CN104363064A - Cooperative spectrum sensing method based on preference users - Google Patents

Abstract

The invention discloses a cooperative spectrum sensing method based on preference users. The cooperative spectrum sensing method includes firstly adopting the preference user computing method to obtain the optimal cooperative sensing node number Kopt, computing detection probabilities of all sensing users within the sensor network and sequencing the detection probabilities from large to small; selecting the Kopt sensing users with large detection probabilities as cooperative sensing nodes; designing the data fusion schemes of the preferred cooperative sensing nodes on the basis of the data fusion decision criterion; finally selecting the optimal cooperative sensing nodes and the optimal data fusion scheme based on the preferred user computing method, and designing the cooperative spectrum sensing processing procedure on the basis of the preferred users. Compared with the prior art, the cooperative spectrum sensing method has the advantages that the preferred user computing method and the data fusion decision criterion are introduced to design the cooperative spectrum sensing processing process, and the cooperative spectrum sensing efficiency is improved. The cooperative spectrum sensing method can be applied to the fields of cognitive radio, the sensor network and the radio-frequency spectrum monitoring but is not limited within the above scope.

Description

A kind of cooperative frequency spectrum sensing method based on preferred, users

Technical field

The present invention relates to a kind of radio spectrum sensing method, particularly relate to a kind of cooperative frequency spectrum sensing method based on preferred, users.

Background technology

Frequency spectrum resource anxiety is on the rise, cognitive radio technology is by real-time perception information frequently, thus distribute flexibly and frequency of utilization resource, unauthorized user is enable to wait for an opportunity to share the frequency spectrum resource of authorized user, thus substantially increase frequency spectrum service efficiency, cognitive radio technology has been acknowledged as one of best-of-breed technology solving frequency spectrum resource anxiety, is the technical foundation of next generation wireless communication network.Frequency spectrum perception is then the top priority of cognitive radio, and single user perception verification and measurement ratio affects with shade by environment is weak, and collaborative spectrum sensing is the effective means solving this impact.When carrying out collaborative spectrum sensing, should consider to improve nodal test performance, avoiding collaborative sensing network overhead too large again and causing collaborative sensing efficiency to reduce.

Publication number CN 103795481A discloses a kind of cooperative frequency spectrum sensing method based on free probability theory, and the method is applicable to MIMO communication environment, and first sample to the Received signal strength of multiple antennas of each base station, sampled signal will focus on; Then according to the noise variance of all reception sampled signals and channel, by means of asymptotic freedom characteristic and the Wishart distribution character of random matrix, the average received signal power of all reception antennas of Algorithm for Solving based on free deconvolution is adopted; Then according to target false alarm probability, Monte Carlo emulation is used to calculate detection threshold in case only there being noise to deposit; Finally target false alarm probability and detection threshold are compared, judge that whether dominant base is at transmission signal.This invention can obtain received power accurately from the Received signal strength of secondary base station, and can effectively improve frequency spectrum perception performance, especially under low signal-to-noise ratio and Small Sample Size.

The present invention relates to a kind of cooperative frequency spectrum sensing method based on preferred, users, the method is transparent to collaborative sensing equipment and technology mode, implementation method does not have the requirement of particular job standard to wireless device, highly versatile, collaborative spectrum sensing verification and measurement ratio is high, and collaborative spectrum sensing network overhead is low, can be widely used in the radio art needing collaborative spectrum sensing, as cognitive radio, sensor network and monitoring radio-frequency spectrum, but do not limit to the above scope enumerated.

Summary of the invention

The object of the present invention is to provide that a kind of collaborative spectrum sensing verification and measurement ratio is high, collaborative spectrum sensing network overhead is low, the cooperative frequency spectrum sensing method based on preferred, users of highly versatile.

Based on the cooperative frequency spectrum sensing method of preferred, users, it is characterized in that, draw optimum collaborative sensing nodes K by preferred, users computational methods _opt, calculate the detection probability of all perception users in perception net and sort from big to small, choosing the K that detection probability is large _optindividual perception user is as collaborative sensing node; Then, the data fusion scheme of preferred node is designed based on data fusion decision rule; Finally, the optimum collaborative sensing node chosen based on preferred, users computational methods and data fusion scheme, design the collaborative spectrum sensing handling process based on preferred, users.

Compared with prior art, its remarkable advantage is in the present invention: 1, introduce preferred, users computational methods and carry out preferably, improve the validity of collaborative sensing data to collaborative spectrum sensing user; 2, based on data fusion decision rule, data integration program is designed, improve fused data efficiency; 3, based on preferred, users computational methods and data fusion conceptual design collaborative spectrum sensing handling process, improve collaborative spectrum sensing efficiency.The method can be applicable to but is not limited to cognitive radio, sensor network and monitoring radio-frequency spectrum field, but do not limit to the above scope enumerated.

Accompanying drawing explanation

Fig. 1 preferred, users choose schematic diagram;

Fig. 2 is based on the data fusion criterion structure chart of preferred, users number;

Fig. 3 is based on the collaborative spectrum sensing process chart of preferred, users.

Embodiment

For further illustrating the present invention, provide an embodiment below in conjunction with accompanying drawing 1, accompanying drawing 2 and accompanying drawing 3, the present embodiment is only limitted to a kind of implementation method of the present invention is described, does not represent the restriction to coverage of the present invention.

It is as follows that preferred, users chooses process: as shown in Figure 1, first solves optimum collaborative sensing nodes K _opt, suppose that collaborative sensing network is made up of N number of sensing node altogether, wherein have K the selected participation collaborative sensing of collaborative sensing node.P _f, P _mrepresent false alarm probability and the false dismissal probability of sensing node respectively, now the cooperation false alarm probability Q of K collaborative sensing node _fand cooperation false dismissal probability Q (K) _m(K) can be expressed as:

Q _f(K)＝1-(1-P _f) ^K(1)

Q _m(K)＝(P _m) ^K(2)

The now probability of false detection Q of system _ecan be expressed as

Q _e＝p ₀·(1-(1-P _f) ^K)+p ₁·(P _m) ^K(3)

Wherein p ₀and p ₁represent that primary user exists and non-existent prior probability respectively.

In order to make the probability of false detection in formula (3) reach minimum, order then

$\frac{{\∂Q}_e\left(K\right)}{\∂K}=-K\·p_0\·{(1-P_f)}^{K-1}+K\·p_1\·P_m^{K-1}=0---\left(4\right)$

Then optimal solution K _optfor

Wherein, represent and round downwards.

Then, the detection probability P of N number of sensing node in collaborative sensing network is calculated _di, i ∈ (1,2 ..., N}, and sort by size, select P _dilarge K _optindividual node carries out collaborative spectrum sensing as preferred node.

Data Fusion process is as follows: as shown in Figure 2, with the K selected _optbased on individual collaborative sensing node, first define u _i(i=1,2 ..., K _opt) be the local court verdict of i-th collaborative sensing node, u ₀for the result of fusion center exports, then

Owing to being separate between each collaborative sensing node, then have

$P(u_1,...,u_{K_{\mathrm{opt}}}|H_1)={\mathrm{\Π}}_S+P_{\mathrm{di}}\·{\mathrm{\Π}}_S-(1-P_{\mathrm{di}})$

$P(u_1,...,u_{K_{\mathrm{opt}}}|H_0)={\mathrm{\Π}}_S+P_{\mathrm{fi}}\·{\mathrm{\Π}}_S-(1-P_{\mathrm{fi}})---\left(7\right)$

Wherein S+ represents all and meets u _ithe set of the i of=1, S-represents all and meets u _ithe set of the i of=0, i ∈ (1,2 ..., K _opt).

In conjunction with log-likelihood ratio criterion:

Obtain:

Order ${\mathrm{\δ}}_0=\log \frac{P_1}{P_0},$ ${\mathrm{\δ}}_i=\left\{\begin{array}{cc}\log \frac{P_{\mathrm{di}}}{P_{\mathrm{fi}}},& u_i=+1\\ \log \frac{1-P_{\mathrm{fi}}}{1-P_{\mathrm{di}}},& u_i=-1\end{array}\right.,$ Then have

Based on preferred, users collaborative spectrum sensing handling process as shown in Figure 3, first defined variable Q _efor system mistake detection probability, Q _ffor system false alarm probability, Q _dfor systems axiol-ogy probability, for normalization throughput of system, p ₀and p ₁represent that primary user occurs and absent variable probability, specifically describes as follows respectively:

Step one: initialization, if iterations is M ₀, sampling number is N ₁, the target detection probability preset value of system is iterations initial value M=1;

Step 2: calculate best collaboration user number K _opt, each sensing node carries out input, calculates the detection probability P of N number of node in perception net _di(i=1,2 ..., N), and sort from big to small;

Step 3: select P _dimaximum K _optindividual collaborative sensing node carries out cooperation and merges; Weights δ is established by optimum fusion scheme _i(i=1,2 ..., K _opt), and carry out data fusion;

Step 4: renewal function value Q _e, Q _dwith if go to step six;

Step 5: m=m+1, if m≤M ₀, then two are gone to step;

Step 6: export Q _e, Q _dwith

Claims (3)

1. based on the cooperative frequency spectrum sensing method of preferred, users, it is characterized in that, the method handling process comprises following steps:

Step one: initialization, if iterations is M ₀, sampling number is N ₁, the target detection probability preset value of system is iterations initial value m=1;

Step 2: calculate best collaboration user number K _opt, each sensing node carries out input, calculates the detection probability P of N number of node in perception net _di(i=1,2 ..., N), and sort from big to small;

Step 3: select P _dimaximum K _optindividual collaborative sensing node carries out cooperation and merges; Weights δ is established by optimum fusion scheme _i(i=1,2 ..., K _opt), and carry out data fusion;

Step 4: renewal function value Q _e, Q _dwith if go to step six;

Step 5: m=m+1, if m≤M ₀, then two are gone to step;

Step 6: export Q _e, Q _dwith

2. best collaborative sensing user according to claim 1 is the detection probability P according to each collaborative sensing node _disort from big to small, choose K _optindividual P _dibe worth large node as collaborative sensing user, wherein K _optby formula enter to obtain, representing and round downwards, is all sensing nodes in collaborative sensing network by N, P _f, P _mrepresent false alarm probability and the false dismissal probability of sensing node respectively.

3. optimum fusion scheme according to claim 1 is by calculating the K participating in merging _optthe local testing result of individual cooperation aggregators is the probability statistics of 1 and 0 number, and computing formula is with wherein S+ represents all and meets u _ithe set of the i of=1, S-represents all and meets u _ithe set of the i of=0, and i ∈ (1,2 ..., K _opt).In conjunction with log-likelihood ratio criterion obtain K _optindividual cooperation aggregators carries out the decision rule of merging: wherein,