KR20170090805A - A Soft-Hard Combination-Based Cooperative Spectrum Sensing Scheme for Cognitive Radio Networks - Google Patents

Abstract

The present invention relates to a cooperative spectrum sensing method based on soft-hard convergence for a cognitive radio network.
A cooperative spectrum sensing method according to the present invention is a spectrum sensing method in a cognitive radio network, wherein in a cognitive radio network composed of a plurality of clusters, a Likelihood Ratio Test (LRT) Performing a cluster decision based on the appearance of a primary user (PU) using a soft combination based on a cluster of the cluster headers (CHs) And performing a full decision using a hard combination based on a weighted decision fusion rule, the decision being made by using a soft-hard fusion method in a cognitive radio network And is provided to improve cooperative spectrum sensing performance.

Description

[0001] A Soft-Hard Combination-Based Cooperative Spectrum Sensing Scheme for Cognitive Radio Networks [

The present invention relates to a spectral sensing method in a cognitive radio network, and more particularly to a soft-hard convergence based cooperative spectrum sensing method for cognitive radio networks.

Cognitive Radio means an intelligent wireless communication technology that intelligently detects and communicates with available frequencies by recognizing a spectrum environment and uses it to perform communication without interfering with existing services. This cognitive radio is a technology that can support spectrum sensing and is emerging as a way to solve the spectrum problem that is lacking in many countries. Therefore, cognitive radio is one of the most promising technologies for future wireless communication.

Spectrum sensing is a key function of cognitive radio to prevent harmful interference with authorized users and to improve spectrum utilization. In general, cooperative spectrum sensing using spatial diversity of secondary users (SUs) is used to further improve sensing performance. By using cooperation, SUs can share their local sensing information to make more accurate decisions than individual decisions can do.

Cooperative Spectrum Sensing is divided into three types as centralized, distributed, and relay-assisted depending on how the SUs share the sensing data in the network. First, in the first centralized cooperative sensing, FC (fusion center) controls cooperative sensing. On the other hand, the second distributed cooperative sensing does not depend on FC. SUs, on the other hand, exchange local spectral decisions between them, and report decisions about the emergence and disappearance of PUs (Primary users). A third cooperative spectral sensing is relay-assisted cooperative sensing. According to these methods, some SU having a weak sensing channel and a strong reporting channel, and some SU having a strong sensing channel and a weak reporting channel can cooperate with each other to improve the performance of the cooperative spectrum.

In this cognitive radio network, a soft combination method based on LRT (Likelihood Ratio Test) is generally applied for cooperative spectrum sensing. This LRT-based soft fusion method has a problem that the cooperative detection process is complicated and the detection probability is low There is a problem.

Chaudhari, S .; Lunden, J .; Koivunen, V .; Poor, H.V. Cooperative Sensing With Imperfect Reporting Channels: Hard Decisions or Soft Decisions? IEEE Trans. Signal Process 2012, 1, 1828. Zarrin, S .; Lim, T.J. Cooperative Spectrum Sensing in Cognitive Radios with Incomplete Likelihood Functions. IEEE Trans.Signal Process. 2010, 6, 32723281. Ujjinimatad, R .; Patil, S.R. Spectrum Sensing in Cognitive Radio Networks with Known and Unknown Noise Levels. IET Commun.2013, 15, 17081714. Reisi, N .; Gazor, S .; Ahmadian, M. Distributed Cooperative Spectrum Sensing in Mixture of Large and Small Scale Fading Channels. IEEE Trans.Wirel.Commun.2013, 11, 54035412. Reisi, N .; Gazor, S .; Ahmadian, M. A Distributed Average Likelihood Ratio Detector for Detection of Signals in Frequency-Selective Nakagami Channels. IEEE Wirel. Commun. Lett. 2014, 3, 245248. Zheng, S .; Kam, P.-Y .; Liang, Y.-C .; Zeng, Y. Spectrum Sensing for Digital Primary Signals in Cognitive Radio: A Bayesian Approach for Maximizing Spectrum Utilization. IEEE Trans. Wirel.Commun.2013, 4, 17741782. Zhou, J .; Shen, Y .; Shao, S .; Tang, Y. Cooperative Spectrum Sensing Scheme with Hard Decision Based on Location Information in Cognitive Radio Networks. Wirel. Pers. Commun.2012, 4, 26372656. Do, T.-N .; An, B. Cooperative Spectrum Sensing Schemes with the Interference Constraint in Cognitive Radio Networks. Sensors 2014, 5, 80378056. Akyildiz, I. F .; Lo, B. F .; Balakrishnan, R. Cooperative spectrum sensing in cognitive radio networks: A survey.Phys. Commun 2011,1, 4062. Nguyen-Thanh, N .; Koo, I. A cluster-based selective cooperative spectrum sensing scheme in cognitive radio. EURASIP J.Wirel.Commun.Netw.2013, 1, 19. Yarkan, S .; Arslan, H. Spectrum Exploiting location awareness toward improved wireless system design in cognitive radio. IEEE Commun. Mag.2008, 1, 128136. Sanchez, S. M .; Souza, R. D .; Fernandez, E.M.G .; Reguera, V.A. Rate and Energy Efficient Power Control in a Cognitive Radio Ad Hoc Network. IEEE Signal Process. Lett.2013, 5, 451454. Letaief, K .; Zhang, W. Cooperative Communications for Cognitive Radio Networks. Proc.IEEE 2009, 5, 878893. Peng, S .; Shu, S .; Yang, X .; Cao, X. Optimization of Log-Likelihood Ratio Test Based Cooperative Spectrum Sensing in Cognitive Radio Networks. In Proceedings of the 7th International Conference on Wireless Communications, Networking and Mobile Computing (WiCOM), Wuhan, China, 2325 September 2011. Peh, E. C. Y .; Liang Y.-C .; Guan Y. L .; Zeng Y. Cooperative Spectrum Sensing in Cognitive Radio Networks with Weighted Decision Fusion Schemes. IEEE Trans.Wirel. Commun.2010, 12, 38383847.

The present invention has been made in order to solve the problems occurring in the cooperative spectrum sensing method in the conventional cognitive radio network, and it is an object of the present invention to provide cooperative spectrum sensing in cognitive radio networks, And to provide a method for sensing a cooperative spectrum based on a soft-hard fusion, which has a high sensing performance.

According to another aspect of the present invention, there is provided a method for sensing spectrum in a cognitive radio network, the method comprising the steps of: Performing a cluster determination based on the appearance of a primary user using an LRT-based soft fusion; and receiving cluster decision from each cluster header in an FU constituting the cluster, and using a hard fusion based on a weight determination fusion rule, And performing a determination.

The step of performing the cluster determination includes receiving a signal transmitted by a primary user during a sample sensing time in SUs existing in each cluster, generating an experimental statistical value indicating energy of the received signal, and transmitting the generated experimental statistic value to a cluster header Wow; The LRT is performed by adding the experimental statistic values of all the SUs existing in the cluster header in the cluster header, and the cluster decision is made one-bit hard decision (Dc = 1 or Dc = 0) based on the appearance of the primary user .

Here, the cluster header calculates a cluster experimental statistic value by adding experimental statistic values of all SUs, calculates an average SNR of a primary user signal of the cluster head, calculates a distribution of cluster experimental statistical values through the calculation, LRT for the cluster experimental statistical value is performed to perform cluster determination for appearance or non-appearance of PU.

The FC selects one of a plurality of SUs as a cluster header in each cluster, periodically performs a spectrum sensing process, and broadcasts the entire decision to all SUs present in the network.

Also, FC collects all cluster crystals and performs a full determination according to a weight determination fusion rule, and the weight determination fusion ratio is expressed by the following equation.

(Here, c is the cluster sequence (c-th cluster) by 1, 2, ..., has a K value, D _c is the clustered crystals (Dc = 1 or Dc = 0) transmitted by each cluster header, P ₀ = P _{r (H 0)} and _{P 1 = P r (H 1} ) is a priori probability for each of the primary user (PU) appearance (H1) and the non-appearance (H0) of the signal is, ω _c is the weight element (weighted factors.

Here, the weight factor? _C is selected through the following equation.

(Where P _{f, c} is the false alarm probability of each cluster header, and P _{d and c} are the detection probabilities)

In order to obtain an optimal cluster boundary value for cluster headers, the FC develops a probability density function (pdf) of an LRT value for the cluster experimental statistical value, The notification probability and the detection probability are determined, and the optimal cluster boundary value is determined.

The false notification probability and detection probability of each cluster header are determined by the following equation.

(Here, Λ _c shows the LRT value, λ _c is a cluster boundary value performed by the cluster header)

Also, the boundary value of the optimal cluster is determined by the following equation according to the minimum error probability criterion.

According to the present invention, cooperative spectrum sensing performance can be improved by using a soft-hard convergence method in a cognitive radio network. That is, the soft-hard convergence method according to the present invention constitutes a cluster-based network, in which soft convergence based on LRT is applied in a cluster, and weight convergence based fusion rules are used in a fusion center, It is possible to reduce the complexity of the cooperative detection.

In addition, LRT can be used to detect primary signals at low SNR (about 15 dB). The use of closed-form for the probability density function of the LRT value reduces the complexity of the LRT computation. As described above, the use of the LRT in comparison with the existing soft fusion methods can reduce the sensing overhead in terms of the reporting time.

1 is a soft-hard fusion concept diagram for cooperative spectrum sensing in a cognitive radio network according to the present invention;
2 is a conceptual diagram showing a reporting mechanism of the SHC method according to the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings

1 illustrates a soft-hard convergence concept for cooperative spectrum sensing in a cognitive radio network according to an embodiment of the present invention.

In FIG. 1, the Cognitive Radio Network is composed of K clusters, and each cluster has the same number of SUs (secondary users). In a cognitive radio network, a Fusion Center (FC) configures clusters, selects a cluster header (CHc), and cooperates with all SUs in the network. The secondary system operates within the radio range of the primary user P (PU). Here, the primary user P may appear or disappear, but the state does not change during one sensing interval time. All SUs in each cluster have the same average signal-to-noise ratio (SNR) as the received primary signal. This assumption is reasonable because the clusters are grouped into neighboring SUs that are equally small. However, each cluster has different link channel conditions with the primary user P. Therefore, each cluster is independent of the primary signal and has a different average SNR.

As shown in FIG. 1, the cooperative spectrum sensing process of the soft-hard combination (SHC) method proposed in the present invention is composed of two steps as follows.

[ Stage 1  ]

Cluster headers perform some cluster decision based on the first activity using a soft combination as follows. First, at the beginning of the sensing process, SU _ci (i-th SU in the c-th cluster) hears the primary signal and produces a regional test statistic ρ _ci that represents the energy of the received signal. Each SU uses M primary signal samples to generate local experimental statistics. And the regional experimental statistical data ρ _ci is transmitted to the cluster header.

Each cluster has one cluster header, and the cluster header can cooperate with all the SUs in the cluster. The cluster header of the c- th cluster is shown as follows. CH _c ( c = 1, 2, ..., K ,) represents the cluster header of the c- th cluster. Cluster header selection is done as follows.

First, a cognitive radio (CR) system performs a spectrum sensing process periodically to detect the appearance of a primary user (PU) P. In general, the structure of CR system is composed of one sensing slot and one data transmission slot. The cooperative spectrum sensing process is performed periodically by FC (Fusion Center) in the sensing slot. The cooperative spectrum sensing period is achieved by the system designer by adjusting the spectrum sensing and spectral division according to application requirements. The FC selects an SU (secondary user) as one cluster header in each cluster. The cluster headers perform Likelihood Ratio Test (LRT) based on experimental statistical data of all SUs existing in the cluster, and determine one-bit hard decision ). D _c ( c = 1, 2, ..., K) represents the cluster decision of the c- th cluster (ie, D _c = 1 or D _c = 0, respectively.

[ Step 2  ]

All cluster headers send their cluster determination information D _c to FC via error-free reporting channels. FC aggregates all cluster decisions and makes the entire decision using a weighted decision fusion rule. As mentioned earlier, clusters use a different average SNR than the received primary signal, so the contribution to the overall decision is different. However, existing fusion rules ( k-out-of-N , eg, OR rule, AND rule or MAJORITY rule) do not consider this aspect. Therefore, the existing k-out-of-N rule can not be applied to the SHC method of the present invention. On the other hand, the weighted decision fusion rule according to the present invention assigns other weighted factors to cluster determination according to sensing reliability.

2 is a conceptual diagram showing a reporting mechanism of the SHC method according to an embodiment of the present invention.

As shown in FIG. 2, in a conventional soft combination scheme, SUs sequentially transmit sensing data to the FC. Here, t _s represents the transmission time required for one SU to transmit the sensing data to the FC.

On the other hand, in the SHC method of the present invention, SUs existing in the same cluster transmit sensing data to the cluster header. For a reasonable comparison with the existing methods, let us denote the time required for an SU to transmit the sensing data to the cluster header as t _s . The cluster header makes a cluster decision with one bit, and sequentially transmits it to the FC. The time required for the cluster header to send the decision to the FC is denoted by t _h . The more data the SU has to report in the cluster header at a given bandwidth and transmission rate of the control channel, the more time it takes. Let ε (ε <0) be the correlation coefficient between the transmission time of the soft sensing data collected by the SU and the transmission time of the one bit decision made by the cluster header. That is, t _s = 竜 t _h .

Finally, the entire decision is made by FC. The total crystal made during each sensing period is denoted by D _g . That is, D _g = 1 or D _g = 0 indicates the occurrence and disappearance of the primary user, respectively. At the end of the spectrum sensing process, the FC broadcasts the entire decision to all SUs present in the network. In the present invention, P _r (A) represents the probability of an arbitrary event A. Likelihood Ratio Test is used for LRT and Log-Likelihood Ratio Test is used for L-LRT.

Hereinafter, in the SHC method of the present invention, the process of soft fusion in the cluster header of each cluster will be described.

First, the cluster SU _ci (i-th secondary user of c-th cluster) measures the received signal r _ci during the M sample sensing period. The signal transmitted by the primary user is S _ci . This signal propagates to SU _ci through a flat fading channel. The m-th sample r _ci (m) of the discrete signal received in the secondary user SU _ci is expressed as Equation (1) below.

(additive white and Gaussian (AWGN) with zero-mean)를 가정한다. h_ci는 채널 이득, s_ci는 전송된 1차 신호를 각각 나타내고, 서로 독립이다. 1차 사용자의 상태는 하나의 센싱 기간 동안 일정하다. Where H ₀ and H ₁ represent the absent and present of PU near SUs, respectively. Also, r _ci represents the primary received signal in the c-th cluster. The noise

(additive white and Gaussian (AWGN) with zero-mean). h _ci denotes a channel gain, and s _ci denotes a transmitted primary signal, and are independent of each other. The state of the primary user is constant during one sensing period.

The local test statistics ρ _ci (estimation of received primary signal power of the SU _ci ) is expressed by the following equation (2).

Where M = 2 TW, which represents the number of samples collected in each SU during a sensing period. T and W represent the detection time and the bandwidth of the signal, respectively. In the SHC method proposed in the present invention, one channel is sensed at a time.

The test statistics (TS) of SUs are combined in the cluster headers using Equal Gain Combining (EGC). The cluster test statistic ρ _c (the estimation of received primary signal power at the cluster head of the c-th cluster) is expressed by the following equation (3).

The experimental statistical value ρ _c is an independent variable under the assumption of H ₀ , and the probability density function has a Chi-square distribution with degrees (L = L) degrees of freedom. Under H ₁ assumption, ρ _c is an independent non-central chi-square random variable with L degrees of freedom and a non-central parameter γ _cL .

The average SNR of the primary user signal measured in the cluster header CH _c is expressed by Equation (4) below. Here, it is assumed that all SUs in the same cluster have the same SNR.

As mentioned earlier, the SNR is obtained using the SU location information. Distribution for ease of analysis noise test statistics by assumed that the unit variance, and using the Central Limit Theorem (CLT) ρ _c is a Gaussian distribution under H ₀ or H ₁ two conditions (H ₀ or H ₁₎ As shown in FIG. Therefore, the distribution of ρ _c can be expressed by the following equation (5).

Where L = NM is the number of samples of the primary signal received, and these samples are collected using soft fusion in the cluster header. The cluster headers of each cluster use ρ _c as a cluster observing means for making a cluster decision. The cluster header performs the LRT to determine the clusters for the appearance or non-occurrence of the primary user. The log-likelihood ratio test (L-LRT) for the binary hypothesis experiment given in Equation (1) can be expressed as Equation (6).

Here, f _p (ρ _c │H ₁ ) and f _p (ρ _c │H ₀ ) are the probability density functions of the cluster experimental statistical data ρ _c under the assumption of H ₁ and H ₀ , respectively, and log means natural logarithm. Because equal a 1 SNRs of the difference signal received within a cluster, A _c (conducted at a certain secondary user i-th in c-th cluster SUci and also at a cluster head) distribution as is f _Λ (Λ _c) . &Lt; / RTI &gt; Therefore, the cluster decision D _c (0, 1) can be expressed by Equation (7) based on L-LRT.

Hereinafter, the process of hard fusion at the FU (Fusion Center) will be described.

As shown in FIG. 1, the network is composed of K clusters, and N SUs exists in each cluster.

The fusion center (FC) receives and aggregates the cluster decisions to determine the state of the primary user (PU). Here, a weighted decision fusion rule (WDFR) is used as a fusion rule. D = [D ₁ , D ₂ . , D _K ] denote the set of cluster decisions received at the FC. FC uses the LRT to make the overall decision as shown in equation (8).

Where P ₀ = P _{r (} H ₀ ) and P ₁ = P _{r (} H ₁ ) are prior probabilities for the appearance and non-occurrence of PU signals, respectively. The determination of the cluster headers is performed independently, and the L-LRT is developed as shown in the following Equation (9).

Therefore, the weighted decision fusion rule can be expressed by Equation (10). &Quot; (10) &quot;

Here, the weighted factors are selected using the following equation (11).

Here, P _{f, c} denotes a false alarm probability of each cluster header, and P _{d, c} denotes a detection probability. (10) is an optimal decision fusion rule, and finally FC broadcasts the entire decision to all SUs in the network.

Hereinafter, a method for calculating an optimal cluster boundary value for cluster headers in the FC will be described. First, an energy detector (ED), which is the most common conventional sensing method, is introduced, and a method for obtaining an optimal cluster boundary value for cluster headers in the SHC method according to the present invention will be described.

[ED (Energy Detector)]

In order to explain the operation of ED, which is an existing sensing method, it is assumed that ED is used in some cluster headers. In this case, CH _c performs the determination according to the following equation (12) based on the energy boundary value? _{ED, c} .

와 local detection probability(LDP)

는 다음의 수학식 13 및 14와 같이 결정된다.Here,? _C is an experimental statistic value described in Equation (3). D _{ED, c} = 1 or D _{ED, c} = 0 means that H ₁ or H ₀ estimation is done in CH _c using ED. Local false alarm probability (LFP)

And local detection probability (LDP)

Is determined according to the following equations (13) and (14).

로 Q-function을 나타낸다..Where γ _c is the average SNR in CH _c . L is the number of primary samples collected and received at each CH _c . Also,

Represents a Q-function.

[Optimal cluster Boundary value  Calculation ]

In the SHC method of the present invention, the pdf (Probability Density Function) of the LRT value Λ _c is developed to calculate an optimal cluster boundary value for the cluster headers.

ρ = [ρ ₁ , ρ ₂ , ... , ρ _c , ... , ρ _K ], μ _{c, j} (j = 0 or j = 1) means the variance of Equation (5). ρ _c is a random variable representing the test statistic for the LRT in the cluster header CH _c . From Equations (5) and (6), the LRT value is developed as shown in the following Equation (15).

The mean value and the variance value in Equation (5) are substituted into Equation (15), and the pdf of the LRT is developed as Equation (16) by applying the opportunistic theory.

here,

And,

here,

Therefore, the false alarm probability P _{f, c} and the detection probability P _{d, c} of each cluster header are expressed by the following equations (20) and (21).

The values of the above equations 20 and 21 can be conveniently obtained using MATLAB. It can be seen from the foregoing description that the false alarm probability (P _{f, c} ) and the detection probability (P _{d, c} ) of each cluster are determined by the channel conditions, that is, the average SNR and the cluster boundary value Able to know. It is very meaningful to find an optimal local sensing boundary value that minimizes the overall sensing error under fixed channel conditions.

In the present invention, a minimum error probability criterion is used to determine the cluster boundary value λ _{opt, c} of the c-th cluster, which can be expressed by the following equation (22).

As can be seen from the above equations (20) to (22), it can be seen that the optimal cluster boundary value? _{Opt, c} can be obtained based on the pdf of the LRT value. Therefore, by using pdf in Equation 16, the cluster header CHc can obtain the optimal cluster boundary value [lambda] _{opt, c} , which is used for comparison in Equation (7).

As described above, the soft-hard fused cooperative spectrum sensing method according to the present invention can enhance the sensing performance by sensing the spectrum using the soft fusion and the hard fusion together.

That is, in the SHC method of the present invention, cluster headers CHc in each cluster fuse experimental statistical values of other SUs and perform LRT with optimal cluster boundary values determined by a minimum error probability criterion. The optimal cluster boundary value is developed using a closed-form representation of the pdf of the LRT value. This simulation result shows that the LRT performs better in comparison with the existing ED, especially in the low SNR system. In addition, the Weighted Decision Fusion Rule (WDFR) can be used by the FC to distinguish the contribution of the other cluster headers. The SHC method can be applied to existing hard fusion methods such as "AND Rule, OR Rule, MAJORITY Rule, LRT" It is possible to obtain a better sensing performance for each case as compared with the case of using each of them. On the other hand, the reporting mechanism of the SHC method can reduce the reporting time compared to the existing soft fusion method.

It is to be understood that the present invention is not limited to the above-described embodiment, and that various modifications and changes may be made by those skilled in the art without departing from the spirit and scope of the appended claims. Of course, can be achieved.

CH: Cluster header
PU: Primary User
SU: Secondary user
FC: Fusion Center

Claims (9)

A method for spectral sensing in Cognitive Radio Networks,
(a) In a cognitive radio network composed of a plurality of clusters, a soft combination based on a Likelihood Ratio Test (LRT) based on each cluster header (CHs) is used for the appearance of a primary user (PU) Based cluster determination;
(b) a cluster decision is received from each cluster header (CHs) in a fusion center (FU) constituting the cluster, and a cluster determination is performed using a hard combination based on a weighted decision fusion rule And performing a soft-hard fusion-based cooperative spectrum sensing method for a cognitive radio network.
The method according to claim 1,
The step (a) of performing the cluster determination
(a-1) to receive signals (S _ci) by the sample sensing time the primary user (PU) for a transmission from the SUs (Secondary Users) present in each cluster, the experiment representing the energy of the received signal (r _ci) Generating test statistics (? _Ci ) and transmitting them to the cluster header;
(LRT) is performed by adding the experimental statistical values (ρ _ci ) of all the SUs present in the cluster in the cluster header (a-2), and one-bit hard decision hard decision (Dc = 1 or Dc = 0) for the soft-hard convergence-based cooperative spectrum sensing method.
3. The method of claim 2,
The cluster header
Calculating a cluster test statistic value (ρ _c) the sum of the test statistic value (ρ _ci) of the all SUs and
Calculating an average signal-to-noise ratio (SNR) (γ _c ) of the primary user signal (S _ci ) of the cluster head, calculating a distribution of the cluster experimental statistical value (ρ _c )
Characterized in that the LRT for the cluster test statistic (r _c ) is performed to perform a cluster decision on the appearance of the PU (Dc = 1) or the non-appearance (Dc = 0) HARD Fusion - based Cooperative Spectrum Sensing Method.
3. The method according to claim 1 or 2,
Wherein the FC selects one of a plurality of SUs as a cluster header in each cluster, periodically performs a spectrum sensing process, and broadcasts the entire decision to all SUs present in the network. Soft - Hard Fusion - based Cooperative Spectrum Sensing Method for Networks.
The method according to claim 1,
Wherein the FC performs a full determination according to a weight determination fusion rule by collecting all cluster decisions, and the weight determination fusion ratio is expressed by the following equation: &lt; EMI ID = Way.

(Here, c is the cluster sequence (c-th cluster) by 1, 2, ..., has a K value, D _c is the clustered crystals (Dc = 1 or Dc = 0) transmitted by each cluster header, P ₀ = P _{r (H 0)} and _{P 1 = P r (H 1} ) is a priori probability for each of the primary user (PU) appearance (H1) and the non-appearance (H0) of the signal is, ω _c is the weight element (weighted factors)
6. The method of claim 5,
Wherein the weight factor? _C is selected through the following equation:? _C ?

(Where P _{f, c} is the false alarm probability of each cluster header, and P _{d and c} are the detection probabilities)
7. The method according to claim 1 or 6,
In order to obtain an optimal cluster boundary value for cluster headers,
The probability density function pdf of the LRT value Λ _c with respect to the cluster experimental statistical value ρ _c is developed (f _Λ (Λ _c )), and the false notification probability of each cluster header (P _{f, c} ) and a detection probability (P _{d, c} ) to determine an optimal cluster boundary value. Cooperative Spectrum Sensing Method.
8. The method of claim 7,
Wherein a false alarm probability (P _{f, c} ) and a detection probability (P _{d, c} ) of each cluster header are determined by the following equation: - HARD Fusion - based Cooperative Spectrum Sensing Method.

(Here, Λ _c shows the LRT value, λ _c is a cluster boundary value performed by the cluster header)
9. The method of claim 8,
Wherein the boundary value of the optimal cluster is determined by the following equation according to a minimum error probability criterion. &Lt; RTI ID = 0.0 &gt; 11. &lt; / RTI &gt;

Images