CN105978646A - Double-threshold cooperative spectrum sensing method based on energy detection of small sample - Google Patents

Abstract

The invention discloses a double-threshold cooperative spectrum sensing method based on energy detection of a small sample. Double thresholds are used to effectively reduce interference of a cognition user in a main user under the condition of low signal to noise ratio, a multidimensional Gaussian approximate processing detection result is used to realize energy detection of the small sample, and the principle that most votes most suitable for practical application in hard decision is used in a fusion center to make a final decision. Compared with a traditional energy detection method, The method of the invention has the advantages that the degree of interference of the cognition user in the main user is effectively reduced in a spectrum sensing process under the condition of small sample and low signal to noise ratio, the possibility of missed energy detection is reduced, and the detection performance of the system is improved.

Description

A kind of double-threshold cooperative spectrum sensing method based on small sample energy measuring

Technical field

The present invention relates to cognitive radio technology field, a kind of double threshold based on small sample energy measuring cooperates Frequency spectrum sensing method.

Background technology

In recent years, due to the fast development of wireless communication technology, the rare problem of radio spectrum resources is more obvious, The most present fixed frequency spectrum distribution policy can not meet people's demand to frequency spectrum resource.Cognitive radio (Cognitive Radio, CR) technology allows the mandate frequency spectrum that cognitive user dynamic access is idle, is that one effectively solves frequency spectrum The method of scarcity of resources.Cognitive user is judged to authorize frequency spectrum the most idle by frequency spectrum perception technology, thus decide whether can To access this mandate frequency spectrum.Continue to monitor in real time the service condition of frequency spectrum after cognitive user insertion authority frequency spectrum, once detect Primary user's signal soon exits, it is to avoid primary user's signal is produced interference.Therefore, frequency spectrum perception be cognitive radio technology One of important basis.

Classical frequency spectrum perception algorithm mainly includes the detection of matched filtering device, cyclostationary characteristic detection and energy measuring. Matched filtering device detection method can obtain and preferably detect performance, but its major defect is cognitive user it is to be appreciated that primary user's signal Priori, the energy consumption of this detection method is many, implements complexity high.Cyclostationary characteristic detection method can well be distinguished Primary user's signal and noise, but the distribution of its Spectral correlation function is sufficiently complex, and this detection method is computationally intensive, and real-time is relatively Difference.Energy measuring method need not the priori of primary user because of it, implements the advantages such as simple and algorithm complex is low, and becomes frequency One of spectrum perception most common method.

The simple gate that traditional energy detection method uses limits in the case of low signal-to-noise ratio, and detection performance is easily subject to noise The impact of fluctuation.The method having research double threshold energy measuring in recent years, but these double threshold energy detection methods are all bases In central limit theorem (Central Limit Theorem, CLT), it is desirable to the sample number of sampling must be sufficiently large.

Summary of the invention

The technical problem to be solved is to overcome the deficiencies in the prior art to provide a kind of based on small sample energy The double-threshold cooperative spectrum sensing method of detection, the inventive method can also be effective in the case of small sample number and low signal-to-noise ratio Ground reduces the secondary user's interference to primary user, reduces false dismissal probability, improves the detection performance of system.

The present invention solves above-mentioned technical problem by the following technical solutions:

A kind of based on small sample energy measuring the double-threshold cooperative spectrum sensing method proposed according to the present invention, including such as Lower step:

Step 1, the binary hypothesis test problem model set up according to the signal detection of secondary user's SU, it is thus achieved that n moment SU Observation x (n)；Wherein n=1,2 ..., N, N are sample number；

Step 2, according to conventional energy detection CED frequency spectrum perception obtain statistic of test T (x)；

Step 3, statistic of test T (x) and two threshold values λ pre-set that will obtain in step 2₀And λ₁According to sentencing Certainly criterion compares respectively, and sends the court verdict compared to system globe area center FC；Decision rule is: if T (x) ≥λ₁, then show that primary user is taking mandate frequency spectrum；If T (x)≤λ₀, then primary user's vacant mandate frequency spectrum is shown；If λ₀＜ T (x) ＜ λ₁, do not enter a judgement, re-execute step 1 to step 2；Wherein, λ₁>λ₀；

Step 4, account for the definition of probability according to the decision rule in step 3 and collision probability and limit, it is thus achieved that double threshold energy Collision probability P of amount detection_cdProbability P is accounted for limit_nd:

$P_{cd}=P\left(T\right(x)<{\mathrm{\&lambda;}}_0|H_1)=F_{{\mathrm{\&chi;}}_{2N}^2}(2{\mathrm{\&lambda;}}_0/{\mathrm{\&sigma;}}_w^2(\mathrm{\&gamma;}+1\left)\right);$

$P_{nd}=P\left(T\right(x)>{\mathrm{\&lambda;}}_0|H_0)=1-F_{{\mathrm{\&chi;}}_{2N}^2}(2{\mathrm{\&lambda;}}_0/{\mathrm{\&sigma;}}_w^2);$

Wherein, P (*) represents probability function, H₁Represent and authorize the situation that in frequency spectrum, primary user's signal exists, H₀Represent and authorize The non-existent situation of primary user's signal in frequency spectrum,Be degree of freedom be 2N center card side distribution cumulative distribution function, σ_w It is the mean-square value of noise,It is average signal-to-noise ratio, σ_sIt it is the mean-square value of primary user's PU signal sampling value；

Step 5, account for probability and multidimensional Gaussian approximation method according to the collision probability tried to achieve in step 4 and limit, it is thus achieved that little The collision probability approximation of sample energy detection double thresholdProbability approximation is accounted for limit

${\tilde{P}}_{cd}\&ap;1-Q\left(\frac{\sqrt[3]{{\mathrm{\&lambda;}}_0/{\mathrm{N\&sigma;}}_w^2(\mathrm{\&gamma;}+1)}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right),{\tilde{P}}_{nd}\&ap;Q\left(\frac{\sqrt[3]{{\mathrm{\&lambda;}}_0/{\mathrm{N\&sigma;}}_w^2}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right);$

Wherein, Q (*) represents Gauss Q function；

Step 6, according to the court verdict sent in step 3 to system globe area center FC, use great majority ballot to merge standard Then, further according to the collision probability approximation of the small sample energy measuring double threshold tried to achieve in step 5Probability approximation is accounted for limitObtain the collision probability of double-threshold cooperative spectrum sensing method based on small sample energy measuringProbability is accounted for limit

$Q_K^{cd}=\underset{m=K}{\overset{M}{\&Sigma;}}\left(\begin{array}{c}M\\ m\end{array}\right){\left({\tilde{P}}_{cd}\right)}^m\&times;{(1-{\tilde{P}}_{cd})}^{M-m},Q_K^{nd}=\underset{m=K}{\overset{M}{\&Sigma;}}\left(\begin{array}{c}M\\ m\end{array}\right){\left({\tilde{P}}_{nd}\right)}^m\&times;{(1-{\tilde{P}}_{nd})}^{M-m};$

Wherein, M represents secondary user's number, and m is variable.

Further as a kind of double-threshold cooperative spectrum sensing method based on small sample energy measuring of the present invention Prioritization scheme, the binary hypothesis test problem model that in described step 1, the signal detection of SU is set up is

$\left\{\begin{array}{cc}x\left(n\right)=w\left(n\right)& H_0\\ x\left(n\right)=s\left(n\right)+w\left(n\right)& H_1\end{array}\right.$

Wherein, w (n) be average be 0, variance isAdditive white Gaussian noise, s (n) is the sampled value of n moment PU signal And to set s (n) as average be 0, variance isStochastic signal, and s (n) and w (n) separate.

Further as a kind of double-threshold cooperative spectrum sensing method based on small sample energy measuring of the present invention Prioritization scheme, statistic of test T (x) obtained in described step 2 is specific as follows,

$T\left(x\right)=\underset{n=0}{\overset{N-1}{\&Sigma;}}{x}^2\left(n\right).$

Further as a kind of double-threshold cooperative spectrum sensing method based on small sample energy measuring of the present invention Prioritization scheme, λ in described step 3₀=0.8 λ, λ₁=1.2 λ, λ are the threshold value of conventional energy detection.

Further as a kind of double-threshold cooperative spectrum sensing method based on small sample energy measuring of the present invention Prioritization scheme, the collision probability in described step 4 refers to that secondary user's mistaken verdict primary user does not exists and takies mandate frequency spectrum, But the probability that actually primary user exists；Limit accounts for probability and refers to that secondary user's mistaken verdict primary user exists and is not take up authorizing frequency Spectrum, but the actually non-existent probability of primary user.

Further as a kind of double-threshold cooperative spectrum sensing method based on small sample energy measuring of the present invention Prioritization scheme, the multidimensional Gaussian approximation method in described step 5 refers to that divided by 2N, card side stochastic variable a is approximately average is 1- (9N)^-1And variance is (9N)^-1Multidimensional Gaussian random variable, i.e.

$F_{{\mathrm{\&chi;}}_{2N}^2}\left(a\right)\&ap;1-Q\left(\frac{\sqrt[3]{a/\left(2N\right)}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right).$

Further as a kind of double-threshold cooperative spectrum sensing method based on small sample energy measuring of the present invention Prioritization scheme, the collision probability of double-threshold cooperative spectrum sensing method based on small sample energy measuring and limit in described step 6 Account for the further expression formula of probability to be respectively as follows:

$\begin{array}{c}{Q}_K^{cd}=\underset{m=K}{\overset{M}{\&Sigma;}}\left(\begin{array}{c}M\\ m\end{array}\right){\left({\tilde{P}}_{cd}\right)}^m\&times;{\left(1-{\tilde{P}}_{cd}\right)}^{M-m}\\ =\underset{m=K}{\overset{M}{\&Sigma;}}\left(\begin{array}{c}M\\ m\end{array}\right){\left(1-Q\left(\frac{\sqrt[3]{{\mathrm{\&lambda;}}_0/{\mathrm{N\&sigma;}}_w^2\left({\mathrm{\&gamma;}}^{\&prime;}+1\right)}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right)\right)}^m\&times;{\left(Q\left(\frac{\sqrt[3]{{\mathrm{\&lambda;}}_0/{\mathrm{N\&sigma;}}_w^2\left({\mathrm{\&gamma;}}^{\&prime;}+1\right)}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right)\right)}^{M-m}\end{array}$

$\begin{array}{c}{Q}_K^{nd}=\underset{m=K}{\overset{M}{\&Sigma;}}\left(\begin{array}{c}M\\ m\end{array}\right){\left({\tilde{P}}_{nd}\right)}^m\&times;{\left(1-{\tilde{P}}_{nd}\right)}^{M-m}\\ =\underset{m=K}{\overset{M}{\&Sigma;}}\left(\begin{array}{c}M\\ m\end{array}\right){\left(Q\left(\frac{\sqrt[3]{{\mathrm{\&lambda;}}_0/{\mathrm{N\&sigma;}}_w^2}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right)\right)}^m\&times;{\left(1-Q\left(\frac{\sqrt[3]{{\mathrm{\&lambda;}}_0/{\mathrm{N\&sigma;}}_w^2}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right)\right)}^{M-m}\end{array}.$

The present invention uses above technical scheme compared with prior art, has following technical effect that

(1) present invention uses double threshold to effectively reduce secondary user's in the case of low signal-to-noise ratio to do primary user Disturb, utilize multidimensional Gaussian approximation to process testing result and achieve small sample energy measuring, and use hard decision at fusion center In be best suitable for " great majority ballot " principle of actual application and make conclusive judgement；

(2) present invention has in the case of small sample and low signal-to-noise ratio and can also efficiently reduce during frequency spectrum perception time Level user's annoyance level to primary user, reduces the advantages such as the false dismissal probability of energy measuring, the detection performance of raising system.

Accompanying drawing explanation

Fig. 1 is double threshold energy measuring model.

Fig. 2 is that the present invention designs the flow process of a kind of double-threshold cooperative spectrum sensing method based on small sample energy measuring and shows It is intended to.

Detailed description of the invention

Below in conjunction with the accompanying drawings technical scheme is described in further detail:

As in figure 2 it is shown, a kind of based on small sample energy measuring the double-threshold cooperative spectrum sensing method of present invention design, Comprise the steps:

The binary hypothesis test that step 1. is set up according to the signal detection of following secondary user's (Secondary User, SU) Observation x (n) of problem model acquisition n moment SU:

$\left\{\begin{array}{cc}x\left(n\right)=w\left(n\right)& H_0\\ x\left(n\right)=s\left(n\right)+w\left(n\right)& H_1\end{array}\right.---\left(1\right)$

Wherein n=1,2 ..., N, N are sample number；W (n) be average be 0, variance isAdditive white Gaussian noise, σ_w Being the mean-square value of noise, s (n) is the sampled value of n moment primary user (Primary User, PU) signal.For sake of simplicity, assume s (n) be average be 0, variance isStochastic signal, σ_sIt is the mean-square value of primary user's PU signal sampling value, and s (n) and w (n) phase The most independent.H₀Represent the non-existent situation of primary user's signal, H₁Represent the situation that primary user's signal exists.

Step 2. obtains according to the frequency spectrum perception of conventional energy detection (Conventional Energy Detection, CED) Obtain statistic of test T (x)

$T\left(x\right)=\underset{n=0}{\overset{N-1}{\&Sigma;}}{x}^2\left(n\right)---\left(2\right)$

Statistic of test T (x) obtained in step 3. comparison step 2 and two threshold values λ pre-set₀And λ₁, and will Result of the comparison sends to system globe area center (Fusion Center, FC)；If T (x) >=λ₁, then primary user is being shown Taking mandate frequency spectrum, court verdict is " 1 "；If T (x)≤λ₀, then showing primary user's vacant mandate frequency spectrum, court verdict is “0”；If λ₀＜ T (x) ＜ λ₁, do not enter a judgement, re-execute step 1 to step 2；Wherein λ₀=0.8 λ, λ₁=1.2 λ, λ are The threshold value of conventional energy detection, double threshold energy measuring model as shown in Figure 1, λ₁>λ₀。

Double threshold energy is tried to achieve in the definition that step 4. accounts for probability according to the decision rule in step 3 and collision probability and limit Collision probability P of detection_cdProbability P is accounted for limit_nd；Collision probability refers to that secondary user's mistaken verdict primary user does not exists and takies Mandate frequency spectrum, but the probability that actually primary user exists, its expression formula is

$P_{cd}=P\left(T\right(x)<{\mathrm{\&lambda;}}_0|H_1)=F_{{\mathrm{\&chi;}}_{2N}^2}(2{\mathrm{\&lambda;}}_0/{\mathrm{\&sigma;}}_w^2(\mathrm{\&gamma;}+1\left)\right)---\left(3\right)$

Collision probability illustrates the annoyance level that primary user is produced by secondary user's due to incorrect noise.Collision is general Rate is the biggest, and secondary user's is the most serious to the interference of primary user；Otherwise, disturb the least.

Limit accounts for probability and refers to that secondary user's mistaken verdict primary user exists and is not take up authorizing frequency spectrum, but the most primary The non-existent probability in family, its expression formula is

$P_{nd}=P\left(T\right(x)>{\mathrm{\&lambda;}}_0|H_0)=1-F_{{\mathrm{\&chi;}}_{2N}^2}(2{\mathrm{\&lambda;}}_0/{\mathrm{\&sigma;}}_w^2)---\left(4\right)$

Limit accounts for probability and illustrates spectrum utilization efficiency.It is the biggest that limit accounts for probability, and spectrum utilization efficiency is the lowest；Otherwise, frequency spectrum Utilization ratio is the highest.

Wherein, P (*) represents probability function,Be degree of freedom be 2N center card side distribution cumulative distribution function,It it is average signal-to-noise ratio.

Step 5. accounts for probability according to the collision probability tried to achieve in step 4 and limit and multidimensional Gaussian approximation method is tried to achieve respectively The collision probability of small sample energy measuring double threshold and limit account for the approximation of probability；Multidimensional Gaussian approximation method refers to that card side is random It is 1-(9N) that variable can be approximated to be average divided by 2N^-1And variance is (9N)^-1Multidimensional Gaussian random variable, i.e.

$F_{{\mathrm{\&chi;}}_{2N}^2}\left(a\right)\&ap;1-Q\left(\frac{\sqrt[3]{a/\left(2N\right)}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right)$

Wherein, a is card side's stochastic variable, and Q (*) is Gauss Q function；

Collision probability approximation based on small sample energy measuring double threshold is

${\tilde{P}}_{cd}\&ap;1-Q\left(\frac{\sqrt[3]{{\mathrm{\&lambda;}}_0/{\mathrm{N\&sigma;}}_w^2(\mathrm{\&gamma;}+1)}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right)---\left(5\right)$

Limit based on small sample energy measuring double threshold accounts for probability approximation

${\tilde{P}}_{nd}\&ap;Q\left(\frac{\sqrt[3]{{\mathrm{\&lambda;}}_0/{\mathrm{N\&sigma;}}_w^2}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right)---\left(6\right)$

Step 6: according to the result sent in step 3 to system globe area center, uses " great majority throwing at fusion center Ticket " criterion makes decisions, further according to the collision probability approximation tried to achieve in step 5 and limit account for probability approximation can obtain based on The collision probability of the double-threshold cooperative spectrum sensing method of small sample energy measuring is

$\begin{array}{c}{Q}_K^{cd}=\underset{m=K}{\overset{M}{\&Sigma;}}\left(\begin{array}{c}M\\ m\end{array}\right){\left({\tilde{P}}_{cd}\right)}^m\&times;{\left(1-{\tilde{P}}_{cd}\right)}^{M-m}\\ =\underset{m=K}{\overset{M}{\&Sigma;}}\left(\begin{array}{c}M\\ m\end{array}\right){\left(1-Q\left(\frac{\sqrt[3]{{\mathrm{\&lambda;}}_0/{\mathrm{N\&sigma;}}_w^2\left({\mathrm{\&gamma;}}^{\&prime;}+1\right)}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right)\right)}^m\&times;{\left(Q\left(\frac{\sqrt[3]{{\mathrm{\&lambda;}}_0/{\mathrm{N\&sigma;}}_w^2\left({\mathrm{\&gamma;}}^{\&prime;}+1\right)}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right)\right)}^{M-m}\end{array}---\left(7\right)$

Limit accounts for probability and is

$\begin{array}{c}{Q}_K^{nd}=\underset{m=K}{\overset{M}{\&Sigma;}}\left(\begin{array}{c}M\\ m\end{array}\right){\left({\tilde{P}}_{nd}\right)}^m\&times;{\left(1-{\tilde{P}}_{nd}\right)}^{M-m}\\ =\underset{m=K}{\overset{M}{\&Sigma;}}\left(\begin{array}{c}M\\ m\end{array}\right){\left(Q\left(\frac{\sqrt[3]{{\mathrm{\&lambda;}}_0/{\mathrm{N\&sigma;}}_w^2}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right)\right)}^m\&times;{\left(1-Q\left(\frac{\sqrt[3]{{\mathrm{\&lambda;}}_0/{\mathrm{N\&sigma;}}_w^2}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right)\right)}^{M-m}\end{array}---\left(8\right)$

Wherein, M represents secondary user's number, and m is variable.

Claims (7)

1. a double-threshold cooperative spectrum sensing method based on small sample energy measuring, it is characterised in that comprise the steps:

Step 1, the binary hypothesis test problem model set up according to the signal detection of secondary user's SU, it is thus achieved that the sight of n moment SU Measured value x (n)；Wherein n=1,2 ..., N, N are sample number；

Step 2, according to conventional energy detection CED frequency spectrum perception obtain statistic of test T (x)；

Step 3, statistic of test T (x) and two threshold values λ pre-set that will obtain in step 2₀And λ₁According to judgement standard Compare the most respectively, and the court verdict compared is sent to system globe area center FC；Decision rule is: if T (x) >= λ₁, then show that primary user is taking mandate frequency spectrum；If T (x)≤λ₀, then primary user's vacant mandate frequency spectrum is shown；If λ₀ ＜ T (x) ＜ λ₁, do not enter a judgement, re-execute step 1 to step 2；Wherein, λ₁>λ₀；

Step 4, account for the definition of probability according to the decision rule in step 3 and collision probability and limit, it is thus achieved that double threshold energy is examined Collision probability P surveyed_cdProbability P is accounted for limit_nd:

$P_{cd}=P\left(T\right(x)<{\mathrm{\&lambda;}}_0|H_1)=F_{{\mathrm{\&chi;}}_{2N}^2}(2{\mathrm{\&lambda;}}_0/{\mathrm{\&sigma;}}_w^2(\mathrm{\&gamma;}+1\left)\right);$

$P_{nd}=P\left(T\right(x)>{\mathrm{\&lambda;}}_0|H_0)=1-F_{{\mathrm{\&chi;}}_{2N}^2}(2{\mathrm{\&lambda;}}_0/{\mathrm{\&sigma;}}_w^2);$

Wherein, P (*) represents probability function, H₁Represent and authorize the situation that in frequency spectrum, primary user's signal exists, H₀Represent and authorize frequency spectrum The middle non-existent situation of primary user's signal,Be degree of freedom be 2N center card side distribution cumulative distribution function, σ_wIt is to make an uproar The mean-square value of sound,It is average signal-to-noise ratio, σ_sIt it is the mean-square value of primary user's PU signal sampling value；

Step 5, account for probability and multidimensional Gaussian approximation method according to the collision probability tried to achieve in step 4 and limit, it is thus achieved that small sample The collision probability approximation of energy measuring double thresholdProbability approximation is accounted for limit

${\tilde{P}}_{cd}\&ap;1-Q\left(\frac{\sqrt[3]{{\mathrm{\&lambda;}}_0/{\mathrm{N\&sigma;}}_w^2(\mathrm{\&gamma;}+1)}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right),{\tilde{P}}_{nd}\&ap;Q\left(\frac{\sqrt[3]{{\mathrm{\&lambda;}}_0/{\mathrm{N\&sigma;}}_w^2}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right);$

Wherein, Q (*) represents Gauss Q function；

Step 6, according to the court verdict sent in step 3 to system globe area center FC, use great majority ballot fusion criterion, then Collision probability approximation according to the small sample energy measuring double threshold tried to achieve in step 5Probability approximation is accounted for limitObtain Obtain the collision probability of double-threshold cooperative spectrum sensing method based on small sample energy measuringProbability is accounted for limit

$Q_K^{cd}=\underset{m=K}{\overset{M}{\&Sigma;}}\left(\begin{array}{c}M\\ m\end{array}\right){\left({\tilde{P}}_{cd}\right)}^m\&times;{(1-{\tilde{P}}_{cd})}^{M-m},Q_K^{nd}=\underset{m=K}{\overset{M}{\&Sigma;}}\left(\begin{array}{c}M\\ m\end{array}\right){\left({\tilde{P}}_{nd}\right)}^m\&times;{(1-{\tilde{P}}_{nd})}^{M-m};$

Wherein, M represents secondary user's number, and m is variable.

A kind of double-threshold cooperative spectrum sensing method based on small sample energy measuring the most according to claim 1, it is special Levying and be, the binary hypothesis test problem model that in described step 1, the signal detection of SU is set up is

$\left\{\begin{array}{cc}x\left(n\right)=w\left(n\right)& H_0\\ x\left(n\right)=s\left(n\right)+w\left(n\right)& H_1\end{array}\right.$

Wherein, w (n) be average be 0, variance isAdditive white Gaussian noise, s (n) is the sampled value of n moment PU signal and sets S (n) be average be 0, variance isStochastic signal, and s (n) and w (n) separate.

A kind of double-threshold cooperative spectrum sensing method based on small sample energy measuring the most according to claim 1, it is special Levying and be, statistic of test T (x) obtained in described step 2 is specific as follows,

$T\left(x\right)=\underset{n=0}{\overset{N-1}{\&Sigma;}}{x}^2\left(n\right).$

A kind of double-threshold cooperative spectrum sensing method based on small sample energy measuring the most according to claim 1, it is special Levy and be, λ in described step 3₀=0.8 λ, λ₁=1.2 λ, λ are the threshold value of conventional energy detection.

A kind of double-threshold cooperative spectrum sensing method based on small sample energy measuring the most according to claim 1, it is special Levying and be, the collision probability in described step 4 refers to that secondary user's mistaken verdict primary user does not exists and takies mandate frequency spectrum, but The probability that actually primary user exists；Limit accounts for probability and refers to that secondary user's mistaken verdict primary user exists and is not take up authorizing frequency Spectrum, but the actually non-existent probability of primary user.

A kind of double-threshold cooperative spectrum sensing method based on small sample energy measuring the most according to claim 1, it is special Levying and be, the multidimensional Gaussian approximation method in described step 5 refers to that divided by 2N, card side stochastic variable a is approximately average is 1- (9N)^-1And variance is (9N)^-1Multidimensional Gaussian random variable, i.e.

$F_{{\mathrm{\&chi;}}_{2N}^2}\left(a\right)\&ap;1-Q\left(\frac{\sqrt[3]{a/\left(2N\right)}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right).$

A kind of double-threshold cooperative spectrum sensing method based on small sample energy measuring the most according to claim 1, it is special Levying and be, in described step 6, collision probability and the limit of double-threshold cooperative spectrum sensing method based on small sample energy measuring account for The further expression formula of probability is respectively as follows:

$\begin{array}{c}{Q}_K^{cd}=\underset{m=K}{\overset{M}{\&Sigma;}}\left(\begin{array}{c}M\\ m\end{array}\right){\left({\tilde{P}}_{cd}\right)}^m\&times;{(1-{\tilde{P}}_{cd})}^{M-m}\\ =\underset{m=K}{\overset{M}{\&Sigma;}}\left(\begin{array}{c}M\\ m\end{array}\right){\left(1-Q\left(\frac{\sqrt[3]{{\mathrm{\&lambda;}}_0/{\mathrm{N\&sigma;}}_w^2({\mathrm{\&gamma;}}^{\&prime;}+1)}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right)\right)}^m\&times;{\left(Q\left(\frac{\sqrt[3]{{\mathrm{\&lambda;}}_0/{\mathrm{N\&sigma;}}_w^2({\mathrm{\&gamma;}}^{\&prime;}+1)}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right)\right)}^{M-m}\end{array}$

$\begin{array}{c}{Q}_K^{nd}=\underset{m=K}{\overset{M}{\&Sigma;}}\left(\begin{array}{c}M\\ m\end{array}\right){\left({\tilde{P}}_{nd}\right)}^m\&times;{(1-{\tilde{P}}_{nd})}^{M-m}\\ =\underset{m=K}{\overset{M}{\&Sigma;}}\left(\begin{array}{c}M\\ m\end{array}\right){\left(Q\left(\frac{\sqrt[3]{{\mathrm{\&lambda;}}_0/{\mathrm{N\&sigma;}}_w^2}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right)\right)}^m\&times;{\left(1-Q\left(\frac{\sqrt[3]{{\mathrm{\&lambda;}}_0/{\mathrm{N\&sigma;}}_w^2}-\&lsqb;1-{\left(9N\right)}^{-1}\&rsqb;}{{\left(9N\right)}^{-1/2}}\right)\right)}^{M-m}.\end{array}$