CN103763048A - Authorized user detection method and system based on characteristic value - Google Patents

Abstract

The invention provides an authorized user detection method and system based on a characteristic value. The method comprises the steps that signals received from a detection channel are sampled to obtain a receiving signal matrix; if authorized user signals exist in the receiving signals of the detection channel, the characteristic value of a statistical covariance matrix of the receiving signal matrix is calculated, and the number of authorized users is pre-estimated according to the characteristic value; the ratio of the receiving signal power maximum likelihood estimation value obtained on the condition that the authorized users with the pre-estimated number exist to the noise power maximum likelihood estimation value obtained on the condition that no authorized user exists is calculated; when the ratio is larger than the judgment threshold determined according to the false alarm probability, it is determined that the authorized users exist in the detection channel. The authorized user detection method and system based on the characteristic value can still detect the authorized users even through the number of the possible authorized users is unknown in advance.

Description

A kind of method and system of the detection authorized user based on characteristic value

Technical field

The present invention relates to the detection field of cognitive radio to authorized user, relate in particular to a kind of method and system of the detection authorized user based on characteristic value.

Background technology

Be accompanied by mobile communication technology development with rapid changepl. never-ending changes and improvements, people also grow with each passing day to the requirement of message transmission rate and service quality.But limited frequency spectrum resource becomes the bottleneck problem of restriction cordless communication network technical development.The frequency spectrum resource major part current due to China is the mode according to fixed allocation, distributes to the system-specific of mandate, and to avoid the phase mutual interference between each system, but frequency spectrum resource is not occupied all the time, thereby caused the waste of frequency spectrum resource.

Cognitive radio technology, as the effective technology that improves the availability of frequency spectrum, had caused in recent years widely and had paid close attention to.The core concept of cognitive radio technology is, adopt suitable frequency spectrum perception mechanism, the frequency spectrum that real-time discovery is not fully utilized, various resources are effectively distributed to user to be accessed, when authorized user need to access, channel is vacated, guarantee the normal work of authorized user not to be exerted an influence simultaneously.Cognitive radio technology, by automatic detection radio environment, reconfigure Internet resources, adjust mode of operation and transformation parameter, realize coexisting of cognitive user and authorized user, not only changed spectrum utilization mode, improved frequency spectrum service efficiency, and whole architectures of communication networks has been brought to potential dramatic change.

Whether detect target frequency bands authorized CU, whether have idle available cavity in frequency spectrum, be that cognitive radio carries out other movable prerequisites.Authorized user detection method based on characteristic value is a kind of brand-new frequency spectrum sensing method of rising in recent years, and the core of algorithm is exactly to utilize the characteristic value of the statistics covariance matrix that receives signal to judge whether to exist authorization user signal.The existing authorized user detection method based on characteristic value mainly contains: maximin is than method (Maximum Minimum Eigenvalue, MME), mean value method (Arithmetic Geometric Mean, AGM), Generalized Likelihood Ratio detection method (Generalized Likelihood Ratio Test, GLRT).

Below these several algorithms are briefly introduced:

MME algorithm is to calculate the eigenvalue of maximum of statistics covariance matrix and the ratio of minimal eigenvalue that receive signal, by this ratio and according to the definite decision threshold of mistake alarm probability, compares, if this ratio is greater than decision threshold, thinks and has authorized user.MME algorithm can, for the situation of single authorized user or the existence of multiple authorized user, still, because it has only used the partial information that receives signal statistics covariance matrix, cause detecting performance accuracy rate not high.

AGM algorithm is the arithmetic average of characteristic value and the ratio of geometric average that calculates the statistics covariance matrix that receives signal, by this ratio and according to the definite decision threshold of mistake alarm probability, compare, if this ratio is greater than decision threshold, thinks and have authorized user.AGM algorithm has been ignored the extreme value of statistics covariance matrix characteristic value, and when detected authorized user number continues to increase, it detects performance and also declines thereupon; When detected authorized user is single authorized user, its performance is not as GLRT algorithm.

The signal power of single authorized user and the ratio of noise power that GLRT algorithm exists hypothesis do maximal possibility estimation, if this ratio is greater than according to the definite decision threshold of mistake alarm probability, determines and have authorized user.Obviously, GLRT algorithm is only applicable to judge whether to exist the situation of single authorized user at present, the inapplicable detection to many authorized users.

In addition, state in the use before the authorized user detection method based on characteristic value, need the known authorized user number that may exist, but when the authorization user signal number that may exist is unknown, above-mentioned authorized user detection method is no longer applicable.

Summary of the invention

The invention provides a kind of method and system of the detection authorized user based on characteristic value, solved the technical problem that how to detect authorized user when not knowing the authorized user number that may exist in advance.

For solving the problems of the technologies described above, the invention provides a kind of method of the detection authorized user based on characteristic value, described method comprises:

To obtaining receiving signal matrix from the signal sampling that detects channel reception;

Suppose that the reception signal that detects channel exists authorization user signal, calculate the characteristic value of the statistics covariance matrix that receives signal matrix, according to described characteristic value, the number of authorized user is carried out to pre-estimation;

There is the power maximal possibility estimation and the ratio that does not have the maximal possibility estimation of noise power under authorized user condition that under the condition of authorized user of pre-estimation number, receive signal in calculating;

When described ratio is greater than according to the definite decision threshold of mistake alarm probability, determine that described detection channel exists authorized user.

Further, the described characteristic value according to statistics covariance matrix is carried out pre-estimation to the number of authorized user, comprising:

The characteristic value of counting statistics covariance matrix;

Characteristic value, by descending, is found to the wherein characteristic value of bust;

Number using the number of the characteristic value larger than this bust characteristic value as the authorized user of pre-estimation.

Further, calculate the power maximal possibility estimation that receives signal under the condition of authorized user that has pre-estimation number and the ratio that does not have the maximal possibility estimation of noise power under authorized user condition, comprising:

In the characteristic value of counting statistics covariance matrix, represent the characteristic value sum of authorized user and the ratio of all the other characteristic value sums of pre-estimation.

For solving the problems of the technologies described above, the present invention also provides a kind of system of the detection authorized user based on characteristic value, and described system comprises: authorized user number pre-estimation module, authorized user judge module, wherein,

Described authorized user number pre-estimation module, for obtaining receiving signal matrix to the signal sampling receiving from detection channel; Calculate the statistics covariance matrix that receives signal matrix, according to the characteristic value of described statistics covariance matrix, the number of authorized user is carried out to pre-estimation; The authorized user number of pre-estimation is sent to authorized user judge module;

Described authorized user judge module, for supposing that the reception signal that detects channel exists authorization user signal, calculate the power maximal possibility estimation and the ratio that does not have the maximal possibility estimation of noise power under authorized user condition that under the condition of the authorized user that has pre-estimation number, receive signal; And when judging described ratio and be greater than according to the definite decision threshold of mistake alarm probability, determine that described detection channel exists authorized user.

Further, described authorized user number pre-estimation module, for the number of authorized user being carried out to pre-estimation according to the characteristic value of described statistics covariance matrix, comprising:

The characteristic value of counting statistics covariance matrix; Characteristic value, by descending, is found to the wherein characteristic value of bust; Number using the number of the characteristic value larger than this bust characteristic value as the authorized user of pre-estimation.

Further, described authorized user judge module, for calculating the power maximal possibility estimation that receives signal under the condition of the authorized user that has pre-estimation number and the ratio that does not have the maximal possibility estimation of noise power under authorized user condition, comprising:

In the characteristic value of counting statistics covariance matrix, represent the characteristic value sum of authorized user and the ratio of all the other characteristic value sums of pre-estimation.

Technique scheme provided a kind of method of estimating authorized user number, solved prior art due to the problem of not knowing authorized user number and cannot utilizing characteristic value to detect authorized user; In addition, above-described embodiment, when whether detection exists authorized user, has considered to receive the All Eigenvalues of signal statistics covariance matrix, has improved the accuracy rate detecting.

Accompanying drawing explanation

Fig. 1 is the method flow diagram of the detection authorized user based on characteristic value of the present embodiment;

Fig. 2 is the system composition module map of the detection authorized user based on characteristic value of the present embodiment.

Embodiment

For making the object, technical solutions and advantages of the present invention clearer, hereinafter in connection with accompanying drawing, embodiments of the invention are elaborated.It should be noted that, in the situation that not conflicting, the combination in any mutually of the feature in embodiment and embodiment in the application.

Fig. 1 is the method flow diagram of the detection authorized user based on characteristic value of the present embodiment.

S101, within a perception cycle of cognitive radio system, utilizes M reception antenna to carrying out n sampling, M>1, n>2M from the signal that detects channel reception;

The signal that S102 obtains n sampling of M antenna is designated as and receives signal matrix X _{m × n};

S103 builds the statistics covariance matrix R that receives signal, R=E[X X*], X* represents the conjugate transpose of X, the eigenvalue λ of compute matrix R _i, i=1 ... M, presses descending λ to characteristic value ₁>=...>=λ _m;

There is authorization user signal in the reception signal that S104 hypothesis detects channel, according to the characteristic value of matrix R, the number of authorized user is carried out to pre-estimation;

Conventionally, if detect the reception signal of channel, have authorization user signal, the signal power of its authorized user is much larger than noise power.Like this, in the characteristic value of matrix R, the characteristic value that numerical value is large should be from authorized user, and the characteristic value that numerical value is little is carried out self noise.In the eigenvalue λ of pressing descending ₁>=...>=λ _mthe characteristic value of middle searching bust, the number using the number of the characteristic value larger than this bust characteristic value as the authorized user of pre-estimation;

Because matrix R obeys the Wishart distribution that parameter is n and M, according to Random Matrices Theory, can use MDL(minimum description length criteria) or AIC(Akaike information criterion) etc. the number of the characteristic value that numerical value is large in algorithm for estimating compute matrix R;

S105 calculates the power maximal possibility estimation that receives signal under the condition of authorized user that has pre-estimation number and the ratio LR (X) of maximal possibility estimation that does not have noise power under authorized user condition;

$\mathrm{LR}\left(X\right)=\frac{{(M-\hat{p})}^{(M-\hat{p})n}}{M^{\mathrm{Mn}}}{\left(\frac{1}{\mathrm{\μ}{(1-\mathrm{\μ})}^{M-\hat{p}}}\right)}^n,\mathrm{\μ}=\frac{\underset{i=1}{\overset{\hat{p}}{\mathrm{\Σ}}}{\mathrm{\λ}}_i}{\mathrm{tr}\left(R\right)},$ The mark of tr () representing matrix,

represent the pre-estimation number of authorized user;

S106 judges whether LR (X) is greater than according to the definite decision threshold of mistake alarm probability, if be greater than, the hypothesis of description of step S104 is correct, detects channel and has authorized user; Otherwise the hypothesis mistake of description of step S104, is noise signal from the signal that detects channel reception;

If the definite decision threshold of mistake alarm probability is γ;

According to mistake alarm probability P _fadefinition: P _fap[LR (X) > γ | H ₀], H ₀represent that reality does not exist authorized user, LR (X) > γ represents that court verdict is for existing authorized user.At given P _faprerequisite under, according to the probability-distribution function of LR (X), can instead release γ.

In order to simplify computational process, can be right

carry out abbreviation

$\mathrm{LR}\left(X\right)=\frac{{(M-\hat{p})}^{(M-\hat{p})N}}{M^{\mathrm{MN}}}{\left(\frac{1}{\mathrm{\&mu;}{(1-\mathrm{\&mu;})}^{M-\hat{p}}}\right)}^N\mathrm{\&gamma;}_{<}^{>}$

h ₁represent physical presence authorized user;

order $f\left(\mathrm{\&mu;}\right)=\mathrm{\&mu;}{(1-\mathrm{\&mu;})}^{M-\hat{p}},f^{\&prime;}\left(\mathrm{\&mu;}\right)={(1-\mathrm{\&mu;})}^{M-\hat{p}}+\mathrm{\&mu;}(M-\hat{p}){(1-\mathrm{\&mu;})}^{M-\hat{p}-1},$ Due to

can obtain 0< μ <1, f ' is >0 (μ);

f (μ) is positive monotonic function, also be monotonic function; Because independent variable and the dependent variable of monotonic function are one to one,

t ₂be equivalent to

η is equivalent decision threshold

According to mistake alarm probability P _fadefinition obtain: P _fa=P[T (x) >, η | H ₀], at given P _faprerequisite under, according to λ in T (x) _iprobability-distribution function, can instead release η,

$\mathrm{\&eta;}=\frac{\hat{p}{(1+\sqrt{\frac{M}{n}})}^2}{M-\hat{p}}+\frac{\hat{p}}{n(M-\hat{p})}\frac{{(\sqrt{M}+\sqrt{N})}^{4/3}}{{\left(\sqrt{\mathrm{Mn}}\right)}^{1/3}}{F}_{\mathrm{TW}2}^{-1}(1-P_{\mathrm{fa}}),$ Wherein

for the inverse function of second order Tracy-Widom distribution function.

Above-described embodiment provided a kind of method of estimating authorized user number, solved prior art due to the problem of not knowing authorized user number and cannot utilizing characteristic value to detect authorized user; In addition, above-described embodiment, when whether detection exists authorized user, has considered to receive the All Eigenvalues of signal statistics covariance matrix, has improved the accuracy rate detecting.

Fig. 2 is the system composition module map of the detection authorized user based on characteristic value of the present embodiment.

Described system comprises authorized user number pre-estimation module, authorized user judge module, wherein,

Described authorized user number pre-estimation module, for obtaining receiving signal matrix to the signal sampling receiving from detection channel; Calculate the statistics covariance matrix that receives signal matrix, according to the characteristic value of described statistics covariance matrix, the number of authorized user is carried out to pre-estimation; The authorized user number of pre-estimation is sent to authorized user judge module;

Described authorized user number pre-estimation module is within a perception cycle of cognitive radio system, utilize M reception antenna to carrying out n sampling from the signal that detects channel reception, M>1, n>2M, the signal that n sampling of M antenna obtained is designated as and receives signal matrix X _{m × n}; According to receiving signal matrix X _{m × n}build the statistics covariance matrix R that receives signal, R=E[X X*], X* represents the conjugate transpose of X, the eigenvalue λ of compute matrix R _i, i1 ... M, presses descending λ to characteristic value ₁>=...>=λ _m; Suppose that the reception signal that detects channel exists authorization user signal, in the eigenvalue λ of pressing descending ₁>=...>=λ _mthe characteristic value of middle searching bust, the number using the number of the characteristic value larger than this bust characteristic value as the authorized user of pre-estimation;

Described authorized user judge module, for supposing that the reception signal that detects channel exists authorization user signal, calculate the power maximal possibility estimation and the ratio that does not have the maximal possibility estimation of noise power under authorized user condition that under the condition of the authorized user that has pre-estimation number, receive signal; And when judging described ratio and be greater than according to the definite decision threshold of mistake alarm probability, determine that described detection channel exists authorized user;

Under the condition of the authorized user that has pre-estimation number that authorized user judge module calculates, receive the power maximal possibility estimation and the ratio that does not have the maximal possibility estimation of noise power under authorized user condition of signal $\mathrm{LR}\left(X\right)=\frac{{(M-\hat{p})}^{(M-\hat{p})n}}{M^{\mathrm{Mn}}}{\left(\frac{1}{\mathrm{\&mu;}{(1-\mathrm{\&mu;})}^{M-\hat{p}}}\right)}^n,\mathrm{\&mu;}=\frac{\underset{i=1}{\overset{\hat{p}}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_i}{\mathrm{tr}\left(R\right)},$ The mark of tr () representing matrix,

represent the pre-estimation number of authorized user.By LR (X) and according to the definite decision threshold γ of mistake alarm probability, compare, if LR (X) is greater than decision threshold γ, explanation detects channel and has authorized user; Otherwise, illustrate that the signal from detecting channel reception is noise signal.

According to mistake alarm probability P _fadefinition: P _fap[LR (X) > γ | H ₀], H ₀represent that reality does not exist authorized user, LR (X) > γ represents that court verdict is for existing authorized user.At given P _faprerequisite under, according to the probability-distribution function of LR (X), can instead release γ;

In order to simplify computational process, can be right

carry out abbreviation.

$\mathrm{LR}\left(X\right)=\frac{{(M-\hat{p})}^{(M-\hat{p})N}}{M^{\mathrm{MN}}}{\left(\frac{1}{\mathrm{\&mu;}{(1-\mathrm{\&mu;})}^{M-\hat{p}}}\right)}^N\mathrm{\&gamma;}_{<}^{>}$ Through simplifying, can obtain

η is equivalent decision threshold, wherein, and H ₁represent physical presence authorized user;

According to mistake alarm probability P _fadefinition obtain: P _fap[T (X) > η | H ₀], at given P _faprerequisite under, according to λ in T (X) _iprobability-distribution function, can instead release η, $\mathrm{\&eta;}=\frac{\hat{p}{(1+\sqrt{\frac{M}{n}})}^2}{M-\hat{p}}+\frac{\hat{p}}{n-(M-\hat{p})}\frac{{(\sqrt{M}+\sqrt{N})}^{4/3}}{{\left(\sqrt{\mathrm{Mn}}\right)}^{1/3}}{F}_{\mathrm{TW}2}^{-1}(1-P_{\mathrm{fa}}),$ Wherein

for the inverse function of second order Tracy-Widom distribution function.

Above-described embodiment provided a kind of method of estimating authorized user number, solved prior art due to the problem of not knowing authorized user number and cannot utilizing characteristic value to detect authorized user; In addition, above-described embodiment, when whether detection exists authorized user, has considered to receive the All Eigenvalues of signal statistics covariance matrix, has improved the accuracy rate detecting.

One of ordinary skill in the art will appreciate that all or part of step in said method can carry out instruction related hardware by program and complete, described program can be stored in computer-readable recording medium, as read-only memory, disk or CD etc.Alternatively, all or part of step of above-described embodiment also can realize with one or more integrated circuits, and correspondingly, the each module/unit in above-described embodiment can adopt the form of hardware to realize, and also can adopt the form of software function module to realize.The present invention is not restricted to the combination of the hardware and software of any particular form.

It should be noted that; the present invention also can have other various embodiments; in the situation that not deviating from spirit of the present invention and essence thereof; those of ordinary skill in the art can make according to the present invention various corresponding changes and distortion, but these corresponding changes and distortion all should belong to the protection range of the appended claim of the present invention.

Claims (6)

1. a method for the detection authorized user based on characteristic value, is characterized in that, described method comprises:

To obtaining receiving signal matrix from the signal sampling that detects channel reception;

Suppose that the reception signal that detects channel exists authorization user signal, calculate the characteristic value of the statistics covariance matrix that receives signal matrix, according to described characteristic value, the number of authorized user is carried out to pre-estimation;

There is the power maximal possibility estimation and the ratio that does not have the maximal possibility estimation of noise power under authorized user condition that under the condition of authorized user of pre-estimation number, receive signal in calculating;

When described ratio is greater than according to the definite decision threshold of mistake alarm probability, determine that described detection channel exists authorized user.

2. the method for claim 1, is characterized in that, the described characteristic value according to statistics covariance matrix is carried out pre-estimation to the number of authorized user, comprising:

The characteristic value of counting statistics covariance matrix;

Characteristic value, by descending, is found to the wherein characteristic value of bust;

Number using the number of the characteristic value larger than this bust characteristic value as the authorized user of pre-estimation.

3. method as claimed in claim 2, is characterized in that, calculates the power maximal possibility estimation that receives signal under the condition of authorized user that has pre-estimation number and the ratio that does not have the maximal possibility estimation of noise power under authorized user condition, comprising:

In the characteristic value of counting statistics covariance matrix, represent the characteristic value sum of authorized user and the ratio of all the other characteristic value sums of pre-estimation.

4. a system for the detection authorized user based on characteristic value, is characterized in that, described system comprises: authorized user number pre-estimation module, authorized user judge module, wherein,

Described authorized user number pre-estimation module, for obtaining receiving signal matrix to the signal sampling receiving from detection channel; Calculate the statistics covariance matrix that receives signal matrix, according to the characteristic value of described statistics covariance matrix, the number of authorized user is carried out to pre-estimation; The authorized user number of pre-estimation is sent to authorized user judge module;

Described authorized user judge module, for supposing that the reception signal that detects channel exists authorization user signal, calculate the power maximal possibility estimation and the ratio that does not have the maximal possibility estimation of noise power under authorized user condition that under the condition of the authorized user that has pre-estimation number, receive signal; And when judging described ratio and be greater than according to the definite decision threshold of mistake alarm probability, determine that described detection channel exists authorized user.

5. system as claimed in claim 4, is characterized in that, described authorized user number pre-estimation module, for the number of authorized user being carried out to pre-estimation according to the characteristic value of described statistics covariance matrix, comprising:

The characteristic value of counting statistics covariance matrix; Characteristic value, by descending, is found to the wherein characteristic value of bust; Number using the number of the characteristic value larger than this bust characteristic value as the authorized user of pre-estimation.

6. system as claimed in claim 5, it is characterized in that, described authorized user judge module, for calculating the power maximal possibility estimation that receives signal under the condition of the authorized user that has pre-estimation number and the ratio that does not have the maximal possibility estimation of noise power under authorized user condition, comprising:

In the characteristic value of counting statistics covariance matrix, represent the characteristic value sum of authorized user and the ratio of all the other characteristic value sums of pre-estimation.

Images