KR101462815B1 - Method for fusion of soft decision in cognitive radio system - Google Patents

Abstract

The present invention relates to a crystal fusing method of soft decision in a cognitive radio system and, more particularly, to a crystal fusing method of soft decision using the weight of a received bit and a sensor performance index. The cognitive radio system according to the embodiment of the present invention is based on a cooperation spectrum sensing method in which devices mounting a plurality of cognitive radio (CR) sensors share the sensing information and includes a fusing center to fuse the sensing information received from the CR sensor. According to the present invention, performance is improved by providing the crystal fusing method of the soft decision in consideration of the weight of the bit and the sensor performance index.

Description

[0001] The present invention relates to a method of fusion determination of a cognitive radio system,

The present invention relates to a decision fusion method of a soft decision method in a cognitive radio system, and more particularly to a decision fusion method of a soft decision method using a sensor performance index and a weight of a received bit.

Currently, wireless communication technology is researching and developing in the direction of ubiquitous network which can exchange information with any person, anytime, anywhere. In recent years, radio waves have been used in mobile communication and broadcasting. However, the demand for frequency resources is increasing rapidly as it extends to all areas of life from traffic, medical science, and security to public safety. Of the total. Cognitive Radio (CR) technology, which is used for finding and communicating an empty frequency that is not actually used, has been attracting attention in order to solve the frequency shortage problem and to maximize the frequency utilization efficiency.

Cognitive radio technology is a technology that automatically detects unused frequencies according to area and time and protects the authorized stations in the surrounding area and makes the desired communication possible. It is possible to find an idle spectrum which is scattered in various widths and occupied time continuously and determines the frequency bandwidth, output, modulation method and the like suitable for the environment and recycles it, thereby improving the efficiency of the frequency resource, which is a limited resource. Since 2004, the IEEE has been working to standardize the technology for using cognitive radio (CR) in the television frequency band.

Among the cognitive radio technologies, in particular, spectrum sensing technology forms the basis of CR technology as a technique of detecting an empty frequency band not used by a primary user by detecting a radio wave environment of a secondary user. The method of spectrum sensing is matched filter, signal feature detection, energy detection, etc. However, it can be implemented even if the type of the signal to be transmitted is unknown and its complexity and sensing time The energy detection method is most suitable. However, if one sub-user performs spectrum sensing independently, it may happen that hidden node problem, shadow fading problem, multi-path fading problem, etc., can not be detected accurately. To solve this problem, Cooperative Spectrum Sensing technology has been developed to determine the spectral occupancy by sharing the result of sensing by the sub users separately in the Fusion Center.

The cooperative spectrum sensing is classified into a hard decision method and a soft decision method according to a method of fusing individual sensing results.

The soft decision method is a method in which the results of several users are transmitted to the convergence center and the final decision is made based on the results from the fusion center. In the hard decision method, each user firstly judges whether or not spectrum is used, To the fusion center and make a final decision.

The soft decision method increases the complexity and the sensing time, while the frequency efficiency of the whole system is reduced as compared with the hard decision method. On the other hand, in the hard decision method, the performance of the sensing is lower than that of the soft decision method, but the system is less burdensome and is a relatively simple method.

Korean Patent Publication No. 10-2013-0008065

SUMMARY OF THE INVENTION It is an object of the present invention to provide a decision fusion method of a soft decision method considering a sensor performance index of a cognitive radio (CR) sensor and a weight of a received bit.

The objects of the present invention are not limited to the above-mentioned objects, and other objects not mentioned can be clearly understood by those skilled in the art from the following description.

)를 추정하는 단계, 추정된 신호

에 대해 1의 개수를 m으로 정의하되,

(수학식 2)로 정의하는 단계, 상기 m을 이용하여 결정 통계량 Λ을 정의하되,

(수학식 3)으로 정의하는 단계 및 상기 Λ을 미리 정해진 임계값과 비교하여, Λ이 임계값보다 크면 신호가 존재하는 것으로 판단하고, 그렇지 않으면 신호가 존재하지 않는 것으로 판단하는 단계를 포함한다.According to an aspect of the present invention, there is provided a method of sensing a sensor based on a cooperative spectrum sensing method in which a plurality of devices equipped with a cognitive radio (CR) sensor share sensing information, in the soft decision method convergence determination method in that the radio system comprising a bit signal to the received signal y _k (

), Estimating the estimated signal

The number of 1s is defined as m,

(2), defining a decision statistic Λ using the m,

(Equation 3), and comparing the Λ with a predetermined threshold to determine that a signal is present if Λ is greater than a threshold, otherwise determining that no signal is present.

는 H₁의 경우 각각의 CR 센서에서 1을 전송할 확률로 올바른 정보를 제공할 확률이고,

는 H₀의 경우 각각의 CR 센서에서 1을 전송할 확률로 잘못된 정보를 제공할 확률이라고 할 때,

이고,

일 수 있다.

Is the probability of providing correct information with probability of transmitting 1 in each CR sensor in case of H ₁ ,

Is the probability that each CR sensor will give false information as a probability to transmit 1 in case of H ₀ ,

ego,

Lt; / RTI >

(수학식 4)로 나타낼 수 있다.
When Λ = Λ ₁ + Λ ₂ , Λ ₁ ≥0, and Λ ₂ ≤0, where Λ ₁ and Λ ₂ are values of

(Equation 4).

According to the present invention, there is an effect that performance can be improved by providing a decision fusion method of a soft decision method in consideration of a sensor performance index and a weight of a bit.

In the proposed LRT-CS, three kinds of information such as channel statistics, sensor performance index, and noise variance are used. However, the method proposed in the present invention uses only two kinds of information such as sensor performance index and noise variance, CS.

In addition, the convergence decision method proposed by the present invention shows excellent performance in SNR excluding high SNR compared with CV using the same information. Therefore, the method proposed by the present invention is a soft decision method with high performance even with a small amount of information, thereby reducing the occurrence of performance degradation in the shadow area and latency terminal of spectrum detection and improving the quality of wireless service There is an effect.

1 is a flowchart showing a decision fusion method of a soft decision method in a cognitive radio system according to an embodiment of the present invention.
2 is a histogram of a decision fusion method of a soft decision method according to an embodiment of the present invention.
3 is a graph illustrating an average of H ₀ and H ₁ distributions according to an SNR according to an embodiment of the present invention.
4 is a graph illustrating distribution of H ₀ and H ₁ distributions according to SNR according to an embodiment of the present invention.
5 is a table summarizing the information required by each fusion determination method.
6 is a graph showing the change of P _D according to the SNR in the fusion determination method.
7 is a graph showing the change of P _D according to the cooperative P _F in the fusion determination method.
8 is a graph showing a change in P _D according to the number K of the local sensors in the fusion determination method.
9 is a graph showing the change of P _D according to another P _d in the fusion determination method.
10 is a graph showing a change in P _D when the channel average is different in the fusion determination method.

While the invention is susceptible to various modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawings and will herein be described in detail. It should be understood, however, that the invention is not intended to be limited to the particular embodiments, but includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention.

The terminology used in this application is used only to describe a specific embodiment and is not intended to limit the invention. The singular expressions include plural expressions unless the context clearly dictates otherwise. In the present application, the terms "comprises" or "having" and the like are used to specify that there is a feature, a number, a step, an operation, an element, a component or a combination thereof described in the specification, But do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, or combinations thereof.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Terms such as those defined in commonly used dictionaries are to be interpreted as having a meaning consistent with the contextual meaning of the related art and are to be interpreted in an ideal or overly formal sense unless expressly defined in the present application Do not.

In the following description of the present invention with reference to the accompanying drawings, the same components are denoted by the same reference numerals regardless of the reference numerals, and redundant explanations thereof will be omitted. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description, well-known functions or constructions are not described in detail since they would obscure the invention in unnecessary detail.

The present invention relates to a decision fusion method of a soft decision method in a cognitive radio system based on a cooperative spectrum sensing method in which a plurality of devices equipped with CR (Cognitive Radio) sensors share sensing information.

In the present invention, the cognitive radio system includes a fusion center for fusing the sensing information received from the CR sensor.

1 is a flowchart showing a decision fusion method of a soft decision method in a cognitive radio system according to an embodiment of the present invention.

)한다(S101). 이는 다음 수학식과 같이 표현될 수 있다.Referring to FIG. 1, first, a bit is estimated on a received signal y _k (

(S101). This can be expressed as the following equation.

에 대해 1의 개수를 m으로 정의한다(S103). 이를 수학식으로 나타내면 다음과 같다. Then, the estimated signal

The number of 1's is defined as m (S103). This can be expressed by the following equation.

Next, a decision statistic Λ is defined using m (S105). This can be expressed by the following equation.

Next, it is compared with a predetermined threshold (S107). If? Is larger than the threshold value, it is determined that a signal exists (S109). Otherwise, it is determined that no signal exists (S111).

The fusion method proposed in the present invention requires sensor performance and noise variance information.

는 H₁의 경우 각각의 CR 센서에서 1을 전송할 확률로 올바른 정보를 제공할 확률이고,

는 H₀의 경우 각각의 CR 센서에서 1을 전송할 확률로 잘못된 정보를 제공할 확률이라고 하자.

Is the probability of providing correct information with probability of transmitting 1 in each CR sensor in case of H ₁ ,

Let H ₀ is for a probability of transfer from one sensor each CR probability of providing erroneous information.

The decision fusion method of the present invention is a method of weighting received y _k by multiplying y _k > 0 or y _k < 0 by the probability of providing correct information and dividing the probability of providing erroneous information.

이므로,

이다. And,

Because of,

to be.

When we define Λ = Λ ₁ + Λ ₂ , Λ ₁ ≥0 and Λ ₂ ≤0.

At this time, Λ ₁ and Λ ₂ can be expressed by the following equations.

In the case of H ₀ , as the SNR increases, the value of m becomes smaller and the value of K m becomes larger, so that the value of Λ ₂ becomes smaller and the value of Λ becomes smaller.

Conversely, in the case of H ₁ , since the value of m is almost the same as Km, the value of Λ becomes larger than that of H ₀ .

Therefore, the fusion method proposed by the present invention is a method of determining the fusion decision of the soft decision method by multiplying different weights according to the received signal.

Likelihood Ratio Test (LRT), Likelihood Ration Test Based on Channel Statistics (LRT-CS), Chair-Varshney (CV), Maximum Ration Combiner (MRC), Equal Gain Combiner (EGC) Hereinafter, the method will be described.

LRT is an optimal crystal fusion method, and uses channel gain, sensor performance index, and noise variance information.

(k=1,...,K)는 지역 센서의 검출확률(detection probability)을 의미하고,

(k=1,...,K)는 지역센서의 오 경보확률(false alarm probability)을 의미한다.y _k is the received signal and h _k is the Rayleigh fading channel gain,

(k = 1, ..., K) denotes a detection probability of the local sensor,

(k = 1, ..., K) denotes a false alarm probability of the local sensor.

At this time, the decision statistics Λ is defined by the following equation.

이다.
At this time,

to be.

LRT-CS is a quasi-optimal crystal fusion method and exhibits almost the same performance as LRT, which is an optimal crystal fusion method. In this method, channel statistics, sensor performance index, and noise variance information are used.

In LRT-CS, the decision statistic Λ is defined as:

이고,

이다.
here,

ego,

to be.

CV shows similar performance to LRT, which is an optimal crystal fusion method at high SNR (signal to noise ratio), and the information used is sensor performance index and noise variance.

In CV, the decision statistic Λ is defined by the following equation.

MRC shows similar performance to LRT, which is the optimal crystal fusion method at low SNR, and the information used is channel gain and variance of noise.

In MRC, the decision statistic Λ is defined as:

EGC is a decision fusion method that assumes that all channels are the same in the MRC decision fusion method. The information used is a variance of noise and has the feature to use the smallest amount of information.

The decision statistic Λ in the EGC is defined as:

The cooperative spectrum is that devices equipped with multiple CR (cognitive radio) sensors share detection information to increase the reliability and accuracy of single spectrum detection.

In the cooperative spectrum detection method, the signal y _k transmitted from each local sensor is integrated according to the fusion rule in the fusion center, and the decision statistic (Λ) is compared with the threshold value Γ. If Λ> Γ, a signal exists And judges that the signal does not exist if Λ <Γ. Here, the threshold value Γ is set to a value satisfying the cooperative P _F according to the SNR in the case of H ₀ . Therefore, the decision fusion method proposed in the present invention determines a new decision statistic and is proposed through the following process.

y _k is the received signal, u _k is the local decision at each sensor, u _k ∈ {-1,1}.

은 k번째 센서에서의 검출확률이고,

은 k번째 센서에서의 오 경보확률이라 정의한다. 이때, k=1,...,K의 값을 갖는다.

Is the detection probability at the k-th sensor,

Is defined as the probability of false alarm at the kth sensor. At this time, k = 1, ..., K.

The decision statistic Λ for making the final decision at the convergence center for the signal yk received from the sensor is defined as follows.

로 y_k>0인 경우, u_k=1로 정의하여 수신된 신호 y_k로부터 추정된 신호의 1의 개수를 의미하고,

으로 정의한다. here,

Y _k > 0, u _k = 1 means the number of 1s of the signal estimated from the received signal y _k ,

.

Here, when Λ = Λ ₁ + Λ ₂ , Λ ₁ and Λ ₂ are defined as follows.

The crystal fusion method of the present invention requires a sensor performance index and noise variance information like the CV crystal fusion method.

는 H₁의 경우 각각의 CR 센서에서 1을 전송할 확률로 올바른 정보를 제공할 확률이고,

는 H₀의 경우 각각의 CR 센서에서 1을 전송할 확률로 잘못된 정보를 제공할 확률이라고 하자.

Is the probability of providing correct information with probability of transmitting 1 in each CR sensor in case of H ₁ ,

Let H ₀ is for a probability of transfer from one sensor each CR probability of providing erroneous information.

The decision fusion method of the present invention is a method of weighting received y _k by multiplying y _k > 0 or y _k < 0 by the probability of providing correct information and dividing the probability of providing erroneous information.

이므로,

이다.
And,

Because of,

to be.

2 is a histogram of a decision fusion method of a soft decision method according to an embodiment of the present invention.

=1의 개수를 H₀, H₁ 각각에서 SNR의 변화에 따른 분포를 살펴보면, H₀의 경우 SNR이 낮으면 가우시안 분포 형태이나, SNR이 높아짐에 따라 익스포넨셜(exponential) 분포의 형태에 가까워지는 것을 확인할 수 있다. Referring to Fig. 2, when K = 8,

= 1, the distribution of H ₀ and H ₁ according to the change of SNR is as follows _{. In} the case of H ₀ , the SNR is low in the Gaussian distribution form. However, as the SNR is high, the SNR is close to the exponential distribution Can be confirmed.

On the other hand, it can be seen that H ₁ follows the Gaussian distribution regardless of the SNR variation.

Figure 3 is showing the average of H _0, H ₁ distribution on SNR in accordance with one embodiment of the present invention, a graph, Figure 4 is the variance of H _0, H ₁ distribution on SNR in accordance with one embodiment of the present invention Fig.

Referring to FIG. 3 and FIG. 4, it can be seen that the mean and variance of the distribution according to the SNR are as follows _{. In} the case of H ₀ , the average and variance start from a certain number at a low SNR and close to 0 as the SNR increases .

On the other hand, in the case of H ₁ , it can be seen that the average and variance are maintained at a constant number regardless of the SNR.

According to this reason, the value of m becomes smaller as the SNR increases in H ₀ , and the value of K is larger than that of H ₀ , so that the value of Λ ₂ becomes smaller and the value of Λ becomes smaller.

On the contrary, in the case of H ₁ , the value of Λ is larger than that of H ₀ because the value of m is almost the same as that of K m.

Accordingly, the fusion determination method proposed by the present invention is a method of reducing the fusion rule of the soft decision method by multiplying different weights according to the received signal.

Hereinafter, the performance of the existing fusion determination method for the fusion determination method proposed in the present invention will be compared and described.

5 is a table summarizing the information required by each fusion determination method.

FIG. 5 shows information required in the conventional fusion determination methods and the fusion determination method proposed in the present invention.

6 is a graph showing the change of P _D according to the SNR in the fusion determination method.

FIG. 6 shows the number of local sensors, the probability of detection of all the local sensors P _d = 0.5, the probability of false alarm of the local sensor P _f = 0.05, the cooperative alarm probability false alarm probability) P _F = 0.01 to indicate the cooperative detection probability P _{D according} to the channel SNR.

Referring to FIG. 6, the convergence determination method proposed by the present invention shows approximately the same performance at a low SNR and a suboptimal fusion decision LRT-CS requiring three pieces of information. As the SNR increases, . However, the method proposed by the present invention uses less information as the amount of information required for LRT-CS is two. And it shows a remarkably higher performance than the CV convergence decision method which requires the same amount of information except for the high SNR. It shows high performance over the entire SNR rather than EGC, except for lower SNR than MRC using the other two information.

7 is a graph showing the change of P _D according to the cooperative P _F in the fusion determination method.

FIG. 7 is a graph showing the results obtained by fixing the number of area sensors to eight, the detection probability P _d of all the local sensors, the false alarm probability P _f of the local sensor P _f = 0.05, and SNR = We compare the performance of the cooperative P _D according to the cooperative P _F of the proposed convergence decision method.

Referring to FIG. 7, the fusion determination method proposed by the present invention shows similar performance to that of the LRT-CS, which is a suboptimal fusion determination method. It can be seen that the CV method using the same information and other MRC and EGC You can see that it shows good performance.

8 is a graph showing a change in P _D according to the number K of the local sensors in the fusion determination method.

FIG. 8 shows the cooperative detection probability P _D as the detection probability P _d of all the local sensors, P _d of false alarms of the local sensor P _f = 0.05, cooperative false alarm probability P _F = 0.01 and SNR = The change with the number K is shown.

Referring to FIG. 8, it can be seen that the convergence determination method proposed by the present invention has a high cooperative detection probability as the number of area sensors increases as a whole. However, when the number of local sensors is the same, the results are similar to those of LRT-CS, which is a suboptimal fusion decision. Also, it shows remarkably better performance than CV method using the same information and other MRC and EGC.

9 is a graph showing the change of P _D according to another P _d in the fusion determination method.

FIG. 9 is a graph showing the results obtained by fixing the number of area sensors to 8, the false alarm probability P _{F of} 0.05 for all the regional sensors and the cooperative P _F = 0.01, and the fusion determination method proposed by the present invention, And the performance of the cooperative P _D according to the change of the SNR by comparing the detection probabilities P _d of the sensors.

Referring to FIG. 9, the fusion determination method proposed in the present invention shows a lower P _D than LRT, which is an optimal fusion decision using a lot of information, and LRT-CS, which is a suboptimal fusion decision, It shows remarkably high performance except high SNR rather than fusion determination method. In addition, it shows high performance over the entire SNR rather than MRC and EGC.

10 is a graph showing a change in P _D when the average SNR of channels is different in the fusion determination method.

Figure 10 is eight the number K of the area sensor, a detection likelihood of the other countries sensor P _d = 0.5, probability of false alarm of the area sensor P _f = 0.05, cooperation probability of false alarm cooperation detection probability is fixed to P _F = 0.01 P _D In accordance with the channel SNR.

Referring to FIG. 10, the convergence determination method proposed by the present invention has lower performance than LRT and LRT-CS which use more information as a whole, but also has a high SNR when compared with a CV using the same information Except for the high performance. Also, when compared with MRC, it shows similar performance at low SNR and shows better performance as SNR increases. And overall performance is better than EGC.

While the present invention has been described with reference to several preferred embodiments, these embodiments are illustrative and not restrictive. It will be understood by those skilled in the art that various changes and modifications may be made therein without departing from the spirit of the invention and the scope of the appended claims.

Claims (3)

In a cognitive radio system including a fusion center for fusing sensing information received from the CR sensor based on a cooperative spectrum sensing method in which devices equipped with a plurality of cognitive radio (CR) sensors share sensing information, In the crystal fusion method of the present invention,
A bit signal (&lt; _{RTI ID = 0.0} &gt;

);
Estimated signal

The number of 1s is defined as m,

(2);
Define a decision statistic Λ using m,

(3); And
Comparing the? With a predetermined threshold to determine that a signal is present if? Is greater than a threshold, and otherwise determining that no signal is present.
In claim 1,

Is the probability of providing correct information with probability of transmitting 1 in each CR sensor in case of H ₁ ,

Is the probability that each CR sensor will give false information as a probability to transmit 1 in case of H ₀ ,

ego,

Wherein the decision fusion method comprises the steps of:
The method of claim 2,
When Λ = Λ ₁ + Λ ₂ , Λ ₁ ≥0, Λ ₂ ≤0,
At this time, Λ ₁ and Λ ₂ ,

(4). &Lt; / RTI &gt;

Images