CN101753174A - Method and system for detecting self-adaption frequency spectrum - Google Patents

Abstract

The invention relates to a method and a system for detecting a self-adaption frequency spectrum. The method comprises the following steps of: detecting by a plurality of cognizing users according to a preset detecting time when the detection begins, sending a detecting message to a centre unit by the cognizing users; acquiring an accessing opportunity judging result of the detection at this time by the centre unit using a judging model which has a certain false dismissal probability and the smallest false-alarm probability, determining the detection starting time for the next detection of the cognizing users according to the detections at this time and at the last time, storing the accessing opportunity judging result of the detection at this time by the centre unit and sending the accessing opportunity judging result and the detection starting time for the next detection to the cognizing users. The method of the invention increases the accessing opportunity by setting the detection time by the judging model which has a certain false dismissal probability and the smallest false-alarm probability and by determining the detection starting time for the next detection of the cognizing users according to the judging result of the detections at this time and at the last time.

Description

Adaptive spectrum detection method and system

Technical field

The present invention relates to a kind of frequency spectrum detecting method and system, particularly a kind of adaptive spectrum detection method and system.

Background technology

Cognitive radio networks is by the perception external environment, and use artificial intelligence technology from environment, to learn, some operating parameter of real time altering (as through-put power, carrier power and modulation etc.) on purpose, make its internal state adapt to the statistics variations of the wireless signal that receives, thereby realize the high reliable communication in any time, any place, and the preferential radio spectrum resources of heterogeneous network environment is utilized efficiently.User in the cognitive radio networks is called cognitive user, buy the usufructuary user of frequency band and be called authorized user, the priority of authorized user service band is than cognitive user height, cognitive user need guarantee authorized user is disturbed less than finding idle frequency spectrum under the condition of certain level substantially, promptly finds to insert chance.Therefore, a key Design problem of cognitive radio networks is the frequency spectrum detection problem, be how cognitive user detects target band and find the problem of usable spectrum, the quality of frequency spectrum detection directly affects the service efficiency of idle frequency spectrum and the right of the normal service band of authorized user.

The prior art frequency spectrum detecting method mainly contains two classes: local frequency spectrum detection and collaboration frequency spectrum detect.Local frequency spectrum detection is meant that cognitive user independently detects frequency spectrum and finds to insert chance, and collaboration frequency spectrum detects and is meant that a plurality of cognitive user insert chance by certain mode cooperative detection frequency spectrum discovery.Local frequency spectrum detection is the basis that collaboration frequency spectrum detects, and collaboration frequency spectrum detects and solved the hidden terminal problem that exists in the local frequency spectrum detection, and the higher detection performance can be provided.

Local frequency spectrum detection mainly contains three classes: matched filter detection, energy measuring and cyclostationary characteristic detect.When authorization user signal was known, it was the better detecting method that matched filter detects.Matched filter detects the features such as pilot code, lead code or spreading code of utilizing authorization user signal, cognitive user can be carried out timing and carrier synchronization with authorized user continuously, even the channel equalization, thereby the signal that demodulate reception arrives is judged whether band occupancy of authorization user signal.But when authorization user signal was unknown, this method need be carried out various receiver algorithms, consumed lot of energy, and the complexity of enforcement is very high.It is the method that the signal statistics according to authorized user detects that cyclostationary characteristic detects.The signal statistics of authorized user generally has periodically, can regard the cyclo-stationary random process as, the spectrum correlation function of unlike signal takes on a different character on " frequency; cycle frequency; amplitude " three dimensions, the spectrum correlation function of authorization signal peak value occurs at different cycle frequencys and frequency place, and noise is to occur peak value at 0 o'clock in the cycle frequency value only, and cyclostationary characteristic detects can detect authorization user signal according to the position of peak value.Because the realization of this method needs twice Fourier transform, so the computational complexity height.Energy measuring is a kind of frequency spectrum detection mode commonly used, use band pass filter intercepting echo signal, after digital-to-analogue (A/D) conversion, sampling and calculating mean sample energy in detection time, obtain test statistics, test statistics and threshold are judged whether band occupancy of authorized user.This method amount of calculation is little, implements simply, and can detect under the condition of not knowing the authorization user signal type.

Because cognitive user is in subsystem, requires its detectability that better sensitivity is arranged.The position difference at each cognitive user place is subjected to the influence of multipath fading and shadow effect, and the computing capability of single cognitive user own is limited, so local detection performance is not high.Collaboration frequency spectrum detects based on local frequency spectrum detection information or result, carry out data fusion and decision-making, utilize the space diversity of the detection node formation of diverse geographic location in the cognitive radio networks, improved the global detection performance greatly, reduce the complexity of single cognitive user, can solve hidden terminal problem effectively, better robustness is arranged when integral body reply decline.Collaboration frequency spectrum detects and mainly contains two kinds of decision-making techniques: soft decision method and hard decision method.Soft decision method is meant that the cognitive user consideration factor (as position, fading channel etc.) relevant with decision-making sends detection information to fusion center, and the decision threshold that the fusion center utilization is set makes a policy.The hard decision method is meant that the result that cognitive user detects this locality sends to fusion center, and fusion center makes a policy according to local testing result.

In above-mentioned frequency spectrum detecting method, because the transmission of hard decision method is testing result, lost a large amount of information, so accuracy of detection is lower, reduced the access chance.Prior art is in order to address the above problem, usually use high specific merging criterion and equal gain combining criterion in hard decision or the soft decision, but the main enforcement purpose of high specific merging criterion and equal gain combining criterion is in order to detect authorized user, so art methods does not improve effectively and inserts chance.

Summary of the invention

The purpose of this invention is to provide a kind of adaptive spectrum detection method and system, can effectively improve frequency spectrum and insert chance.

To achieve these goals, the invention provides a kind of adaptive spectrum detection method, comprising:

Several cognitive user detect the zero hour according to detecting predefined detection time at this;

Described cognitive user sends to center cell with detection information;

Described center cell uses the access chance court verdict that false dismissal probability is certain, judgement model that minimize false alarm probability obtains this detection, determines the detection zero hour that cognitive user detects next time by the court verdict of this detection and preceding one-time detection;

The access chance court verdict of described this detection of described center cell storage, and will the access chance court verdict of described this detection and the detection of detection next time send to described cognitive user the zero hour.

Described false dismissal probability judgement model certain, that minimize false alarm probability is:

$\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}\frac{{\mathrm{\&gamma;}}_i}{1+{\mathrm{\&gamma;}}_i}{u}_i\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">H</figure-callout>}}_1\\ >\\ <\\ H_0\end{array}\mathrm{\&eta;}$

Wherein, threshold value η is:

$\mathrm{\&eta;}=Q^{-1}(1-a){\mathrm{\&sigma;}}_{c,{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">H</figure-callout>}}_1}+{\stackrel{\&OverBar;}{u}}_{c,H_1}$

In the formula, M is the number of cognitive user, γ _iBe the signal to noise ratio of i cognitive user, H ₀Be the hypothesis of the vacant frequency band of authorized user, H ₁Be the hypothesis of authorized user band occupancy, u _iBe that i cognitive user detects the test statistics that obtains, Q ^-1(.) is the inverse function of Q function, and a is predefined false dismissal probability value,

With

When being the authorized user band occupancy respectively, center cell place overall situation test statistics

Average and standard deviation.

Described court verdict by this detection and preceding one-time detection determines that the detection that cognitive user detects next time comprises the zero hour: when the access chance court verdict of the K time detection is the vacant frequency band of authorized user, and the detection T zero hour that cognitive user detects for the K+1 time _K+1=T _K+ j _KT, wherein T _KThe detection zero hour that the expression cognitive user detects for the K time, T represents the frame length of the MAC layer of cognitive user, j _KBe time interval between twice detection divided by the multiple of T, be 1≤j _KThe positive integer of≤D, D is that predefined maximum time is at interval divided by the multiple of T; When the access chance court verdict of the K time detection is the authorized user band occupancy, the detection T zero hour that cognitive user detects for the K+1 time _K+1=T _K+ T, wherein T _KThe detection zero hour that the expression cognitive user detects for the K time, T represents the frame length of the MAC layer of cognitive user.

On the technique scheme basis, the step of the judgement model specification cognitive user detection time that comprises also that center cell is certain according to false dismissal probability, minimize false alarm probability, described detection time is for making following access chance expression formula maximum:

$C=(1-\frac{T_s}{T})Q\left(\frac{Q^{-1}(1-a){\mathrm{\&sigma;}}_{c,H_1}+{\stackrel{\&OverBar;}{u}}_{c,H_1}-{\stackrel{\&OverBar;}{u}}_{c,H_0}}{{\mathrm{\&sigma;}}_{c,H_0}}\right)$

In the formula, C is for inserting chance, T _sBe the detection time of cognitive user, T is the frame length of the MAC layer of cognitive user,

With

When being the vacant frequency band of authorized user respectively, the average and the standard deviation of center cell place overall situation test statistics.

To achieve these goals, the present invention also provides a kind of adaptive spectrum detection system, comprising:

Cognitive user is used for detecting the zero hour according to detecting predefined detection time at this, and detection information is sent;

Center cell, be used to receive the detection information that cognitive user sends, use the access chance court verdict that false dismissal probability is certain, judgement model that minimize false alarm probability obtains this detection, court verdict by this detection and preceding one-time detection is determined the detection zero hour that cognitive user detects next time, the access chance court verdict of described this detection of storage, and will the access chance court verdict of described this detection and the detection of detection next time send to described cognitive user the zero hour.

Described center cell comprises:

Receiving element is connected with described cognitive user, is used to receive the detection information that cognitive user sends;

Judging unit, be connected with described receiving element, be used to use the access chance court verdict that false dismissal probability is certain, judgement model that minimize false alarm probability obtains this detection, determine the detection zero hour that cognitive user detects next time by the court verdict of this detection and preceding one-time detection;

Memory cell is connected with described judging unit, the access chance court verdict of one-time detection before being used to store;

Transmitting element is connected with described judging unit, is used for the access chance court verdict of described this detection and the detection that next time detects sends to described cognitive user the zero hour.

Described false dismissal probability judgement model certain, that minimize false alarm probability is:

$\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}\frac{{\mathrm{\&gamma;}}_i}{1+{\mathrm{\&gamma;}}_i}{u}_i\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">H</figure-callout>}}_1\\ >\\ <\\ H_0\end{array}\mathrm{\&eta;}$

Wherein, threshold value η is:

$\mathrm{\&eta;}=Q^{-1}(1-a){\mathrm{\&sigma;}}_{c,{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">H</figure-callout>}}_1}+{\stackrel{\&OverBar;}{u}}_{c,H_1}$

In the formula, M is the number of cognitive user, γ _iBe the signal to noise ratio of i cognitive user, H ₀Be the hypothesis of the vacant frequency band of authorized user, H ₁Be the hypothesis of authorized user band occupancy, u _iBe that i cognitive user detects the test statistics that obtains, Q ^-1(.) is the inverse function of Q function, and a is predefined false dismissal probability value, With

When being the authorized user band occupancy respectively, center cell place overall situation test statistics

Average and standard deviation.

Described judging unit comprises:

First judgment sub-unit is connected with described receiving element, is used to use the access chance court verdict that false dismissal probability is certain, judgement model that minimize false alarm probability obtains this detection;

Second judgment sub-unit, be connected with transmitting element with described first judgment sub-unit, memory cell respectively, be used for determining the detection zero hour that cognitive user detects next time by the court verdict of this detection and preceding one-time detection, comprise: when the access chance court verdict of the K time detection is the vacant frequency band of authorized user, the detection T zero hour that cognitive user detects for the K+1 time _K+1=T _K+ j _KT, wherein T _KThe detection zero hour that the expression cognitive user detects for the K time, T represents the frame length of the MAC layer of cognitive user, j _KBe time interval between twice detection divided by the multiple of T, be 1≤j _KThe positive integer of≤D, D is that predefined maximum time is at interval divided by the multiple of T; When the access chance court verdict of the K time detection is the authorized user band occupancy, the detection T zero hour that cognitive user detects for the K+1 time _K+1=T _K+ T, wherein T _KThe detection zero hour that the expression cognitive user detects for the K time, T represents the frame length of the MAC layer of cognitive user.

On the technique scheme basis, also comprise setup unit, described setup unit is connected with receiving element, judging unit and the transmitting element of described center cell, be used for certain according to false dismissal probability, as to minimize false alarm probability judgement model specification cognitive user detection time, described detection time is for making following access chance expression formula maximum:

$C=(1-\frac{T_s}{T})Q\left(\frac{Q^{-1}(1-a){\mathrm{\&sigma;}}_{c,H_1}+{\stackrel{\&OverBar;}{u}}_{c,H_1}-{\stackrel{\&OverBar;}{u}}_{c,H_0}}{{\mathrm{\&sigma;}}_{c,H_0}}\right)$

In the formula, C is for inserting chance, T _sBe the detection time of cognitive user, T is the frame length of the MAC layer of cognitive user,

With

When being the vacant frequency band of authorized user respectively, the average and the standard deviation of center cell place overall situation test statistics.Certain according to false dismissal probability, as to minimize the judgement model specification cognitive user of false alarm probability detection time.

The present invention proposes a kind of adaptive spectrum detection method and system, judgement model certain by false dismissal probability on the one hand, that minimize false alarm probability is provided with and increases the access chance detection time, determine the zero hour that cognitive user detects next time by the court verdict of this and preceding one-time detection on the other hand, make the time interval between twice detection obtain adjusting according to court verdict, by not detecting at the mac layer frame that has, being used for carrying out transfer of data detection time, therefore effectively increased the access chance.

Description of drawings

Fig. 1 is the flow chart of adaptive spectrum detection method of the present invention;

Fig. 2 is the structural representation of adaptive spectrum detection system of the present invention.

Embodiment

Below by drawings and Examples, technical scheme of the present invention is described in further detail.

Fig. 1 is the flow chart of adaptive spectrum detection method of the present invention, comprising:

Step 1, several cognitive user detect the zero hour according to detecting predefined detection time at this;

Step 2, described cognitive user send to center cell with detection information;

Step 3, described center cell use the access chance court verdict that false dismissal probability is certain, judgement model that minimize false alarm probability obtains this detection, determine the detection zero hour that cognitive user detects next time by the court verdict of this detection and preceding one-time detection;

The access chance court verdict of described this detection of step 4, described center cell storage, and will the access chance court verdict of described this detection and the detection of detection next time send to described cognitive user the zero hour.

In the technique scheme of the present invention, several cognitive user have been allocated in advance T detection time _S, each cognitive user is according at T detection time _SIn detect.When detecting for the K time, several cognitive user detect the T zero hour at the K time _KBegin to detect, and at T _K+ T _SDetect constantly and finish.After detection was finished, several cognitive user sent to center cell with detection information, and detection information comprises test statistics and signal to noise ratio.Center cell receives the detection information of each cognitive user, use false dismissal probability judgement model certain, that minimize false alarm probability to judge, obtain the access chance court verdict of the K time detection, it is whether band occupancy or the access chance is not arranged of authorized user, the court verdict that detects for the K-1 time of the center cell reading pre-stored is determined the detection T zero hour that cognitive user detects for the K+1 time by court verdict that the K-1 time is detected and the court verdict combination that detects for the K time simultaneously _K+1If the authorized user band occupancy shows and do not insert chance, then cognitive user finishes to detect (the T zero hour next time from the center cell judgement _K+ T _SConstantly～T _K+1Can not use this frequency band constantly); If authorized user is band occupancy not, showing has the access chance, and then cognitive user finishes to detect (the T zero hour next time from the center cell judgement _K+ T _SConstantly～T _K+1Can use this frequency band constantly).At last, center cell is stored the court verdict that detects the K time, and with the court verdict of the K time detection and the detection T zero hour of the K+1 time detection _K+1Send to cognitive user, make cognitive user next time according to the detection T zero hour that detects for the K+1 time _K+1Begin to detect.

In the technique scheme of the present invention, the local frequency spectrum detection of cognitive user adopts energy measuring.The energy measuring of single cognitive user is to adopt band pass filter intercepting echo signal, after digital-to-analogue (A/D) conversion, at T detection time _sInterior sampling N time and calculated population sampling energy obtain test statistics u _i(being described detection information).The problem of frequency spectrum detection can be regarded the binary hypothesis test problem as:

H ₀：Y[n]＝W[n] n＝0，1，......，N

H ₁：Y[n]＝X[n]+W[n] n＝0，1，......，N

In the formula, H ₀The hypothesis of the vacant frequency band of expression authorized user, H ₁The hypothesis of expression authorized user band occupancy, Y[n] sampling of expression received signal, W[n] the expression noisy sample, X[n] sampling of expression authorization user signal, n represents frequency in sampling, N represents the number of sampling.Suppose W[n] be Gauss's independent same distribution random process, W[n]: N (0, σ _w ²), σ _w ²The expression noise power; X[n] be the independent same distribution random process, average is 0, variance $E\left[\right|X{\left[n\right]}^2\left|\right]={\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="x"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">x</figure-callout>}}^2,$ σ _x ²The signal power of expression authorized user; X[n] and W[n] independent same distribution.

The test statistics u of i cognitive user energy measuring then _iFor:

$u_i=\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{\left(y\right[n\left]\right)}^2$

Define the signal to noise ratio of i cognitive user ${\mathrm{\&gamma;}}_i={\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="x"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">x</figure-callout>}}^2/{\mathrm{\&sigma;}}_w^2$ (γ _iUsually can obtain in advance, be known quantity), detection time T _s, sampling frequency f _s, the number of then sampling N=T _sf _sObvious test statistics u _iObey card side and distribute, then test statistics u _iProbability density function can be write as:

$f\left(u_i\right)=\left\{\begin{array}{cc}\frac{{(1/2)}^{N/2}}{\mathrm{\&Gamma;}(N/2)}{u_i}^{\frac{N}{2}-1}{e}^{-\frac{u_i}{2}}& H_0\\ \frac{1}{1+{\mathrm{\&gamma;}}_i}\frac{{(1/2)}^{N/2}}{\mathrm{\&Gamma;}(N/2)}{\left(\frac{u_i}{1+{\mathrm{\&gamma;}}_i}\right)}^{\frac{N_i-1}{2}}{e}^{-\frac{1}{2}\frac{u_i}{1+{\mathrm{\&gamma;}}_i}}& {\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">H</figure-callout>}}_1\end{array}\right.$

N is enough big when the sampling number, knows test statistics u by central limit law _iApproximate Gaussian distributed:

$u_i~\left\{\begin{array}{cc}N({\mathrm{N\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="w"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="w"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">w</figure-callout>\; </figure-callout>}}^2,{2\mathrm{N\&sigma;}}_w^4)& H_0\\ N((N+{\mathrm{\&gamma;}}_j){\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="w"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">w</figure-callout>}}^2,2(N+{2\mathrm{\&gamma;}}_j){\mathrm{\&sigma;}}_w^4)& {\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">H</figure-callout>}}_1\end{array}\right.$

If select the threshold value of i cognitive user for use is λ _i, then:

False alarm probability $P_f^i=P_r(u_i>{\mathrm{\&lambda;}}_i|H_0)$

Detection probability $P_d^i=P_r(u_i>{\mathrm{\&lambda;}}_i|H_1)$

False alarm probability P _f ⁱDuring the vacant frequency band of expression authorized user, cognitive user is thought the probability of authorized user band occupancy by detection, detection probability P _d ⁱDuring expression authorized user band occupancy, cognitive user detects the probability of authorized user band occupancy.

In the technique scheme of the present invention, center cell receives the test statistics u that cognitive user sends by Common Control Channel _i, use false dismissal probability judgement model certain, that minimize false alarm probability to judge, obtain to insert the court verdict of chance.False dismissal probability judgement specification of a model certain, that minimize false alarm probability is as follows:

At false dismissal probability P (D ₀| H ₁Under the condition of)=a (a is a definite value), make false alarm probability P (D ₁| H ₀) reaching minimum, this is a change integration problem that constraints is arranged, and solves necessary condition and should make target function type (1) reach minimum:

Q＝μP(D ₀|H ₁)+P(D ₁|H ₀)

In the formula, μ is a Lagrange's multiplier, D ₀Represent that detected result shows the vacant frequency band of authorized user, D ₁Represent that detected result shows the authorized user band occupancy, therefore, P (D ₀| H ₁) but detect the probability (false dismissal probability) of the vacant frequency band of authorized user, P (D during expression authorized user band occupancy ₁| H ₀) but detect the probability (false alarm probability) of authorized user band occupancy during the vacant frequency band of expression authorized user.

On the other hand, in the average risk expression formula of bayesian criterion, if make c ₀₀=c ₁₁=0, P (H ₀) c ₁₀=1, P (H ₁) c ₀₁=μ, then bayesian criterion should make average risk

R＝μP(D ₀|H ₁)+P(D ₁|H ₀)

Reach minimum.In the formula, c ₀₀Expression correctly detects the risk that inserts chance, c ₁₁Expression correctly detects the risk of authorization user signal, c ₁₀The risk of expression false-alarm, c ₀₁Risk is surveyed in the expression omission.Therefore target function is exactly as P (H ₀) c ₁₀=1, P (H ₁) c ₀₁Bayesian criterion under the=μ condition, then likelihood ratio threshold value:

∧ ₀＝P(H ₀)c ₁₀/P(H ₁)c ₀₁＝1/μ

So false dismissal probability judgement model certain, that minimize false alarm probability can be write as:

$\mathrm{LR}\left(u\right)\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">H</figure-callout>}}_1\\ >\\ <\\ H_0\end{array}\frac{1}{\mathrm{\&mu;}}$

Likelihood ratio:

$\mathrm{LR}\left(u\right)=\frac{P\left(u\right|H_1)}{P\left(u\right|H_0)}=\underset{i=1}{\overset{M}{\mathrm{\&Pi;}}}\frac{P\left(u_i\right|H_1)}{P\left(u_i\right|H_0)}={\left(\underset{i=1}{\overset{\mathrm{<figure-callout\; id="1"\; label="M"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">M</figure-callout>}}{\mathrm{\&Pi;}}}\frac{1}{1+{\mathrm{\&gamma;}}_i}\right)}^{\frac{N}{2}}{\mathrm{<figure-callout\; id="1"\; label="e"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">e</figure-callout>}}^{\frac{1}{2}{\mathrm{\&Sigma;}}_{i=1}^M\frac{{\mathrm{\&gamma;}}_i}{1+\mathrm{\&gamma;i}}{u}_i}$

Ground of equal value, false dismissal probability of the present invention judgement model certain, that minimize false alarm probability is:

$\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}\frac{{\mathrm{\&gamma;}}_i}{1+{\mathrm{\&gamma;}}_i}{u}_i\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">H</figure-callout>}}_1\\ >\\ <\\ H_0\end{array}\mathrm{\&eta;}---\left(1\right)$

In the formula, M represents the number of cognitive user.

In the technique scheme of the present invention, center cell receives the test statistics u that cognitive user sends by Common Control Channel _cAfter, obtain overall test statistics u by fusion _c, judge the access chance of frequency band afterwards again with threshold ratio.Wherein overall test statistics u _cFor:

$u_c=\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}{\mathrm{\&omega;}}_i{u}_i={\left(\mathrm{\&omega;}\right)}^{T}u---\left(2\right)$

In the formula, ${\mathrm{\&omega;}}_i=\frac{{\mathrm{\&gamma;}}_i}{1+{\mathrm{\&gamma;}}_i}$ Be the weight coefficient of i cognitive user test statistics, ω is the weight vectors of frequency band, and u is the test statistics vector of frequency band, ω=(ω ₁, ω ₂..., ω _M) ^T, u=(u ₁, u ₂..., u _M) ^T, M is the number of the cognitive user of detection frequency band.Because the linear superposition of Gaussian Profile remains Gaussian distributed, then average:

${\stackrel{\&OverBar;}{u}}_c=\left\{\begin{array}{cc}{\mathrm{NI}}^T\mathrm{\&omega;}{\mathrm{\&sigma;}}_w^2& H_0\\ {({\mathrm{NI}}^T+\mathrm{\&gamma;})}^T\mathrm{\&omega;}{\mathrm{\&sigma;}}_w^2& {\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">H</figure-callout>}}_1\end{array}\right.$

In the formula, vectorial I ^T=[1,1 ..., 1], signal to noise ratio vector γ=[γ ₁, γ ₂..., γ _M] ^TVariance:

${\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="c"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">c</figure-callout>}}^2=E{(u-{\stackrel{\&OverBar;}{u}}_c)}^2$

$={\mathrm{\&omega;}}^{T}E\left[(u-{\stackrel{\&OverBar;}{u}}_c){(u-{\stackrel{\&OverBar;}{u}}_c)}^T\right]\mathrm{\&omega;}$

Then:

${\mathrm{\&sigma;}}_{c,H_0}^2={\mathrm{\&omega;}}^{T}E\left[(u-{\stackrel{\&OverBar;}{u}}_{c,H_0}){(u-{\stackrel{\&OverBar;}{u}}_{c,H_0})}^T\right|H_0]\mathrm{\&omega;}$

$={\mathrm{\&omega;}}^T\left(2N{\mathrm{\&sigma;}}_w^{4}I\right)\mathrm{\&omega;}$

$={\mathrm{\&omega;}}^T\left({2T}_s{f}_s{\mathrm{\&sigma;}}_w^{4}I\right)\mathrm{\&omega;}$

${\mathrm{\&sigma;}}_{c,{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">H</figure-callout>}}_1}^2={\mathrm{\&omega;}}^{T}E\left[(u-{\stackrel{\&OverBar;}{u}}_{c,H_1}){(u-{\stackrel{\&OverBar;}{u}}_{c,H_1})}^T\right|H_1]\mathrm{\&omega;}$

$=2N{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="4"\; label="w"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">w</figure-callout>}}^4{\mathrm{\&omega;}}^T\mathrm{\&omega;}+4{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="4"\; label="w"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">w</figure-callout>}}^4{\mathrm{\&omega;}}^T\mathrm{diag}\left(\mathrm{\&gamma;}\right)\mathrm{\&omega;}$

$=2T_s{f}_s{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="4"\; label="w"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">w</figure-callout>}}^4{\mathrm{\&omega;}}^T\mathrm{\&omega;}+4{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="4"\; label="w"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">w</figure-callout>}}^4{\mathrm{\&omega;}}^T\mathrm{diag}\left(\mathrm{\&gamma;}\right)\mathrm{\&omega;}$

Then:

$P_f^c\left(T_S\right)=Q\left(\frac{\mathrm{\&eta;}-{\stackrel{\&OverBar;}{u}}_{c,H_0}}{{\mathrm{\&sigma;}}_{c,H_0}}\right)$

$P_d^c\left(T_s\right)=Q\left(\frac{\mathrm{\&eta;}-{\stackrel{\&OverBar;}{u}}_{c,{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">H</figure-callout>}}_1}}{{\mathrm{\&sigma;}}_{c,H_1}}\right)$

In the formula, P _f ^c(T _S) be the false alarm probability of fusion center, P _d ^c(T _s) be the detection probability of fusion center, η is the decision threshold of fusion center.

Because it is certain that the present invention sets false dismissal probability, so threshold value η should satisfy the constraints of false dismissal probability:

$a=1-P_d^c\left(T_s\right)=1-Q\left(\frac{\mathrm{\&eta;}-{\stackrel{\&OverBar;}{u}}_{c,{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">H</figure-callout>}}_1}}{{\mathrm{\&sigma;}}_{c,H_1}}\right)---\left(3\right)$

In the formula, a is predefined false dismissal probability value.

The threshold value at fusion center place then:

$\mathrm{\&eta;}=Q^{-1}(1-a){\mathrm{\&sigma;}}_{c,{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">H</figure-callout>}}_1}+{\stackrel{\&OverBar;}{u}}_{c,H_1}---\left(4\right)$

The false alarm probability at fusion center place:

$P_f^c\left(T_s\right)=Q\left(\frac{Q^{-1}(1-a){\mathrm{\&sigma;}}_{c,{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">H</figure-callout>}}_1}+{\stackrel{\&OverBar;}{u}}_{c,H_1}-{\stackrel{\&OverBar;}{u}}_{c,H_0}}{{\mathrm{\&sigma;}}_{c,H_0}}\right)---\left(5\right)$

Therefore, formula (1), (2) and (4) are the judgement model that false dismissal probability of the present invention is certain, minimize false alarm probability.In the practical application, center cell receives the test statistics u that M cognitive user sends _i, obtain overall test statistics u by formula (2) _c, obtain threshold value η by formula (4) then, judge according to formula (1) at last and insert chance.

On the technique scheme basis, also comprise before the step 1 and set the cognitive user step of detection time, that is: center cell, detection time of minimizing the judgement model specification cognitive user of false alarm probability certain, and the access chance maximum of setting that makes the every frame of MAC layer detection time according to false dismissal probability.

According to said process, the false alarm probability of center cell as can be known

$P_f^c\left(T_s\right)=Q\left(\frac{Q^{-1}(1-a){\mathrm{\&sigma;}}_{c,{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">H</figure-callout>}}_1}+{\stackrel{\&OverBar;}{u}}_{c,H_1}-{\stackrel{\&OverBar;}{u}}_{c,H_0}}{{\mathrm{\&sigma;}}_{c,H_0}}\right)---\left(6\right)$

Therefore, inserting chance can be expressed as:

$C=(1-\frac{T_s}{T})P_f^c\left(T_s\right)=(1-\frac{T_s}{T})Q\left(\frac{Q^{-1}(1-a){\mathrm{\&sigma;}}_{c,{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="H2009102425630E00011.png"\; state="\{\{state\}\}">H</figure-callout>}}_1}+{\stackrel{\&OverBar;}{u}}_{c,H_1}-{\stackrel{\&OverBar;}{u}}_{c,H_0}}{{\mathrm{\&sigma;}}_{c,H_0}}\right)---\left(7\right)$

In the formula, T represents the frame length of the MAC layer of cognitive user, and it is known that T is generally.

Usually, the system requirements false dismissal probability less than certain set point, therefore utilizes formula (3) and (5) just can obtain T detection time less than certain set point and false alarm probability _sScope, in the scope of detection time, T detection time of cognitive user is set rationally _sMake formula (7) maximum, can increase frequency spectrum and insert chance.In the practical application, the detection time in the time of can adopting multiple mode (as setting little time step) through type (7) to obtain to insert the chance maximum.

In technique scheme, the detection T zero hour that cognitive user detects for the K+1 time _K+1Relevant with the court verdict that court verdict and the K-1 time of the K time detection detect.Among the present invention, set the detection T zero hour that cognitive user detects for the K+1 time _K+1=T _K+ j _KT, wherein T _KThe detection zero hour that the expression cognitive user detects for the K time, T represents the frame length of the MAC layer of cognitive user, j _KBe time interval between twice detection divided by the multiple of T, be positive integer, j _KThe relevant variable of court verdict with the K time detection and the K-1 time detection.That is to say T _K+1-T _K=j _KT, i.e. the detection T zero hour that detects for the K+1 time _K+1The detection T zero hour with the K time detection _KBetween the time interval be the integral multiple of frame length of the MAC layer of cognitive user.According to existing detection experience, this time interval is set within the specific limits usually simultaneously, and therefore 1≤j is arranged _K≤ D, D are that predefined maximum time is at interval divided by the multiple of T.

Usually the mean down time of frequency band can obtain by statistical method, supposes that the mean down time is T _Off, because mean down time T _OffBe far longer than the long T of mac layer frame, those skilled in the art can reasonably be provided with the D value guaranteeing not influence under the normal prerequisite of using of authorized user.

For initialized the 1st detection case, set T ₁=0, j ₁=1, then cognitive user begins to detect immediately.After finishing the 1st detection, center cell obtains to insert the chance court verdict according to the 1st testing result, regardless of court verdict, and the detection T zero hour of the 2nd detection ₂The detection T zero hour with the 1st detection ₁Between the time interval be T, i.e. the detection T zero hour that detects of cognitive user the 2nd time ₂=T.If the authorized user band occupancy, then cognitive user is from T _SConstantly～T can not use this frequency band constantly; If authorized user is band occupancy not, then cognitive user is from T _SConstantly～T can use this frequency band constantly.

For the K time detection case, when cognitive user at T _KConstantly begin to detect and at T _K+ T _SConstantly finish detect for the K time after, center cell obtains the access chance court verdict of the K time detection, the detection T zero hour of the K+1 time detection of acquisition cognitive user according to the K time testing result _K+1Be divided into following several situation:

(1) when the access chance court verdict that detects for the K time is the authorized user band occupancy,, gets j regardless of court verdict last time _K=1, i.e. the detection T zero hour that detects for the K+1 time _K+1The detection T zero hour with the K time detection _KBetween the time interval be T, cognitive user T _K+1The detection T zero hour of inferior detection _K+1=T _K+ T; If the authorized user band occupancy, then cognitive user is from T _K+ T _SConstantly～T _K+1Can not use this frequency band constantly; If authorized user is band occupancy not, then cognitive user is from T _K+ T _SConstantly～T _K+1Can use this frequency band constantly.

(2) when the access chance court verdict that detects for the K time be the vacant frequency band of authorized user, and the access chance court verdict of the K-1 time detection is got j when also being the vacant frequency band of authorized user _K=j _K-1+ 1, promptly detect for the K+1 time and the time interval between the K time is detected detection zero hour has increased a frame length, cognitive user T than the time interval between detection zero hour of the K time detection and the K-1 time detection _K+1The detection T zero hour of inferior detection _K+1=T _K+ j _KT, the j that cognitive user is finished in detection _KT-T _SCan use this frequency band in time period.In the practical application, center cell is at the access chance court verdict that obtains the K time detection and obtain the K time j _KAfter, also to judge j _KWhether greater than predefined D value, if j _K≤ D, j _KCalculate the detection zero hour and the transmission that next time detects according to said process, if j _K＞D then sets J _K=j _K-1-1, calculate next time the detection zero hour and transmission then, so circulation is gone down.

(3) when the access chance court verdict that detects for the K time be the vacant frequency band of authorized user, and the access chance court verdict of the K-1 time detection is got j when being the authorized user band occupancy _K=2, i.e. the detection T zero hour that detects for the K+1 time _K+1The detection T zero hour with the K time detection _KBetween the time interval be 2T, cognitive user T _K+1The detection T zero hour of inferior detection _K+1=T _K+ 2T; If the authorized user band occupancy, the 2T-T that finishes in detection of cognitive user then _SCan not use this frequency band in time period; If authorized user band occupancy not, the 2T-T that finishes in detection of cognitive user then _SCan use this frequency band in time period.In fact, situation (3) is a kind of special case of situation (2), because the access chance court verdict that detects for the K-1 time is the authorized user band occupancy, therefore j is arranged _K-1=1, so j in the situation (3) _K=j _K-1+ 1=2.

From technique scheme as can be seen, one aspect of the present invention judgement model certain by false dismissal probability, that minimize false alarm probability is provided with and increases the access chance detection time, determine the zero hour that cognitive user detects next time by the court verdict of this and preceding one-time detection on the other hand, make the time interval between twice detection obtain adjusting according to court verdict, by not detecting at the mac layer frame that has, being used for carrying out transfer of data detection time, therefore effectively increased the access chance.

The present invention also provides a kind of adaptive spectrum detection system, as shown in Figure 2, structural representation for adaptive spectrum detection system of the present invention, this system comprises cognitive user and center cell, cognitive user is used for detecting the zero hour according to detecting predefined detection time at this, and detection information is sent; Center cell is used to receive the detection information that cognitive user sends, use the access chance court verdict that false dismissal probability is certain, judgement model that minimize false alarm probability obtains this detection, court verdict by this detection and preceding one-time detection is determined the detection zero hour that cognitive user detects next time, the access chance court verdict of described this detection of storage, and will the access chance court verdict of described this detection and the detection of detection next time send to described cognitive user the zero hour.Particularly, center cell of the present invention comprises receiving element, judging unit, memory cell and transmitting element, and receiving element is connected with described cognitive user, is used to receive the detection information that cognitive user sends; Judging unit is connected with described receiving element, be used to use the access chance court verdict that false dismissal probability is certain, judgement model that minimize false alarm probability obtains this detection, determine the detection zero hour that cognitive user detects next time by the court verdict of this detection and preceding one-time detection; Memory cell is connected with described judging unit, the access chance court verdict of one-time detection before being used to store; Transmitting element is connected with described judging unit, is used for the access chance court verdict of described this detection and the detection that next time detects sends to described cognitive user the zero hour.Further, judging unit comprises first judgment sub-unit and second judgment sub-unit, first judgment sub-unit is connected with described receiving element, is used to use the access chance court verdict that false dismissal probability is certain, judgement model that minimize false alarm probability obtains this detection; Second judgment sub-unit respectively with described first judgment sub-unit, memory cell is connected with transmitting element, be used to receive the access chance court verdict of this detection that first judgment sub-unit sends, the access chance court verdict of the preceding one-time detection of reading cells storage simultaneously, court verdict by this detection and preceding one-time detection is determined the detection zero hour that cognitive user detects next time, access chance court verdict with this detection sends to cell stores afterwards, with the access chance court verdict of this detection and the detection that next time detects send to transmitting element the zero hour.Wherein, court verdict by this detection and preceding one-time detection determines that the detection that cognitive user detects next time comprises the zero hour: when the access chance court verdict of the K time detection is the vacant frequency band of authorized user, and the detection T zero hour that cognitive user detects for the K+1 time _K+1=T _K+ j _KT, wherein T _KThe detection zero hour that the expression cognitive user detects for the K time, T represents the frame length of the MAC layer of cognitive user, j _KBeing the time interval multiple between twice detection, is 1≤j _KThe positive integer of≤D, D is that predefined maximum time is at interval divided by the multiple of T; When the access chance court verdict of the K time detection is the authorized user band occupancy, the detection T zero hour that cognitive user detects for the K+1 time _K+1=T _K+ T, wherein T _KThe detection zero hour that the expression cognitive user detects for the K time, T represents the frame length of the MAC layer of cognitive user.False dismissal probability in the adaptive spectrum detection system of the present invention contents such as judgement model certain, that minimize false alarm probability have illustrated in aforementioned adaptive spectrum detection method technical scheme, repeat no more.

One of ordinary skill in the art will appreciate that: all or part of step that realizes said method embodiment can be finished by the relevant hardware of program command, aforesaid program can be stored in the computer read/write memory medium, this program is carried out the step that comprises said method embodiment when carrying out; And aforesaid storage medium comprises: various media that can be program code stored such as ROM, RAM, magnetic disc or CD.

It should be noted that at last: above embodiment is only unrestricted in order to technical scheme of the present invention to be described, although the present invention is had been described in detail with reference to preferred embodiment, those of ordinary skill in the art is to be understood that, can make amendment or be equal to replacement technical scheme of the present invention, and not break away from the spirit and scope of technical solution of the present invention.

Claims (9)

1. an adaptive spectrum detection method is characterized in that, comprising:

Several cognitive user detect the zero hour according to detecting predefined detection time at this;

Described cognitive user sends to center cell with detection information;

Described center cell uses the access chance court verdict that false dismissal probability is certain, judgement model that minimize false alarm probability obtains this detection, determines the detection zero hour that cognitive user detects next time by the court verdict of this detection and preceding one-time detection;

The access chance court verdict of described this detection of described center cell storage, and will the access chance court verdict of described this detection and the detection of detection next time send to described cognitive user the zero hour.

2. adaptive spectrum detection method according to claim 1 is characterized in that, described false dismissal probability judgement model certain, that minimize false alarm probability is:

$\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}\frac{{\mathrm{\&gamma;}}_i}{1+{\mathrm{\&gamma;}}_i}{u}_i\begin{array}{c}{H}_1\\ >\\ <\\ H_0\end{array}\mathrm{\&eta;}$

Wherein, threshold value η is:

$\mathrm{\&eta;}=Q^{-1}(1-a){\mathrm{\&sigma;}}_{c,H_1}+{\stackrel{\&OverBar;}{u}}_{c,H_1}$

In the formula, M is the number of cognitive user, γ _iBe the signal to noise ratio of i cognitive user, H ₀Be the hypothesis of the vacant frequency band of authorized user, H ₁Be the hypothesis of authorized user band occupancy, u _iBe that i cognitive user detects the test statistics that obtains, Q ^-1(.) is the inverse function of Q function, and a is predefined false dismissal probability value,

With

When being the authorized user band occupancy respectively, center cell place overall situation test statistics

Average and standard deviation.

3. adaptive spectrum detection method according to claim 1, it is characterized in that, described court verdict by this detection and preceding one-time detection determines that the detection that cognitive user detects next time comprises the zero hour: when the access chance court verdict of the K time detection is the vacant frequency band of authorized user, and the detection T zero hour that cognitive user detects for the K+1 time _K+1=T _K+ j _KT, wherein T _KThe detection zero hour that the expression cognitive user detects for the K time, T represents the frame length of the MAC layer of cognitive user, j _KBe time interval between twice detection divided by the multiple of T, be 1≤j _KThe positive integer of≤D, D is that predefined maximum time is at interval divided by the multiple of T; When the access chance court verdict of the K time detection is the authorized user band occupancy, the detection T zero hour that cognitive user detects for the K+1 time _K+1=T _K+ T, wherein T _KThe detection zero hour that the expression cognitive user detects for the K time, T represents the frame length of the MAC layer of cognitive user.

4. according to arbitrary described adaptive spectrum detection method in the claim 1～3, it is characterized in that, the step of the judgement model specification cognitive user detection time that comprises also that center cell is certain according to false dismissal probability, minimize false alarm probability, described detection time is for making following access chance expression formula maximum:

$C=(1-\frac{T_s}{T})Q\left(\frac{Q^{-1}(1-a){\mathrm{\&sigma;}}_{c,H_1}+{\stackrel{\&OverBar;}{u}}_{c,H_1}-{\stackrel{\&OverBar;}{u}}_{c,H_0}}{{\mathrm{\&sigma;}}_{c,H_0}}\right)$

In the formula, C is for inserting chance, T _sBe the detection time of cognitive user, T is the frame length of the MAC layer of cognitive user,

With

When being the vacant frequency band of authorized user respectively, the average and the standard deviation of center cell place overall situation test statistics.

5. an adaptive spectrum detection system is characterized in that, comprising:

Cognitive user is used for detecting the zero hour according to detecting predefined detection time at this, and detection information is sent;

Center cell, be used to receive the detection information that cognitive user sends, use the access chance court verdict that false dismissal probability is certain, judgement model that minimize false alarm probability obtains this detection, court verdict by this detection and preceding one-time detection is determined the detection zero hour that cognitive user detects next time, the access chance court verdict of described this detection of storage, and will the access chance court verdict of described this detection and the detection of detection next time send to described cognitive user the zero hour.

6. adaptive spectrum detection system according to claim 5 is characterized in that, described center cell comprises:

Receiving element is connected with described cognitive user, is used to receive the detection information that cognitive user sends;

Judging unit, be connected with described receiving element, be used to use the access chance court verdict that false dismissal probability is certain, judgement model that minimize false alarm probability obtains this detection, determine the detection zero hour that cognitive user detects next time by the court verdict of this detection and preceding one-time detection;

Memory cell is connected with described judging unit, the access chance court verdict of one-time detection before being used to store;

Transmitting element is connected with described judging unit, is used for the access chance court verdict of described this detection and the detection that next time detects sends to described cognitive user the zero hour.

7. adaptive spectrum detection system according to claim 5 is characterized in that, described false dismissal probability judgement model certain, that minimize false alarm probability is:

$\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}\frac{{\mathrm{\&gamma;}}_i}{1+{\mathrm{\&gamma;}}_i}{u}_i\begin{array}{c}{H}_1\\ >\\ <\\ H_0\end{array}\mathrm{\&eta;}$

Wherein, threshold value η is:

$\mathrm{\&eta;}=Q^{-1}(1-a){\mathrm{\&sigma;}}_{c,H_1}+{\stackrel{\&OverBar;}{u}}_{c,H_1}$

In the formula, M is the number of cognitive user, γ _iBe the signal to noise ratio of i cognitive user, H ₀Be the hypothesis of the vacant frequency band of authorized user, H ₁Be the hypothesis of authorized user band occupancy, u _iBe that i cognitive user detects the test statistics that obtains, Q ^-1(.) is the inverse function of Q function, and a is predefined false dismissal probability value,

With

When being the authorized user band occupancy respectively, center cell place overall situation test statistics

Average and standard deviation.

8. adaptive spectrum detection system according to claim 6 is characterized in that, described judging unit comprises:

First judgment sub-unit is connected with described receiving element, is used to use the access chance court verdict that false dismissal probability is certain, judgement model that minimize false alarm probability obtains this detection;

Second judgment sub-unit, be connected with transmitting element with described first judgment sub-unit, memory cell respectively, be used for determining the detection zero hour that cognitive user detects next time by the court verdict of this detection and preceding one-time detection, comprise: when the access chance court verdict of the K time detection is the vacant frequency band of authorized user, the detection T zero hour that cognitive user detects for the K+1 time _K+1=T _K+ j _KT, wherein T _KThe detection zero hour that the expression cognitive user detects for the K time, T represents the frame length of the MAC layer of cognitive user, j _KBe time interval between twice detection divided by the multiple of T, be 1≤j _KThe positive integer of≤D, D is that predefined maximum time is at interval divided by the multiple of T; When the access chance court verdict of the K time detection is the authorized user band occupancy, the detection T zero hour that cognitive user detects for the K+1 time _K+1=T _K+ T, wherein T _KThe detection zero hour that the expression cognitive user detects for the K time, T represents the frame length of the MAC layer of cognitive user.

9. according to the arbitrary described adaptive spectrum detection system of claim 5～8, it is characterized in that, also comprise setup unit, described setup unit is connected with receiving element, judging unit and the transmitting element of described center cell, be used for certain according to false dismissal probability, as to minimize false alarm probability judgement model specification cognitive user detection time, described detection time is for making following access chance expression formula maximum:

$C=(1-\frac{T_s}{T})Q\left(\frac{Q^{-1}(1-a){\mathrm{\&sigma;}}_{c,H_1}+{\stackrel{\&OverBar;}{u}}_{c,H_1}-{\stackrel{\&OverBar;}{u}}_{c,H_0}}{{\mathrm{\&sigma;}}_{c,H_0}}\right)$

In the formula, C is for inserting chance, T _sBe the detection time of cognitive user, T is the frame length of the MAC layer of cognitive user,

With

When being the vacant frequency band of authorized user respectively, the average and the standard deviation of center cell place overall situation test statistics.Certain according to false dismissal probability, as to minimize the judgement model specification cognitive user of false alarm probability detection time.

Images