompanying drawing of the present invention for the purpose of illustration, but it will be appreciated by those skilled in the art that: without departing from the spirit and scope of the invention and the appended claims, various replacements, to change and modifications be all possible.Therefore, the technical scheme that the present invention protects should not be limited to preferred example and accompanying drawing disclosure of that.

Claims (5)

1. the commonly used digital Modulation Signals Recognition method being correlated with based on Cyclic Spectrum, it is characterised in that: include following step:

(1) after sampled for the signal to noise ratio modulation signal of intercepting and capturing, it is divided into n equal portions, as the input signal carrying out spectrum related operation；

(2) choose suitable Fourier transform to count and smoothing windows width, respectively this n part signal is made by the spectrum correlation estimation algorithm based on frequency domain smoothing spectrum related operation；

The described spectrum correlation estimation algorithm based on frequency domain smoothing is as follows:

$S_{x\mathrm{\&Delta;}t}^{\mathrm{\&alpha;}}{(t,f)}_{\mathrm{\&Delta;}f}=\frac{1}{M}\underset{v=-(M-1)/2}{\overset{(M-1)/2}{\&Sigma;}}\frac{1}{\mathrm{\&Delta;}t}{X}_{\mathrm{\&Delta;}t}(t,f+\frac{\mathrm{\&alpha;}}{2}+{\mathrm{vF}}_S)X_{\mathrm{\&Delta;}t}^{*}(t,f-\frac{\mathrm{\&alpha;}}{2}+{\mathrm{vF}}_S)$

$X_{\mathrm{\&Delta;}t}(t,f)=\underset{K=0}{\overset{N-1}{\&Sigma;}}{a}_{\mathrm{\&Delta;}t}\left({\mathrm{KT}}_S\right)x_k(t-{\mathrm{KT}}_S)\exp (-j2\mathrm{\&pi;}f(t-{\mathrm{KT}}_S\left)\right)$

Wherein,For the result after spectrum related operation；X_Δt(t, f) for signal x_kThe result of (t) short time discrete Fourier transform；Signal x_kOne of t n equal portions that () is institute's intercepted signal；Δ t is x_k(t) persistent period；A_ΔtIt it is window function；Δ f is frequency domain smoothing interval；F_SFor frequency domain smallest incremental unit；T_SIt it is time-domain sampling interval；N is signal x_k(t) sample number；M is smoothing windows spread factor；α is cycle frequency；F is frequency；T is the time；

(3) operation result in step (2) is added after alpha cross section Wavelet Denoising Method and averages, the output result of the spectrum correlation estimation algorithm being improved；

The spectrum correlation estimation algorithm of described improvement is as follows:

$S_{x\mathrm{\&Delta;}t}^{\mathrm{\&alpha;}}\left(f\right)=\frac{1}{N}\underset{k=1}{\overset{N}{\&Sigma;}}{\mathrm{WDN}}_{\mathrm{\&alpha;}}\&lsqb;S_{x\mathrm{\&Delta;}t}^{\mathrm{\&alpha;}}(t_k,f)\&rsqb;$

Wherein, WDN_α() represents rightIn cycle frequency α carry out Wavelet Denoising Method；

(4) extract 5 kinds of Modulation Identification features according to the alpha cross section of step (3) gained Spectral correlation function and f cross section, and in conjunction with a kind of signal characteristic structure sorting technique based on Time-domain Statistics, several digital modulation signals are identified；Described 5 kinds of Modulation Identification are characterized as: Spectral correlation function alpha cross section and f cross section maximum value ratio (are designated as R₁), alpha cross section intense line number (be designated as R₂), alpha cross section coefficient of variation (be designated as R₃), f cross section normalized area (be designated as R₄), the significance ratio of alpha cross section spectral line (be designated as R₅)；Described a kind of statistical nature based on signal time domain is: the standard deviation of zero center normalization instantaneous amplitude absolute value (is designated as R₆)；Described several digital modulation signals includes 2ASK, 4ASK, BPSK, QPSK, 8PSK, MSK, 2FSK, 4FSK, 2FSK*, 4FSK*, and wherein FSK* represents the incoherent frequency shift keying of code element initial phase；

The described identification process to several digital modulation signals is:

IfThen judge that the modulation system of this signal is as MSK；

If $\left\{\begin{array}{cc}{\hat{R}}_1>r_{12},& {\hat{R}}_2>r_{21}\end{array}\right\}$ Or $\left\{\begin{array}{cc}{\hat{R}}_1<r_{12},& {\hat{R}}_3<r_{31}\end{array}\right\},$ Then judge that the modulation methods of this signal is as 4FSK；

IfThen judge that the modulation system of this signal is as 2FSK；

IfThen judge that the modulation system of this signal is as BPSK；

If $\left\{\begin{array}{cccc}{\hat{R}}_1>r_{11},& {\hat{R}}_2<r_{22},& {\hat{R}}_5<r_{51},& {\hat{R}}_6<r_{61}\end{array}\right\},$ Then judge that the modulation system of this signal is as 2ASK；

If $\left\{\begin{array}{cccc}{\hat{R}}_1>r_{11},& {\hat{R}}_2<r_{22},& {\hat{R}}_5<r_{51},& {\hat{R}}_6>r_{61}\end{array}\right\},$ Then judge that the modulation system of this signal is as 4ASK；

IfThen judge the modulation system 2FSK* of this signal；

IfThen judge that the modulation system of this signal is as 4FSK*；

If $\left\{\begin{array}{cccc}{\hat{R}}_1<r_{12},& {\hat{R}}_3>r_{31},& {\hat{R}}_4<r_{42},& {\hat{R}}_5>r_{52}\end{array}\right\},$ Then judge the modulation system QPSK of this signal；

If $\left\{\begin{array}{cccc}{\hat{R}}_1<r_{12},& {\hat{R}}_3>r_{31},& {\hat{R}}_4<r_{42},& {\hat{R}}_5<r_{52}\end{array}\right\},$ Then judge the modulation system 8PSK of this signal.

2. a kind of commonly used digital Modulation Signals Recognition method relevant based on Cyclic Spectrum according to claim 1, it is characterised in that: in described step (2), the sampling number of every part of signal is taken as N=1024, and smoothing windows spread factor is M=63.

3. a kind of commonly used digital Modulation Signals Recognition method relevant based on Cyclic Spectrum according to claim 1, it is characterized in that: in described step (3), the principle of Wavelet Denoising Method is: use sym8 small echo that noisy sequence is decomposed, on the layer 5 decomposed, use soft sure thresholding system of selection that sequence carries out denoising, and this thresholding adjusts with the noise variance of ground floor wavelet decomposition.

4. a kind of commonly used digital Modulation Signals Recognition method relevant based on Cyclic Spectrum according to claim 1, it is characterised in that: described 5 kinds in step (4) are as follows based on the definition of the Modulation Identification feature of modulation signal spectrum correlation function:

1. Spectral correlation function alpha cross section and f cross section maximum value compare R₁It is defined as:

$R_1=\underset{N\&RightArrow;\mathrm{\&infin;}}{\lim }\frac{1}{N}\underset{n=1}{\overset{N}{\&Sigma;}}\frac{\max \left(\right|S_{xn}^{\mathrm{\&alpha;}}\left(0\right)\left|\right)}{\max \left(\right|S_{xn}^0\left(f\right)\left|\right)}$

WhereinWithRespectively Spectral correlation functionAlpha cross section and f cross-section function；

2. alpha cross section intense line number R₂It is defined as:

$R_2=\underset{N\&RightArrow;\mathrm{\&infin;}}{lim}\frac{1}{N}\underset{n=1}{\overset{N}{\&Sigma;}}count({\mathrm{\&rho;}}_n\&GreaterEqual;{\mathrm{\&rho;}}_{th}*{\mathrm{\&rho;}}_{max})$

Wherein count () represents the notable peak value seeking the relevant figure alpha cross section of spectrum, ρ_nFor the significance of cross section crest, ρ_maxFor the maximum significance of cross section crest, ρ_thFor significance threshold value, take definite value 0.27；

Its medium wave peak significance ρ is defined as:

$\mathrm{\&rho;}=\frac{h^2}{l*max\left(h\right)}$

Wherein h is the relevant figure alpha cross section medium wave peak amplitude of spectrum and the difference of the greater in adjacent two trough amplitudes, and l is the width value of crest, and max (h) is the maximum h value in alpha cross section；

3. alpha cross section coefficient of variation R₃It is defined as:

$R_3=\underset{N\&RightArrow;\mathrm{\&infin;}}{lim}\frac{1}{N}\underset{n=1}{\overset{N}{\&Sigma;}}count({\mathrm{\&beta;}}_n\&GreaterEqual;{\mathrm{\&beta;}}_{th}*{\mathrm{\&beta;}}_{max})$

Wherein count () represents that statistics Spectral correlation function alpha cross section waviness is more than β_th*β_maxCrest number, β_nFor the waviness of cross section crest, β_maxFor the maximum waviness of cross section crest, β_thFor waviness threshold value, take definite value 0.1；

Its medium wave peak waviness β is defined as:

$\mathrm{\&beta;}=\frac{h}{l}$

Wherein the implication of h and l is with described in 2.；

4. f cross section normalized area R₄It is defined as:

$R_4=\underset{N\&RightArrow;\mathrm{\&infin;}}{lim}\frac{1}{N}\underset{n=1}{\overset{N}{\&Sigma;}}\left\{\frac{1}{max\left(S_{xn}^0\right(f))}{\&Integral;}_{-f_0}^{f_0}\right|S_{xn}^0\left(f\right)\left|df\right\}$

WhereinRepresent the maximum asking for Spectral correlation function f cross-section function,Represent the area asking for f cross section；

5. the significance of alpha cross section spectral line compares R₅It is defined as:

$R_5=\underset{N\&RightArrow;\mathrm{\&infin;}}{lim}\frac{1}{N}\underset{n=1}{\overset{N}{\&Sigma;}}\frac{\sec \left({\mathrm{\&rho;}}_n\right)}{max\left({\mathrm{\&rho;}}_n\right)}$

Wherein max (ρ_n) for the maximum notable angle value of alpha cross section spectral line, sec (ρ_n) for time big notable angle value of alpha cross section spectral line；

The standard deviation R of the statistical nature zero center normalization instantaneous amplitude absolute value based on signal time domain in described step (4)₆It is defined as:

$\mathrm{\&delta;}=\sqrt{\frac{1}{N}\&lsqb;\underset{i=1}{\overset{N}{\&Sigma;}}{A}_{cn}^2\left(i\right)\&rsqb;-{\&lsqb;\frac{1}{N}\underset{i=1}{\overset{N}{\&Sigma;}}\left|A_{cn}\left(i\right)\right|\&rsqb;}^2}$

Wherein N is observation sample number, A_cnI () is zero center normalization instantaneous amplitude, be defined as A_cn(i)=A_n(i)-1；WhereinDivisorAnd A (i) is sample point instantaneous amplitude.

5. a kind of commonly used digital Modulation Signals Recognition method relevant based on Cyclic Spectrum according to claim 1, it is characterised in that: each debugging in described step (4) identifies that the threshold value value of feature is as follows:

Feature R₁Threshold value: r₁₁=0.67, r₁₂=0.36；Feature R₂Threshold value: r₂₁=4.8, r₂₂=3.2；Feature R₃Threshold value: r₃₁=60.46；；Feature R₄Threshold value: r₄₁=903.51, r₄₂=498.54；Feature R₅Threshold value: r₅₁=0.073, r₅₂=7.95 × 10^-4；Feature R₆Threshold value: r₆₁=0.29.