CN102065056A - Method for realizing MQAM (Multiple Quadrature Amplitude Modulation) signal modulation mode identification of any constellation diagram on basis of clustering - Google Patents

Abstract

The invention discloses a method for realizing MQAM (Multiple Quadrature Amplitude Modulation) signal modulation mode identification of any constellation diagram on the basis of clustering. For an MQAM signal, a modulation order can be big theoretically, the modulated constellation diagram can have various types, thus the possible range of the modulation mode of a signal to be identified is hard to determine, and people may face unknown modulation modes. The method puts forward a two-step clustering algorithm to reconstruct the constellation diagram, and the modulation mode identification is finished on the basis of the reconstructed constellation diagram; the self-adaption subtractive clustering based on a signal to noise ratio gives an initial clustering result, and the primary constellation diagram reconstruction is finished; on the basis, fuzzy C mean value clustering is adopted for finishing the constellation point reconstruction of the final modulation signal; after constellation point reconstruction is finished, the modulation order of the signal is judged according to the number of the constellation points; and for a modulation mode set of different orders, the modulation mode identification is carried out by a generalized likelihood ratio. The method disclosed by the invention can realize the MQAM signal modulation mode identification of any constellation diagram by the reconstructed constellation diagram.

Description

Realize the method for the MQAM signal Modulation Mode Recognition of any planisphere based on cluster

Technical field

The present invention is a kind of Digital Communication Signal Modulation recognition methods, belongs to signal of communication and handles and the detection technique field.

Background technology

The automatic identification of modulation system is a technology between input and signal demodulation, and main task is that the intelligence that realizes modulation signal receives, handles.Aspect civilian, government is in order to implement effective radio spectrum management, usually needs the monitor civil signals transmission, so that to their retentive controls or discovery with monitor the transmitter of unregistered registration.Aspect military and national security, the modulation automatic identification technology is used more extensive.For obtaining communication information, at first to judge the modulation system of signal, could implement correct demodulation and information processing subsequently and analysis afterwards; In electronic warfare,, all need to understand fully the parameter and the character of related communication or electronic signal by Modulation identification technology for implementing electronic countermeasures, electronic counter-countermea-sures, threat detection, alarm, target acquisition and search etc.Except that multicarrier modulated signal MFSK, remaining modulation system MPSK, MASK and MQAM signal can be obtained the baseband complex signal of sign synchronization by all purpose communication signal processing algorithm, and we can utilize patent ground method of the present invention to finish the identification of modulation system to utilize this baseband complex signal.

Summary of the invention

The invention provides a kind of method that realizes the MQAM signal Modulation Mode Recognition of any planisphere based on cluster.

The object of the present invention is achieved like this:

For the MQAM signal, order of modulation can be very big in theory, and the pattern of modulation constellation can have multiple arbitrarily, so Modulation Identification is very difficult; Modulation constellation has been represented different modulation systems, and the present invention proposes the reconstruct that two step clustering algorithms are realized planisphere, finishes Modulation Mode Recognition based on the reconstruct planisphere then, and concrete steps are:

Step 1: Signal Pretreatment: mainly be by the universal demodulation algorithm, from intermediate-freuqncy signal, demodulate the complex baseband signal of sign synchronization;

Step 2: use the self adaptation subtractive clustering based on signal-to-noise ratio (SNR) estimation, this is a first step clustering algorithm, and main purpose is initial number and the initial position that provides cluster centre,

The complex baseband signal r of sign synchronization _kFor

$r_k=R_k{e}^{j{\mathrm{\θ}}_k}+n_k,$ k＝1，2，…，N. (1)

Wherein, wherein The modulation constellation points of representation signal, n _kRepresented Gaussian noise, N represents sample sequence length; Signal to noise ratio is SNR, and this value obtains by blind signal-to-noise ratio (SNR) estimation; By SNR, calculate average noise power P _n, and then calculate " density " index D of each sampled point _k:

$D_k=\underset{j=1}{\overset{N}{\mathrm{\Σ}}}\exp (-\frac{{\left|\right|r_k-r_j\left|\right|}^2}{K_a*P_n}).---\left(2\right)$

Wherein, K _aBe one and adjust coefficient.Result according to formula (2) selects D _kMaximum D _{C, 1}Corresponding sampling points r _{C, 1}Be first cluster centre; Then, to top " density " index correction;

$D_k=D_k-D_{c,1}\exp (-\frac{{\left|\right|r_k-r_{c,1}\left|\right|}^2}{K_b*P_n}).---\left(3\right)$

Wherein, K _bBe one and adjust coefficient.Use the same method and select second cluster centre r _{C, 2}

From the 3rd cluster centre r _{C, 3}New cluster centre r is selected in beginning at every turn _{C, l+1}After, judge whether cluster centre is all selected:

min{‖r _c，k-r _c，l+1‖ ²}＞K _c*P _n，k＝1，2，…，l. (4)

Wherein, K _bBe one and adjust coefficient,, then proceed subtractive clustering, otherwise finish if formula (4) is set up;

Step 3: use C means clustering algorithm or fuzzy C-means clustering algorithm reconstruct planisphere.This is the second step clustering algorithm, and main purpose is based on the reconstruct that first step cluster result is finished planisphere;

R={r _k, k=1,2 ..., N} is the sample set that N baseband complex signal sample formed, C is predetermined classification number, m _i, i=1,2 ... C is the center of each cluster, μ _i(r _k) be the membership function of k sample for i cluster;

According to known initial cluster center, calculate the degree of membership size of sample to all initial clustering points, membership function is defined as follows

${\mathrm{\μ}}_i\left(r_k\right)=\frac{{(1/{\left|\right|r_k-m_i\left|\right|}^2)}^{1/(b-1)}}{\underset{l=1}{\overset{C}{\mathrm{\Σ}}}{(1/{\left|\right|r_k-m_l\left|\right|}^2)}^{1/(b-1)}},$ k＝1，2，…，N，i＝1，2，…，C. (5)

After degree of membership is calculated and finished, calculate the value of the cluster centre under this degree of membership

$m_i=\frac{\underset{k=1}{\overset{N}{\mathrm{\Σ}}}{\left[{\mathrm{\μ}}_i\left(r_k\right)\right]}^b{r}_k}{\underset{k=1}{\overset{N}{\mathrm{\Σ}}}{\left[{\mathrm{\μ}}_i\left(r_k\right)\right]}^b},$ i＝1，2，…，C. (6)

After cluster centre and degree of membership have been arranged, calculate cluster cost function with the membership function definition

$J_f=\underset{i=1}{\overset{C}{\mathrm{\Σ}}}\underset{k=1}{\overset{N}{\mathrm{\Σ}}}{\left[{\mathrm{\μ}}_i\left(r_k\right)\right]}^b{\left|\right|r_k-m_i\left|\right|}^2.---\left(7\right)$

Wherein, b＞1 is the constant that can control the fog-level of cluster result; Iterate by top formula (5), (6), (7), make cluster centre converge to the point of cluster cost function minimum gradually, this moment, the degree of membership value tended towards stability, and the termination condition of iteration is

J _f(k+1)≈J _f(k). (8)

Can finish star map reconstruction by above two clustering algorithms.

Step 4: use Generalized Likelihood Ratio test (GLRT:General LikelihoodRate Test), finish the identification of modulation system:, judge the modulation system exponent number at first according to the constellation point number that reconstructs based on the reconstruct planisphere; If it is the same in the modulation system set to be identified the exponent number of two kinds of modulation systems being arranged, continue to use the GLRT classification so;

By reconstruct constellation point m _iAnd modulation constellation points

Between Euclidean distance nearest, estimate for just the maximum likelihood of modulation constellation point Subscript c has represented different modulation types; Use GLRT classification same order modulation type then.Expression formula is as follows

Wherein,

The order of modulation that expression estimates,

Expression makes l _{GLR, m}Obtain the modulation system c of minimum value; Through type (10) has been finished final Modulation Mode Recognition.

The benefit of this method is to finish the identification of modulation system of the MQAM signal of any planisphere.The order of modulation of MQAM signal can be from 4 rank to 1024 rank (even higher), and modulation constellation also can arrange to change according to both sides, traditional Modulation Identification method, and the MQAM range of signal that can discern is very limited.The method of using the present invention to propose is finished Modulation Mode Recognition and is had: advantages such as MQAM signal constellation which scope is unrestricted, discrimination is high, anti-Gaussian noise ability is strong.

Description of drawings

Fig. 1 is the digital modulation signals Modulation Mode Recognition method handling process that the present invention is based on star map reconstruction;

Fig. 2 is to use the present invention to carry out the design sketch of 16QAM signal constellation which reconstruct;

Fig. 3 is to use the present invention to carry out the V.29/16QAM design sketch of signal constellation which reconstruct;

Embodiment

Below in conjunction with accompanying drawing, the present invention is further illustrated:

For the MQAM signal, order of modulation can be very big in theory, and the pattern of modulation constellation can have multiple arbitrarily, so Modulation Identification is very difficult; Modulation constellation has been represented different modulation systems, and the present invention proposes the reconstruct that two step clustering algorithms are realized planisphere, finishes Modulation Mode Recognition based on the reconstruct planisphere then, and concrete steps are:

Step 1: Signal Pretreatment: mainly be by the universal demodulation algorithm, from intermediate-freuqncy signal, demodulate the complex baseband signal of sign synchronization;

Step 2: use the self adaptation subtractive clustering based on signal-to-noise ratio (SNR) estimation, this is a first step clustering algorithm, and main purpose is initial number and the initial position that provides cluster centre,

The complex baseband signal r of sign synchronization _kFor

$r_k=R_k{e}^{j{\mathrm{\θ}}_k}+n_k,$ k＝1，2，…，N. (11)

Wherein, wherein

The modulation constellation points of representation signal, n _kRepresented Gaussian noise, N represents sample sequence length; Signal to noise ratio is SNR, and this value obtains by blind signal-to-noise ratio (SNR) estimation; By SNR, calculate average noise power P _n, and then calculate " density " index D of each sampled point _k:

$D_k=\underset{j=1}{\overset{N}{\mathrm{\Σ}}}\exp (-\frac{{\left|\right|r_k-r_j\left|\right|}^2}{K_a*P_n}).---\left(12\right)$

Wherein, K _aBe one and adjust coefficient.Result according to formula (2) selects D _kMaximum D _{C, 1}Corresponding sampling points r _{C, 1}Be first cluster centre; Then, to top " density " index correction;

$D_k=D_k-D_{c,1}\exp (-\frac{{\left|\right|r_k-r_{c,1}\left|\right|}^2}{K_b*P_n}).---\left(13\right)$

Wherein, K _bBe one and adjust coefficient.Use the same method and select second cluster centre r _{C, 2}

From the 3rd cluster centre r _{C, 3}New cluster centre r is selected in beginning at every turn _{C, l+1}After, judge whether cluster centre is all selected:

min{‖r _c，k-r _c，l+1‖ ²}＞K _c*P _n，k＝1，2，…，l. (14)

Wherein, K _bBe one and adjust coefficient,, then proceed subtractive clustering, otherwise finish if formula (4) is set up;

Step 3: use C means clustering algorithm or fuzzy C-means clustering algorithm reconstruct planisphere.This is the second step clustering algorithm, and main purpose is based on the reconstruct that first step cluster result is finished planisphere;

R={r _k, k=1,2 ..., N} is the sample set that N baseband complex signal sample formed, C is predetermined classification number, m _i, i=1,2 ... C is the center of each cluster, μ _i( _rK) be the membership function of k sample for i cluster;

According to known initial cluster center, calculate the degree of membership size of sample to all initial clustering points, membership function is defined as follows

${\mathrm{\μ}}_i\left(r_k\right)=\frac{{(1/{\left|\right|r_k-m_i\left|\right|}^2)}^{1/(b-1)}}{\underset{l=1}{\overset{C}{\mathrm{\Σ}}}{(1/{\left|\right|r_k-m_l\left|\right|}^2)}^{1/(b-1)}},$ k＝1，2，…，N，i＝1，2，…，C. (15)

After degree of membership is calculated and finished, calculate the value of the cluster centre under this degree of membership

$m_i=\frac{\underset{k=1}{\overset{N}{\mathrm{\Σ}}}{\left[{\mathrm{\μ}}_i\left(r_k\right)\right]}^b{r}_k}{\underset{k=1}{\overset{N}{\mathrm{\Σ}}}{\left[{\mathrm{\μ}}_i\left(r_k\right)\right]}^b},i=\mathrm{1,2},...,C.---\left(16\right)$

After cluster centre and degree of membership have been arranged, calculate cluster cost function with the membership function definition

$J_f=\underset{i=1}{\overset{C}{\mathrm{\Σ}}}\underset{k=1}{\overset{N}{\mathrm{\Σ}}}{\left[{\mathrm{\μ}}_i\left(r_k\right)\right]}^b{\left|\right|r_k-m_i\left|\right|}^2.---\left(17\right)$

Wherein, b＞1 is the constant that can control the fog-level of cluster result; Iterate by top formula (5), (6), (7), make cluster centre converge to the point of cluster cost function minimum gradually, this moment, the degree of membership value tended towards stability, and the termination condition of iteration is

J _f(k+1)≈J _f(k). (18)

Can finish star map reconstruction by above two clustering algorithms.

Step 4: use Generalized Likelihood Ratio test (GLRT:General LikelihoodRate Test), finish the identification of modulation system:, judge the modulation system exponent number at first according to the constellation point number that reconstructs based on the reconstruct planisphere; If it is the same in the modulation system set to be identified the exponent number of two kinds of modulation systems being arranged, continue to use the GLRT classification so;

By reconstruct constellation point m _iAnd modulation constellation points

Between Euclidean distance nearest, estimate for just the maximum likelihood of modulation constellation point

Subscript c has represented different modulation types; Use GLRT classification same order modulation type then.Expression formula is as follows

Wherein,

The order of modulation that expression estimates, Expression makes l _{GLR, m}Obtain the modulation system c of minimum value; Through type (10) has been finished final Modulation Mode Recognition.

Fig. 1 is the digital modulation signals Modulation Mode Recognition method handling process that the present invention is based on star map reconstruction, use the universal demodulation algorithm to draw needed sign synchronization baseband signal from intermediate-freuqncy signal, use the self adaptation subtractive clustering to finish initial clustering then, provide the needed initial cluster center value of fuzzy C-means clustering.Then use the fuzzy C-means clustering algorithm, can reconstruct the very little signal modulation constellation of deviation.Utilize the reconstruct planisphere at last, adopt the GLRT grader, identify modulation system.

Illustrate, for the 16QAM of 5dB (square), V.29/16QAM signal carries out star map reconstruction respectively, this moment, the error rate of two modulation signals all was 16% to bit signal to noise ratio (Eb/No).Fig. 2 is to use the present invention to carry out the design sketch of 16QAM signal constellation which reconstruct.Fig. 3 is to use the present invention to carry out the V.29/16QAM design sketch of signal constellation which reconstruct.Can find out that from two figure received signal points is dispersed on the complex plane because of containing noise.Provided initial cluster center by the self adaptation subtractive clustering, made reconstruct planisphere and actual modulated constellation point very approaching by fuzzy C-means clustering.Therefore we can use the reconstruct planisphere to finish Modulation Identification.

The benefit of this method is to finish the identification of modulation system of the MQAM signal of any planisphere.The order of modulation of MQAM signal can be from 4 rank to 1024 rank (even higher), and modulation constellation also can arrange to change according to both sides, traditional Modulation Identification method, and the MQAM range of signal that can discern is very limited.The method of using the present invention to propose is finished Modulation Mode Recognition and is had: advantages such as MQAM signal constellation which scope is unrestricted, discrimination is high, anti-Gaussian noise ability is strong.

Above embodiment only is used to illustrate preferred implementation of the present invention; but the present invention is not limited to above-mentioned execution mode; in the ken that described field those of ordinary skill is possessed; any modification of being done within the spirit and principles in the present invention, be equal to and substitute and improvement etc., it all should be encompassed within the technical scheme scope that the present invention asks for protection.

Claims (1)

1. realize the method for the MQAM signal Modulation Mode Recognition of any planisphere based on cluster, it is characterized in that comprising the steps:

Step 1: Signal Pretreatment: mainly be by the universal demodulation algorithm, from intermediate-freuqncy signal, demodulate the complex baseband signal of sign synchronization;

Step 2: self adaptation subtractive clustering reconstruct planisphere, this is a first step clustering algorithm, main purpose is initial number and the initial position that provides cluster centre,

The complex baseband signal r of sign synchronization _kFor

$r_k=R_k{e}^{j{\mathrm{\θ}}_k}+n_k,$ k＝1，2，…，N. (1)

Wherein, wherein

The modulation constellation points of representation signal, n _kRepresented Gaussian noise, N represents sample sequence length; Signal to noise ratio is SNR, and this value obtains by blind signal-to-noise ratio (SNR) estimation; By SNR, calculate average noise power P _nAnd then calculate " density " index D of each sampled point _k

$D_k=\underset{j=1}{\overset{N}{\mathrm{\Σ}}}\exp (-\frac{{\left|\right|r_k-r_j\left|\right|}^2}{K_a*P_n}).---\left(2\right)$

Wherein, K _aBe one and adjust coefficient.Result according to formula (2) selects D _kMaximum D _{C, 1}Corresponding sampling points r _{C, 1}Be first cluster centre; Then, to top " density " index correction;

$D_k=D_k-D_{c,1}\exp (-\frac{{\left|\right|r_k-r_{c,1}\left|\right|}^2}{K_b*P_n}).---\left(3\right)$

Wherein, K _bBe one and adjust coefficient.Use the same method and select second cluster centre r _{C, 2}

From the 3rd cluster centre r _{C, 3}New cluster centre r is selected in beginning at every turn _{C, l+1}After, judge whether cluster centre is all selected;

min{‖r _c，k-r _c，l+1‖ ²}＞K _c*P _n，k＝1，2，…，l. (4)

Wherein, K _bBe one and adjust coefficient.If formula (4) is set up, then proceed subtractive clustering, otherwise finish;

Step 3, use fuzzy C-means clustering algorithm or C means clustering algorithm reconstruct planisphere, this is the second step clustering algorithm, main purpose is based on the reconstruct that first step cluster result is finished planisphere;

{ r _k, k=1,2 ..., N} is the sample set that N baseband complex signal sample formed,

Be the cluster centre number that subtractive clustering estimates, m _i, i=1,2, Be the center of each cluster, μ _i(r _k) be the membership function of k sample for i cluster;

According to known initial cluster center, calculate the degree of membership size of sample to all initial clustering points, membership function is defined as follows

${\mathrm{\μ}}_i\left(r_k\right)=\frac{{(1/{\left|\right|r_k-m_i\left|\right|}^2)}^{1/(b-1)}}{\underset{l=1}{\overset{\hat{C}}{\mathrm{\Σ}}}{(1/{\left|\right|r_k-m_l\left|\right|}^2)}^{1/(b-1)}},$ k＝1，2，…，N，i＝1，2，…， $\hat{C}.---\left(5\right)$

Degree of membership is calculated the cluster centre value under this degree of membership after calculating and finishing

$m_i=\frac{\underset{k=1}{\overset{N}{\mathrm{\Σ}}}{\left[{\mathrm{\μ}}_i\left(r_k\right)\right]}^b{r}_k}{\underset{k=1}{\overset{N}{\mathrm{\Σ}}}{\left[{\mathrm{\μ}}_i\left(r_k\right)\right]}^b},$ i＝1，2，…， $\hat{C}.---\left(6\right)$

After cluster centre and degree of membership have been arranged, calculate cluster cost function with the membership function definition

$J_f=\underset{i=1}{\overset{\hat{C}}{\mathrm{\Σ}}}\underset{k=1}{\overset{N}{\mathrm{\Σ}}}{\left[{\mathrm{\μ}}_i\left(r_k\right)\right]}^b{\left|\right|r_k-m_i\left|\right|}^2.---\left(7\right)$

Wherein, b＞1 is the constant that can control the fog-level of cluster result.Iterate by top formula (5) (6) (7), make cluster centre converge to the point of cluster cost function minimum gradually, this moment, the degree of membership value tended towards stability, and the termination condition of iteration is

J _f(k+1)≈J _f(k). (8)

Can finish star map reconstruction by above two clustering algorithms;

Step 4, use are finished Modulation Mode Recognition based on the Generalized Likelihood Ratio test (GLRT) of reconstruct planisphere:

At first, judge the modulation system exponent number according to the constellation point number that reconstructs.If it is the same in the modulation system set to be identified the exponent number of two kinds of modulation systems being arranged, continue to use the GLRT classification so;

By reconstruct constellation point m _iAnd modulation constellation points

Between Euclidean distance nearest, the maximum likelihood that provides modulation constellation points is estimated

Subscript c has represented different modulation types; Use GLRT classification same order modulation type then; Expression formula is as follows

Wherein,

The order of modulation that expression estimates,

Expression makes l _{GLR, m}Obtain the modulation system c of minimum value.Through type (10) has been finished final Modulation Mode Recognition.

Images