Disclosure of Invention
The technical problem to be solved by the present invention is to provide a multi-band cooperative spectrum sensing method based on clustering optimization, which can adapt to the energy change of signals received from users, improve the detection performance of secondary users, reduce the computation complexity of a spectrum sensing fusion center, and improve the cooperative detection efficiency, aiming at the prior art.
The technical scheme adopted by the invention for solving the technical problems is as follows: a multi-band cooperative spectrum sensing method based on clustering optimization is used for spectrum sensing fusion center and N secondary users with spectrum sensing function to perform spectrum detection, and is characterized by sequentially comprising the following steps:
(1) establishing a cooperative sensing model consisting of a spectrum sensing fusion center, N secondary users and authorized users; the spectrum sensing fusion center is marked as FC, and N secondary users are respectively marked as CRi(i ═ 1,2, …, N ≧ 3), the authorized user is denoted as PU;
(2) n sub-users CRiRespectively and independently acquiring SNRiAnd respectively obtaining SNR of the signal to noise ratios obtained by the two sensorsiSending the signal to a spectrum sensing fusion center FC for clustering;
(3) presetting the SNR thresholds of M clusters according to the ascending sequence of the SNR thresholdsWall,m(M-1, 2, …, M and 0.5N. ltoreq.M<N), spectrum sensing fusion center FC distributes each user CRiself-SNR sentiSNR respectively corresponding to M signal-to-noise ratio thresholdsWall,mJudging and comparing to obtain M1Each cluster containing secondary users, and the obtained cluster is marked as Cl,l=1,2,…,M1,1<M1≤M,SNRWall,1<SNRWall,2<…<SNRWall,M(ii) a Frequency spectrum perception fusion center FC for each user CRiSNRiWith each signal-to-noise ratio threshold SNRWall,mThe judgment and comparison process of (2) is as follows:
(3-1) according to MIndividual cluster SNR thresholdWall,mSetting M +1 clustering SNR intervals as [ - ∞, SNRWall,1)、[SNRWall,1,SNRWall,2)、…、[SNRWall,M-1,SNRWall,M) And [ SNRWall,MInfinity), wherein the signal-to-noise ratio of the secondary users located within the first cluster is at [ - ∞, SNRWall,1) Within the cluster SNR interval, the SNR of the secondary users in the second cluster is at [ SNR ]Wall,1,SNRWall,2) Within the cluster SNR interval, and so on, the SNR of the secondary user in the Mth cluster is [ SNR ]Wall,M-1,SNRWall,M) Within the cluster SNR interval, the SNR of the secondary users in the M +1 th cluster is [ SNR ]Wall,MAnd ∞) within a clustering signal-to-noise ratio interval;
(3-2) the spectrum sensing fusion center FC respectively leads each user CRiSNRiWith M signal-to-noise ratio thresholds SNRWall,mComparing to determine the SNRiThe cluster signal-to-noise ratio interval section is located; wherein:
when the signal-to-noise ratio SNRiThe cluster SNR interval is [ - ∞, SNRWall,1) Then the SNR is not giveniThe corresponding secondary user participates in the cooperative perception; if the signal-to-noise ratio SNRiThe cluster signal-to-noise ratio interval is [ SNR ]Wall,MInfinity), then the SNR will beiThe corresponding secondary user is placed in the Mth cluster;
(4) at M1In each cluster containing the secondary users, selecting the secondary user with the largest signal-to-noise ratio as the cluster primary user in the cluster according to the sequence of the signal-to-noise ratios of the secondary users from large to small, thereby obtaining M1Individual cluster first-time users;
(5) in a second cluster containing secondary users, the cluster primary user is used as a fusion center of the cluster, and spectrum sensing results of other secondary users in the cluster are received and fused to obtain a cooperative detection result of the cluster; wherein the cooperative detection process in the cluster comprises the following steps (5-1) to (5-3):
(5-1) setting K sub-users CR in the second clusterk(K-1, 2, …, K), K sub-users CRkRespectively carrying out spectrum sensing based on energy and independently acquiring signal-to-noise ratio SNRkAnd respectively obtaining SNRkAnd sending the spectrum sensing result to the cluster primary user CR1(ii) a Wherein the spectrum sensing result comprises a secondary user CRkIs detected with probability Pd,kAnd false alarm probability Pf,k;
(5-2) Cluster first user CR1Receiving other K-1 secondary users CRkSNR of transmitted signal to noise ratiokSumming the spectrum sensing results and judging the SNRkSNR larger than preset SNR screening valuechoseIf so, selecting the secondary user corresponding to the signal-to-noise ratio as a member of the cognitive group participating in cooperative detection, and executing the step (5-3); otherwise, selecting the spectrum sensing result corresponding to the secondary user with the highest signal-to-noise ratio as the cluster primary user CR1The final detection result of (1);
(5-3) Cluster Primary user CR1Performing self-adaptive sensing fusion according to the spectrum sensing result of the selected cognitive group members participating in the cooperation; the adaptive perception fusion process comprises the following steps (5-31) to (5-33):
(5-31) Cluster first user CR1According to the spectrum sensing result sent by K-1 secondary users, counting the number m (m is more than or equal to 1 and less than or equal to K-1) of the secondary users sensing the PU spectrum of the authorized user to be in an occupied state and the number K-1-m of the secondary users sensing the PU spectrum of the authorized user to be in an idle state in the K-1 secondary users; the PU frequency spectrum of the authorized user is recorded as H in the occupied state1Authorizing the user PU1The spectrum is in idle state and is marked as H0;
(5-32) Cluster first user CR1According to the signal-to-noise ratio sent by K-1 secondary users, calculating the PU frequency spectrum of m perceived authorized users as an occupied state H1Sub-user integrity factor k1,jAnd K-1-m sensing authorized users PU frequency spectrum to be in idle state H0Sub-user integrity factor k2,t(ii) a Wherein the integrity factor k1,jAnd kappa2,tThe calculation formula of (a) is as follows:
(5-33) Cluster first user CR1According to the respective sensing results of m secondary users and the integrity coefficient k1,jRespectively calculating the PU frequency spectrum of the authorized user as the occupation state H1Average detection probability ofGlobal detection probabilityAnd this occupation state H1Corresponding global miss probability Dundet,H1And authorizing the spectrum of the user PU to be in an idle state H0Average detection probability ofGlobal detection probabilityThis idle state H0Corresponding global miss probabilityAnd global false alarm probabilityWherein the process comprises the following steps (a) to (g):
(a) establishing global error detection probability P of m secondary user cooperative perceptioneObtaining an energy detection optimization function gamma with respect to a decision threshold*And an optimum threshold value gamma for energy detectionoptAnd calculating the PU frequency spectrum of the authorized user as an occupation state H1Average detection probability ofWherein,
global error detection probability P of m secondary user cooperative perceptioneThe calculation formula is as follows:
wherein,PU frequency spectrum is in idle state H for authorized user0The probability of (a) of (b) being,PU frequency spectrum is in occupied state H for authorized user1The probability of (d); pfIs the global false alarm probability, PdFor global detection probability, PmIs the global miss probability;PU frequency spectrum is in occupied state H for corresponding authorized user1The average signal-to-noise ratio of the m secondary users, wherein,snrito the secondary user CRiSignal-to-noise ratio of itself, q (z) represents a normal gaussian complementary integration function;
energy detection optimization function gamma with respect to decision threshold*Is defined as:
optimal threshold value gamma for energy detectionoptComprises the following steps:
authorizing user PU frequency spectrum to be in occupied state H1Average detection probability ofThe calculation formula is as follows:
(b) according to the obtained PU frequency spectrum of the authorized user as an occupation state H1Average detection probability ofAnd the integrity factor k of m secondary users1,jCalculating the PU frequency spectrum of the authorized user as the occupation state H1Global detection probability ofAnd this occupation state H1Corresponding global miss probabilityWherein the global detection probabilityAnd global miss probabilityThe calculation formula is as follows:
(c) the PU frequency spectrum of the authorized user is idleState H0Average detection probability ofAnd the integrity factor K of K-1-m sub-users2,tCalculating the PU frequency spectrum of the authorized user to be in an idle state H0Global detection probability ofAnd this idle state H0Corresponding global miss probabilityGlobal false alarm probabilityWherein the average detection probabilityGlobal detection probabilityGlobal miss probabilityAnd global false alarm probabilityThe calculation formulas of (A) are respectively as follows:
(d) cluster headSecondary user CR1According to the PU frequency spectrum of the authorized user as the occupied state H1Corresponding global miss probabilityAnd authorizing the user PU frequency spectrum to be in an idle state H0Corresponding global false alarm probabilityEstablishing a spectrum sensing error function Fun (m) based on the number of the secondary users; the spectral perception error function fun (m) is calculated as follows:
wherein, PpuRepresenting the probability of the PU signal of the authorized user appearing in its authorized spectrum;
(e) calculating a minimum value Fun (m) of the spectrum sensing error function Fun (m)0) And using the minimum value Fun (m) of the spectrum sensing error function0) Corresponding number m0(m0M) is used as the optimal number of cooperative sub-users participating in cooperative sensing, and the m sub-users are subjected to signal-to-noise ratio snr according to the corresponding signal-to-noise ratio snriPerforming descending order arrangement to obtain descending order arrangement groups of m sub-users;
(f) selecting the first m in the descending order arrangement group of the secondary users0The individual secondary user is used as the optimal cooperative secondary user participating in cooperative perception; wherein the selected best cooperative secondary user is marked as CR'rWherein r is 1,2, …, m0;
(g) Cluster first user CR1According to m in step (f)0Performing cooperative sensing based on an OR criterion on the frequency spectrum sensing result of the optimal cooperative secondary user, and taking the detection result of the cooperative sensing as the final detection result of the K secondary users in the cluster; wherein the OR criterion is as follows:
wherein, Pd,rFor optimal cooperative sub-users CR within this cluster "rProbability of detection of, Pfa,jFor optimal cooperative sub-users CR within this cluster "rFalse alarm probability of (d); qd,1Is the global detection probability, Q, after the cooperative detection of the clusterfa,1The global false alarm probability after the cluster cooperative detection is obtained; omegarRepresenting the signal-to-noise ratio SNR "rCoefficient of weight, SNR "maxRepresents m in the present cluster0Maximum signal-to-noise ratio, SNR, of the best cooperative sub-user "minRepresents m in the present cluster0The minimum signal-to-noise ratio of the best cooperative secondary user;
(6) respectively acquiring a third cluster to an Mth cluster according to the process of the step (5)1Intra-cluster global detection probability Q within a clusterd,3ToAnd a global false alarm probability Qfa,2To
(7) Spectrum sensing fusion center FC according to M1Global detection probability Q in corresponding cluster sent by first user of each clusterd,sAnd global false alarm probability Qfa,sPerforming fusion detection based on an AND criterion, AND taking a fusion detection result as a final multi-band cooperative spectrum detection result; the AND criterion is as follows:
wherein Q isdFor global detection probability after cooperation, QfaIs the global false alarm probability after cooperation.
Compared with the prior art, the invention has the advantages that: clustering secondary users by the spectrum sensing fusion center according to the self signal-to-noise ratio of each secondary user and a preset clustering signal-to-noise ratio threshold, selecting the secondary user with the maximum signal-to-noise ratio as a cluster primary user corresponding to the clustering in each cluster containing the secondary users, taking the cluster primary user as the fusion center corresponding to the cluster, and adaptively adjusting and acquiring the optimal threshold of energy detection by the secondary users so as to adapt to the requirement of dynamic change of signal energy received by the secondary users and improve the energy detection probability of the secondary users; then, the primary cluster user fuses the spectrum sensing results of other secondary users in the cluster, so that the calculation amount of the spectrum sensing fusion center for fusing the detection results of all the secondary users in the traditional cooperative sensing method is reduced, and the storage space of the spectrum sensing fusion center is saved; and then the spectrum sensing fusion center performs fusion detection according to the corresponding intra-cluster global detection probability and the global false alarm probability sent by each cluster primary user, and takes the fusion detection result as a final multi-band cooperative spectrum detection result. The clustering multi-band cooperative spectrum sensing method can adapt to the energy change of signals received from users, improve the detection performance of secondary users, reduce the calculation complexity of a spectrum sensing fusion center and improve the cooperative detection efficiency.
Detailed Description
The invention is described in further detail below with reference to the accompanying examples.
In order to realize that a spectrum sensing fusion center FC and N secondary users with a spectrum sensing function detect the spectrum situation of multiple frequency bands, as shown in fig. 1, the method for sensing a multi-frequency band cooperative spectrum based on cluster optimization in this embodiment sequentially includes the following steps:
(1) establishing a cooperative sensing model consisting of a spectrum sensing fusion center FC, N secondary users and authorized users; wherein, N sub-users are respectively marked as CRi(i ═ 1,2, …, N ≧ 3), the authorized user is denoted as PU;
(2) n sub-users CRiRespectively and independently acquiring SNRiAnd respectively obtaining SNR of the signal to noise ratios obtained by the two sensorsiSending the signal to a spectrum sensing fusion center FC for clustering; for example, the secondary user CR1The self signal-to-noise ratio obtained independently is SNR1Sub-user CR2The self signal-to-noise ratio obtained independently is SNR3;
(3) Presetting the SNR thresholds of M clusters according to the ascending sequence of the SNR thresholdsWall,m(M-1, 2, …, M and 0.5N. ltoreq.M<N), that is, the SNR thresholds in the M preset clusters are respectively SNRWall,1、SNRWall,2… and SNRWall,MThe spectrum sensing fusion center FC is used for enabling each user CRiself-SNR sentiSNR respectively corresponding to M signal-to-noise ratio thresholdsWall,mJudging and comparing to obtain M1Each cluster containing secondary users, and the obtained cluster is marked as Cl,l=1,2,…,M1,1<M1≤M,SNRWall,1<SNRWall,2<…<SNRWall,M;
For example, the secondary user CR1Is SNR1Respectively with signal-to-noise ratio threshold SNRWall,1To SNRWall,MMake a size judgment and comparison, and then make the secondary user CR2Is SNR1Respectively with signal-to-noise ratio threshold SNRWall,1To SNRWall,MMaking size judgment and comparison, analogizing in turn, and finally making secondary user CRNIs SNRNRespectively with signal-to-noise ratio threshold SNRWall,1To SNRWall,MJudging and comparing the sizes;
wherein, the spectrum sensing fusion center FC is used for each user CRiSNRiWith each signal-to-noise ratio threshold SNRWall,mThe judgment and comparison process of (2) is as follows:
(3-1) SNR threshold based on M clustersWall,mSetting M +1 clustering SNR intervals as [ - ∞, SNRWall,1)、[SNRWall,1,SNRWall,2)、…、[SNRWall,M-1,SNRWall,M) And [ SNRWall,MInfinity), wherein the signal-to-noise ratio of the secondary users located within the first cluster is at [ - ∞, SNRWall,1) Within the cluster SNR interval, the SNR of the secondary users in the second cluster is at [ SNR ]Wall,1,SNRWall,2) Within the cluster SNR interval, and so on, the SNR of the secondary user in the Mth cluster is [ SNR ]Wall,M-1,SNRWall,M) Within the cluster SNR interval, the SNR of the secondary users in the M +1 th cluster is [ SNR ]Wall,MAnd ∞) within a clustering signal-to-noise ratio interval;
for example, now five SNR thresholds are set to SNR respectivelyWall,1=1dB、SNRWall,2=3dB、SNRWall,3=5dB、SNRWall,4=8dB、SNRWall,511dB, the sub-user SNR in the first cluster is in the [ - ∞,1dB) cluster SNR period, the sub-user SNR in the second cluster is in the [1dB,3dB) cluster SNR period, the sub-user SNR in the third cluster is in the [3dB,5dB) cluster SNR period, the sub-user SNR in the fourth cluster is in the [5dB,8dB) cluster SNR period, the sub-user SNR in the fifth cluster is in the [8dB,11dB) cluster SNR period, and the sub-user SNR in the sixth cluster is in the [11dB, infinity) cluster SNR period;
(3-2) the spectrum sensing fusion center FC respectively leads each user CRiSNRiWith M signal-to-noise ratio thresholds SNRWall,mComparing to determine the SNRiThe cluster signal-to-noise ratio interval section is located; wherein:
when the signal-to-noise ratio SNRiThe cluster SNR interval is [ - ∞, SNRWall,1) Then the SNR is not giveniThe corresponding secondary user participates in the cooperative perception; if the signal-to-noise ratio SNRiThe cluster signal-to-noise ratio interval is [ SNR ]Wall,MInfinity), indicating that the sub-user corresponding to the SNR has very good detection performance, the SNR of the SNR is determinediThe corresponding secondary users are placed in the Mth cluster, so that the number of clusters can be reduced, the operation speed is improved, and the overall cooperative sensing performance in the Mth cluster can be improved;
for example, now five SNR thresholds are set to SNR respectivelyWall,1=1dB、SNRWall,2=3dB、SNRWall,3=5dB、SNRWall,4=10dB、SNRWall,511dB, six sub-users CR1To CR6Corresponding self signal-to-noise ratios are SNR respectively1=-1dB、SNR2=1.5dB、SNR3=2dB、SNR4=6dB、SNR5=7dB、SNR614 dB; by comparison, SNR is known1Within the [ - ∞,1dB) clustering SNR interval, the secondary user CR is not granted1Participating in cooperative sensing; due to SNR6Within a [11dB, ∞) clustering signal-to-noise ratio interval, the secondary user CR6Placing the cluster in a cluster corresponding to the cluster signal-to-noise ratio interval section of [10dB,11 dB);
(4) at M1In each cluster containing the secondary users, selecting the secondary user with the largest signal-to-noise ratio as the cluster primary user in the cluster according to the sequence of the signal-to-noise ratios of the secondary users from large to small, thereby obtaining M1Individual cluster first-time users;
(5) in a second cluster containing secondary users, the cluster primary user is used as a fusion center of the cluster, and spectrum sensing results of other secondary users in the cluster are received and fused to obtain a cooperative detection result of the cluster;
the cluster primary users in each cluster are used as the fusion center of the cluster, so that the fusion calculation amount of the spectrum sensing fusion center FC on the detection results of all secondary users can be reduced, the storage space of the spectrum sensing fusion center FC is saved, and each cluster can independently perform cooperative sensing, so that the cooperative detection time of all secondary users is effectively prolonged, the requirement of cognitive radio on the spectrum switching efficiency of the secondary users is met, and the interference of the secondary users on the frequency band occupied by authorized users is avoided; wherein the cooperative detection process in the cluster comprises the following steps (5-1) to (5-3):
(5-1) setting K sub-users CR in the second clusterk(K-1, 2, …, K), K sub-users CRkRespectively carrying out spectrum sensing based on energy and independently acquiring signal-to-noise ratio SNRkAnd respectively obtaining SNRkAnd sending the spectrum sensing result to the cluster primary user CR1(ii) a Wherein the spectrum sensing result comprises a secondary user CRkIs detected with probability Pd,kAnd false alarm probability Pf,k;
(5-2) Cluster first user CR1Receiving other K-1 secondary users CRkSNR of transmitted signal to noise ratiokSumming the spectrum sensing results and judging the SNRkSNR larger than preset SNR screening valuechoseIf so, selecting the secondary user corresponding to the signal-to-noise ratio as a member of the cognitive group participating in cooperative detection, and executing the step (5-3); otherwise, selecting the spectrum sensing result corresponding to the secondary user with the highest signal-to-noise ratio as the cluster primary user CR1The final detection result of (1);
wherein, in the step (5-2), the SNR selection value is presetchoseThe reason is that, among secondary users participating in cooperative sensing, if there is a secondary user (also referred to as a "bad user") with a lower signal-to-noise ratio, the detection accuracy of the "bad user" is extremely low, and once participating in cooperative sensing, the detection probability of the overall cooperative sensing by the spectrum sensing fusion center FC is lowered, thereby reducing the sensing efficiency. Therefore, in cooperative sensing, it is necessary to set a signal-to-noise threshold to match these "Bad users are "rejected.
(5-3) Cluster Primary user CR1Performing self-adaptive sensing fusion according to the spectrum sensing result of the selected cognitive group members participating in the cooperation; the adaptive perception fusion process comprises the following steps (5-31) to (5-33):
(5-31) Cluster first user CR1According to the spectrum sensing result sent by K-1 secondary users, counting the number m (m is more than or equal to 1 and less than or equal to K-1) of the secondary users sensing the PU spectrum of the authorized user to be in an occupied state and the number K-1-m of the secondary users sensing the PU spectrum of the authorized user to be in an idle state in the K-1 secondary users; the PU frequency spectrum of the authorized user is recorded as H in the occupied state1Authorizing the user PU1The spectrum is in idle state and is marked as H0;
(5-32) Cluster first user CR1According to the signal-to-noise ratio sent by K-1 secondary users, calculating the PU frequency spectrum of m perceived authorized users as an occupied state H1Sub-user integrity factor k1,jAnd K-1-m sensing authorized users PU frequency spectrum to be in idle state H0Sub-user integrity factor k2,t(ii) a The integrity coefficient represents the credibility of detection made by the corresponding secondary user and also represents the detection capability of the secondary user; the higher the integrity coefficient is, the higher the detection probability of the corresponding secondary user is; wherein the integrity factor k1,jAnd kappa2,tThe calculation formula of (a) is as follows:
(5-33) Cluster first user CR1According to the respective sensing results of m secondary users and the integrity coefficient k1,jRespectively calculating the PU frequency spectrum of the authorized user as the occupation state H1Average detection probability ofGlobal detection probabilityAnd this occupation state H1Corresponding global miss probabilityAnd authorizing the frequency spectrum of the user PU to be in an idle state H0Average detection probability ofGlobal detection probabilityThis idle state H0Corresponding global miss probabilityAnd global false alarm probabilityWherein the process comprises the following steps (a) to (g):
(a) establishing global error detection probability P of m secondary user cooperative perceptioneObtaining an energy detection optimization function gamma with respect to a decision threshold*And an optimum threshold value gamma for energy detectionoptAnd calculating the PU frequency spectrum of the authorized user as an occupation state H1Average detection probability ofWherein,
global error detection probability P of m secondary user cooperative perceptioneThe calculation formula is as follows:
wherein,PU frequency spectrum is in idle state H for authorized user0The probability of (a) of (b) being,PU frequency spectrum is in occupied state H for authorized user1The probability of (d); pfIs the global false alarm probability, PdFor global detection probability, PmIs the global miss probability;PU frequency spectrum is in occupied state H for corresponding authorized user1The average signal-to-noise ratio of the m secondary users, wherein,snrito the secondary user CRiSignal-to-noise ratio of itself, q (z) represents a normal gaussian complementary integration function;
energy detection optimization function gamma with respect to decision threshold*Is defined as:
by optimizing the function gamma for energy detection with respect to decision thresholds*Extremizing to obtain optimal threshold value gamma for energy detectionoptComprises the following steps:
that is, in the process of detecting the energy utilized by each user, when the decision threshold value aiming at the signal energy is taken as gammaoptIn time, the secondary user can accurately detect the existence of the received signal and adapt to the change condition of the energy of the signal received by the secondary user, so that the accuracy of the secondary user based on energy detection is improved;
authorizing user PU frequency spectrum to be in occupied state H1Average detection probability ofThe calculation formula is as follows:
(b) according to the obtained PU frequency spectrum of the authorized user as an occupation state H1Average detection probability ofAnd the integrity factor k of m secondary users1,jCalculating the PU frequency spectrum of the authorized user as the occupation state H1Global detection probability ofAnd this occupation state H1Corresponding global miss probabilityWherein the global detection probabilityAnd global miss probabilityThe calculation formula is as follows:
(c) according to the obtained PU frequency spectrum of the authorized user, the PU frequency spectrum is in an idle state H0Average detection probability ofAnd the integrity factor K of K-1-m sub-users2,tCalculating the PU frequency spectrum of the authorized user to be emptyIdle state H0Global detection probability ofAnd this idle state H0Corresponding global miss probabilityGlobal false alarm probabilityWherein the average detection probabilityGlobal detection probabilityGlobal miss probabilityAnd global false alarm probabilityThe calculation formulas of (A) are respectively as follows:
(d) cluster first user CR1According to the PU frequency spectrum of the authorized user as the occupied state H1Corresponding global miss probabilityAnd authorized user PU spectrumIs in an idle state H0Corresponding global false alarm probabilityEstablishing a spectrum sensing error function Fun (m) based on the number of the secondary users; the spectrum sensing error function fun (m) represents the error condition of spectrum sensing when the number of secondary users is m; the spectral perception error function fun (m) is calculated as follows:
wherein, PpuRepresenting the probability of the PU signal of the authorized user appearing in its authorized spectrum;
(e) calculating a minimum value Fun (m) of the spectrum sensing error function Fun (m)0) And using the minimum value Fun (m) of the spectrum sensing error function0) Corresponding number m0(m0M) is used as the optimal number of cooperative sub-users participating in cooperative sensing, and the m sub-users are subjected to signal-to-noise ratio snr according to the corresponding signal-to-noise ratio snriPerforming descending order arrangement to obtain descending order arrangement groups of m sub-users;
wherein, the number of the secondary users participating in the cooperative sensing is m0Then, the cooperative sensing of the secondary users in the cluster has the minimum spectrum sensing error, and the detection performance corresponding to the cooperative spectrum sensing is stronger; because the signal-to-noise ratio of each secondary user is still the key for influencing the spectrum detection performance of each secondary user, the performance of each secondary user after sequencing can be conveniently compared by performing descending sequencing according to the signal-to-noise ratio value so as to select the secondary user with high detection performance;
(f) selecting the first m in the descending order arrangement group of the secondary users0The individual secondary user is used as the optimal cooperative secondary user participating in cooperative perception; wherein the selected best cooperative secondary user is marked as CR'rWherein r is 1,2, …, m0;
E.g. when descending order according to signal-to-noise ratioThe rank-ordered sequence group of the sub-users obtained after the ranking is { CR1,CR2、…、CRm0、CRm0+1,…,CRmAt this time, the front m is selected0Individual secondary users, { CR1,CR2、…、CRm0The symbols are used as optimal cooperative sub-users participating in cooperative sensing and respectively correspond to marks CR1To CRm0Is best cooperative secondary user CR'1To CR'm0;
(g) Cluster first user CR1According to m in step (f)0Performing cooperative sensing based on an OR criterion on the frequency spectrum sensing result of the optimal cooperative secondary user, and taking the detection result of the cooperative sensing as the final detection result of the K secondary users in the cluster; wherein the OR criterion is as follows:
wherein, Pd,rFor optimal cooperative sub-users CR within this cluster "rProbability of detection of, Pfa,jFor optimal cooperative sub-users CR within this cluster "rFalse alarm probability of (d); qd,1Is the global detection probability, Q, after the cooperative detection of the clusterfa,1The global false alarm probability after the cluster cooperative detection is obtained; omegarRepresenting the signal-to-noise ratio SNR "rWeight coefficient of (a), ωrThe larger the weight coefficient is, the stronger the detection performance of the optimal cooperative secondary user corresponding to the weight coefficient is; SNR "maxRepresents m in the present cluster0Maximum signal-to-noise ratio, SNR, of the best cooperative sub-user "minRepresents m in the present cluster0The minimum signal-to-noise ratio of the best cooperative secondary user;
(6) respectively acquiring a third cluster to an Mth cluster according to the process of the step (5)1Intra-cluster global detection probability Q within a clusterd,3ToAnd global false alarm probabilityToWherein, the step (6) completes cooperative sensing in the rest clusters according to the cooperative process in the second cluster;
(7) spectrum sensing fusion center FC according to M1Global detection probability Q in corresponding cluster sent by first user of each clusterd,sAnd global false alarm probability Qfa,sPerforming fusion detection based on an AND criterion, AND taking a fusion detection result as a final multi-band cooperative spectrum detection result; the AND criterion is as follows:
wherein Q isdFor global detection probability after cooperation, QfaIs the global false alarm probability after cooperation. In step (7), the spectrum sensing fusion center FC only needs to be applied to M1(1<M1≤M<N) Global detection probability Q sent by first-time users of clustersd,sAnd global false alarm probability Qfa,sFusion calculation is carried out without fusing the detection results of the N secondary users, so that the fusion calculation amount is reduced to a great extent, and the fusion efficiency is improved.