CN101808334B - Spectrum perception method for detecting angle of arrival of authorized user in cognitive radio - Google Patents

Abstract

The invention discloses a spectrum sensing method for detecting the angle of arrival of authorized users in a cognitive radio system, relates to a spectrum sensing method in a cognitive radio system, and more specifically relates to a method capable of detecting signals and angles of arrival of authorized users. The steps of the method are as follows: the signal is received by the linear array antenna system; the spatial spectrum of the signal is obtained by a multiple signal classification algorithm; the statistical average value λ of the spatial spectrum is used as a judgment parameter _; Value γ; compare λ with the threshold value γ to determine whether the authorized user exists; if the authorized user exists, use the minimum optimization search to obtain the angle of arrival of the authorized user. This method of detecting authorized user multi-dimensional information provides conditions for further flexible use of authorized user frequency bands.

Description

Detect the frequency spectrum sensing method of angle of arrival of authorized user in a kind of cognitive radio

Technical field

The present invention relates to a kind of cognitive radio system intermediate frequency spectrum cognitive method, relate to a kind of method that can detect the authorization user signal and the angle of arrival more specifically.

Background technology

Along with the fast development of wireless communication technology and greatly enriching of wireless traffic, the radio subscriber that increases day by day makes frequency spectrum resource more and more nervous, and present used frequency spectrum only accounts for 2% to 6% of all usable spectrums.The cognitive radio technology of management of frequency spectrum resource is just becoming the focus that addresses this problem dynamically.

Cognitive radio is as a kind of frequency spectrum share technology of intelligence; Can rely on the support of artificial intelligence; The cognitive radio communication environment, according to certain study and decision making algorithm, real-time adaptive ground changes system operational parameters; Dynamically detect and effectively utilize idle frequency spectrum, allow on time, frequency and space, to carry out the spectrum reuse of multidimensional in theory.

And the frequency spectrum perception technology is one of key technology of cognitive radio system, and the purpose of frequency spectrum perception is to keep watch on and detect the active situation of authorization user signal on the special frequency, and when detecting this frequency range during the free time, cognitive user can be used this frequency range; When authorization user signal occurred once more, cognitive user must withdraw from this frequency range at the appointed time.In order authorized user not to be caused harmful interference, this just requires cognitive user can intercept frequency spectrum in real time continuously, to improve the reliability that detects.

In general, the frequency spectrum perception technology can be divided into emission source detection, cooperative detection and Interference Detection.But present method can not detect the multidimensional information of authorized user, and method itself also has certain restriction.The researcher is trying to explore the new technology of frequency spectrum perception more reliably at present.

Summary of the invention

The present invention proposes a kind of authorized user that can detect and exist, and the multidimensional frequency spectrum perception algorithm of the user's that can obtain the authorization angle of arrival information.The present invention adopts following technical scheme to realize:

(1) form linear array antenna by M array element and receive signal, at a time between in the section, the signal of l array element reception is x _l(t) (l=1,2 ... M).

(2) be arranged in a column vector to M signal, vector form is expressed as X (t)=AS (t)+N (t).

(3) each signal with X (t) carries out the sampling that hits is L, passes through the band pass filter of a frequency range to be detected then.Can receive the sampling array data X of signal ₁(t).

(4) ask sampling array data X ₁(t) covariance matrix,

(5) right

Carrying out feature decomposition has, In the formula, U _SBe the signal subspace that big characteristic value characteristic of correspondence vector is opened, U _NIt is the noise subspace that little characteristic value characteristic of correspondence vector is opened.

(6) can compose estimation formulas by steering vector and noise subspace does,

With spatial spectrum P _MUSICAssembly average as judging parameter because the signal parameter hunting zone of linear array is [90 °, 90 °], so

(7) try to achieve through test and satisfy false alarm probability P _fThreshold value γ.

(8) λ and threshold value γ are compared, exist if λ, explains main subscriber signal greater than threshold value γ; Otherwise authorization user signal does not exist.

(9) when authorization user signal exists, the angle of arrival of authorized user realizes with minimum optimization searching, i.e. θ _MUSIC=arg _θMina ^H(θ) U _NU _N ^HA (θ).

Advantage of the present invention is to have proposed a kind of frequency spectrum perception technology that can detect the authorized user multidimensional information.Whether typical frequency spectrum perception technology can only detect authorized user and exist, and the present invention is when detecting the signal of authorized user, the user's that can obtain the authorization the angle of arrival.And this method with the assembly average of spatial spectrum as judging parameter, do not rely on higher signal to noise ratio, than traditional energy measuring better detection performance is arranged.

Description of drawings

Fig. 1 is the flow chart of this method.

Fig. 2 has shown that respectively under the situation of different antennae number, signal to noise ratio is to the influence of angle of arrival estimated mean-square.

Fig. 3 be the present invention under the condition of signal to noise ratio snr=-5, ROC curve when different antenna element is counted and ROC curve contrast that can quantity algorithm.

Fig. 4 is the influence of signal to noise ratio to detection probability.

Embodiment

Come the present invention is done further detailed description below in conjunction with accompanying drawing and instantiation.

(1) as shown in Figure 1, at first receive the interior arrowband far-field signal of section sometime by the linear multiaerial system of M array element, because the signal phase that each array element receives is different, can reach the purpose of performance optimization through rational combined treatment.Suppose that the signal that l array element receives does

$x_l\left(t\right)=\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{g}_{\mathrm{li}}{s}_i(t-{\mathrm{\τ}}_{\mathrm{li}})+n_l\left(t\right)l=\mathrm{1,2},...M$

In the formula, g _LiBe of the gain of l array element to i signal, n _l(t) l array element of expression supposes that at t noise constantly noise is a Gaussian noise in the environment, and average is zero, and variance is σ ²τ _LiRepresent the time delay of l array element with respect to reference array element.

(2) signal that M array element was received in particular moment is arranged in a column vector, can get

$\left[\begin{array}{c}{x}_1\left(t\right)\\ x_2\left(t\right)\\ .\\ .\\ .\\ x_3\left(t\right)\end{array}\right]=\left[\begin{array}{cccc}{g}_{11}{e}^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{11}}& g_{12}{e}^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{12}}& ...& g_{1N}{e}^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{1N}}\\ g_{21}{e}^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{21}}& g_{22}{e}^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{22}}& ...& g_{2N}{e}^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{2N}}\\ .& .& .& .\\ .& .& .& .\\ .& .& .& .\\ g_{M1}{e}^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{M1}}& g_{M2}{e}^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{\mathrm{<figure-callout\; id="2"\; label="M"\; filenames="HSA00000047394400021.png"\; state="\{\{state\}\}">M</figure-callout>}2}}& ...& g_{\mathrm{MN}}{e}^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{\mathrm{MN}}}\end{array}\right]\left[\begin{array}{c}{s}_1\left(t\right)\\ s_2\left(t\right)\\ .\\ .\\ .\\ s_N\left(t\right)\end{array}\right]+\left[\begin{array}{c}{n}_1\left(t\right)\\ n_2\left(t\right)\\ .\\ .\\ .\\ n_M\left(t\right)\end{array}\right]$

In the formula, ω ₀For receiving the frequency of signal, suppose that each array element is isotropic and does not exist passage inconsistent in the array, the influence of factors such as mutual coupling, then the gain in the following formula can be omitted (promptly being normalized to 1), and like this, following formula can be reduced to

$\left[\begin{array}{c}{x}_1\left(t\right)\\ x_2\left(t\right)\\ .\\ .\\ .\\ x_3\left(t\right)\end{array}\right]=\left[\begin{array}{cccc}{e}^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{11}}& e^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{12}}& ...& e^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{1N}}\\ e^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{21}}& e^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{22}}& ...& e^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{2N}}\\ .& .& .& .\\ .& .& .& .\\ .& .& .& .\\ e^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{M1}}& e^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{\mathrm{<figure-callout\; id="2"\; label="M"\; filenames="HSA00000047394400021.png"\; state="\{\{state\}\}">M</figure-callout>}2}}& ...& e^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{\mathrm{MN}}}\end{array}\right]\left[\begin{array}{c}{s}_1\left(t\right)\\ s_2\left(t\right)\\ .\\ .\\ .\\ s_N\left(t\right)\end{array}\right]+\left[\begin{array}{c}{n}_1\left(t\right)\\ n_2\left(t\right)\\ .\\ .\\ .\\ n_M\left(t\right)\end{array}\right]$

Being write as vector form does, X (t)=AS (t)+N (t).In the formula, X (t) is the M * 1 dimension sampled data vector of array, and N (t) is the M * 1 dimension noise data vector of array, and S (t) is N * 1 n dimensional vector n of signal, and A is the M * N dimension steering vector battle array of space array, and

A＝[a ₁(ω ₀)a ₂(ω ₀)…a _N(ω ₀)]， $a_i\left({\mathrm{\&omega;}}_0\right)=\left[\begin{array}{c}{e}^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{1i}}\\ e^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{2i}}\\ .\\ .\\ .\\ e^{-j{\mathrm{\&omega;}}_0{\mathrm{\&tau;}}_{\mathrm{Mi}}}\end{array}\right]$

(3) each signal with X (t) carries out the sampling that hits is L, passes through the band pass filter of a frequency range to be detected then.Can receive the sampling array data X of signal ₁(t).

(4) ask sampling array data X ₁(t) covariance matrix,

(5) right

Carrying out feature decomposition has, In the formula, U _SBe the signal subspace that big characteristic value characteristic of correspondence vector is opened, U _NIt is the noise subspace that little characteristic value characteristic of correspondence vector is opened.

(6) calculate judgement parameter λ; Mountain steering vector and noise subspace can be composed estimation formulas; we with the assembly average of spatial spectrum as judging parameter; Again because the signal parameter hunting zone of linear array is [90 °; 90 °], so

(7) try to achieve through test and satisfy false alarm probability P _fThreshold value γ.Under varying environment, threshold value is different.Different time, frequency range or geographical position all can impact threshold value.So confirming of a certain environment Xiamen limit value need have only there not being authorized user under the situation of white Gaussian noise, obtains through a large amount of calculating to observation data.Concrete implementation method does, reuses the step of (1)～(6), obtains the set { γ of the assembly average of corresponding spatial noise spectrum ₁, γ ₂γ _N, utilize the resampling bootstrapping method in the statistics to be met false alarm probability P _fThreshold value γ.

(8) λ and threshold value γ are compared, if λ, explains the measured value that receives signal greater than threshold value γ on significance level (being false alarm probability), then authorization user signal exists; Otherwise as λ during less than threshold value γ, authorization user signal does not exist.

(9) when when confirming that authorized user exists, can asking the angle of arrival of authorized user by (8), the angle of arrival of authorized user realizes with minimum optimization searching, i.e. θ _MUSIC=arg _θMina ^H(θ) U _NU _N ^HA (θ).

Following instance emulation the random signal under QPSK (QPSK) modulation under Gaussian noise (average is 0, and variance is 1) environment, the detection performance of algorithm and angle of arrival estimated accuracy.The hypothesis information source number is less than the bay number in the simulation process.

Fig. 2 has shown that respectively under the situation of different antennae number, signal to noise ratio is to the influence of angle of arrival estimated mean-square.As can be seen from the figure, along with the increase of signal to noise ratio, angle of arrival estimated accuracy is high more.And the increase of bay number also can improve the precision that the angle of arrival is estimated.

Fig. 3 be the present invention under the condition of signal to noise ratio snr=-5, the ROC curve when different antenna element is counted.The ROC curve that has added energy measuring among the figure is to contrast.Under the condition of identical signal to noise ratio; Detection performance of the present invention obviously is superior to energy measuring, and under the situation of different antenna element number, the ROC curve does not have significant change; So under the prerequisite that guarantees the precision that the angle of arrival is estimated, the bay number can reduce as far as possible.

Fig. 4 is the influence of signal to noise ratio to detection probability.

Claims (1)

1. detect the frequency spectrum sensing method of angle of arrival of authorized user in the cognitive radio, it is characterized in that: the concrete performing step of said method is following:

(1) form linear array antenna by M array element and receive signal, at a time between in the section, the signal of l array element reception is x _l(t), l=1,2 ... M;

(2) be arranged in a column vector to M signal, vector form is expressed as X (t)=AS (t)+N (t); In the formula, X (t) is the M * 1 dimension sampled data vector of array, and N (t) is the M * 1 dimension noise data vector of array, and S (t) is N * 1 n dimensional vector n of signal, and A is the M * N dimension steering vector battle array of space array;

(3) each signal with X (t) carries out the sampling that hits is L, then through the band pass filter of a frequency range to be detected, can receive the sampling array data X of signal ₁(t);

(4) ask sampling array data X ₁(t) covariance matrix,

(5) right

Carrying out feature decomposition has,

In the formula, U _SBe the signal subspace that big characteristic value characteristic of correspondence vector is opened, U _NIt is the noise subspace that little characteristic value characteristic of correspondence vector is opened;

(6) can compose estimation formulas by steering vector and noise subspace does,

With spatial spectrum P _MUSICAssembly average as judging parameter

Because the signal parameter hunting zone of linear array is [90 °, 90 °], so

(7) try to achieve through test and satisfy false alarm probability P _fThreshold value γ;

(8)

compared with threshold value γ; If

, explains that authorization user signal exists greater than threshold value γ; Otherwise authorization user signal does not exist;

(9) when authorization user signal exists, the angle of arrival of authorized user realizes with minimum optimization searching, promptly

θ _MUSIC＝arg _θmina ^H(θ)U _NU _N ^Ha(θ)。

Images