CN103281141A - Blind spectrum sensing method and device - Google Patents

Abstract

The invention relates to the technical field of wireless communication and discloses a blind spectrum sensing method and a blind spectrum sensing device. The method comprises the steps that spectrum sensing equipment receives signals at an authorized frequency band, the self-correlation coefficient of the received signals is calculated after the received signals are sampled and filtered, the covariance matrix of the signals is constructed, the approximate 2-norm condition number T of the covariance matrix and finally whether signals of authorized users exist or not are judged according to the T. The blind spectrum sensing method and the blind spectrum sensing device disclosed by the invention have the advantages that the calculation complexity is low, the signal characteristic is not required to be authorized, the device is not sensitive to noise uncertainty and the performance is excellent.

Description

A kind of blind frequency spectrum sensing method and device

Technical field

The present invention relates to the frequency spectrum perception technology in the cognitive radio system, relate in particular to a kind of blind frequency spectrum sensing method that need not to send any characteristic information of signal.

Background technology

Along with society and expanding economy, people are increasing to the demand of wireless communication data amount.Yet frequency spectrum resource is limited, and at present, most frequency range has been distributed to authorized user, and this frequency spectrum that causes distributing to new wireless system is very few.And on the other hand, actual measurement shows that the mandate frequency range that these have distributed is not fully used, and their mosts of the time all are in the frequency range of idle condition, particularly some function admirable, and utilance is very low.The scarcity of frequency spectrum resource and the contradiction of poor efficiency have been impelled the proposition of cognitive radio.The thought of cognitive radio is that cognitive user is passed through frequency spectrum perception, and then knows the frequency spectrum operating position, and the frequency sub-band of selecting authorized user temporarily not have to use transmits.

As seen, frequency spectrum perception is a basic module of cognitive radio system.Frequency spectrum detecting method commonly used comprises: energy measuring, matched filtering detection and cyclo-stationary detect.The energy measuring complexity is low, but is subjected to probabilistic influence of noise, and its mis-behave is serious.Matched filtering excellent performance, but the feature of the known transmission signal of needs.The performance that cyclo-stationary detects is also very excellent, but complexity is higher, when practical application, is subjected to certain limitation.

Can judge on this frequency range it is signal or noise according to the covariance matrix that receives signal.Traditional technical scheme has proposed a kind of characteristic value of the covariance matrix that receives signal of utilizing and has carried out frequency spectrum perception, and judgment variables can be made of the characteristic value of covariance matrix, and has provided a kind of judgment variables building method, the i.e. ratio of minimax characteristic value.As can be seen, ideally, when having only noise to exist, the ratio of this minimax characteristic value is 1, and when having signal to exist, this ratio is greater than 1.The advantage of this method is any prior information that need not to send signal, also need not any statistical property of noise.But finding the solution the minimax characteristic value needs the complex features value to decompose, and computation complexity is very high, and Project Realization is very difficult.

Summary of the invention

Technical problem: in order to overcome existing energy measuring to the uncertain influence of noise, the present invention proposes a kind of blind frequency spectrum sensing method and device that utilizes the covariance matrix of wireless signal.

Technical scheme: a kind of blind frequency spectrum sensing method comprises the steps:

(1) receives the wireless signal for the treatment of on the perception frequency range;

(2) carry out sampling filter to received signal, calculate the auto-correlation coefficient of signal, and the structure covariance matrix, being designated as R, its dimension is L;

N behind a sampling filter sample signal is expressed as, x (0), and x (1) ..., x (N-1), described coefficient correlation obtains by the following method: choosing calculation window length is L, from l=0 ..., L-1, the coefficient correlation of calculating sample signal

${\mathrm{\λ}}_l=\frac{1}{N}\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}x\left(n\right)x^{*}(n-l)$

Wherein, when n-l＜0, x (0)=0, oeprator * represents to ask conjugation, and number of samples N is the positive integer greater than 1, and length of window L is the positive integer greater than 1; Corresponding covariance matrix R is expressed as,

$R=\left[\begin{array}{cccc}{\mathrm{\λ}}_0& {\mathrm{\λ}}_1& \·\·\·& {\mathrm{\λ}}_{L-1}\\ {\mathrm{\λ}}_1^{*}& {\mathrm{\λ}}_0& \·\·\·& {\mathrm{\λ}}_{L-2}\\ \·& \·& & \·\\ \·& \·& \·\·\·& \·\\ \·& \·& & \·\\ {\mathrm{\λ}}_{L-1}^{*}& {\mathrm{\λ}}_{L-2}^{*}& \·\·\·& {\mathrm{\λ}}_0\end{array}\right].$

(3) calculate the conditional number that covariance matrix R is similar under two norms, the structure judgment variables is designated as T; The building method of described T is:

Step I: to the summation of all diagonal entries of matrix R and divided by L, remember that this value is A;

Step I i: square sum of compute matrix R all elements, then divided by L, deduct again A square, ask evolution again, its result is designated as B;

Step I ii: calculate B divided by

Its result is designated as C;

Step I v: calculate B and multiply by Its result is designated as D;

Step v: calculate T ₁=A+C calculates T ₂If=A-D is T ₂Smaller or equal to zero, then to T ₂Give a very little number, obtain:

$T=\frac{T_1}{{\mathrm{<figure-callout\; id="2"\; label="T"\; filenames="HDA00003244302800012.png"\; state="\{\{state\}\}">T</figure-callout>}}_2}=\frac{A+C}{A-D}.$

The calculating of A and B obtains by the following method:

A＝λ ₀, $B=\sqrt{\frac{2}{L}\underset{i=1}{\overset{L-1}{\mathrm{\&Sigma;}}}(L-i){{\mathrm{\&lambda;}}_i}^2}.$

(4) judge according to judgment variables T whether frequency spectrum idle: when T greater than predefined decision threshold, then judging has authorization signal to exist on this frequency spectrum, when T less than predefined decision threshold, then judging does not have authorization signal, i.e. this frequency spectrum free time.

Number of samples N determines according to the cycle of frequency spectrum perception, the precision of frequency spectrum perception.Length of window L determines according to computing capability, frequency spectrum perception precision and the complexity of awareness apparatus.

Described decision threshold obtains by theory or emulation according to desired false alarm probability or detection probability.

This method is applicable to the frequency spectrum perception of multiaerial system or the cooperation perception of multinode.

A kind of blind frequency spectrum sensing device comprises: wireless signal samples and filtration module, auto-correlation coefficient computing module, judgment variables computing module and judging module;

Described wireless signal samples and filtration module are used for obtaining the wireless signal of institute's perception frequency range;

Described auto-correlation coefficient computing module is used for calculating the auto-correlation coefficient of signal;

Described judgment variables computing module is used for calculating the conditional number that covariance matrix R is similar under two norms, the structure judgment variables;

Described judging module comprises comparator, is used for relatively judgment variables and thresholding.

The present invention adopts technique scheme, has following beneficial effect: in the judgment variables calculation stages, this method only needs the lower add operation of complexity and a spot of multiplication division arithmetic, compare in the background technology frequency spectrum perception based on characteristic value, complexity is very low, and uncertain insensitive to noise.And the present invention also is applicable to frequency spectrum perception and the collaborative spectrum sensing of multiaerial system.

Description of drawings

Fig. 1 is the frequency spectrum sensing method flow chart of the embodiment of the invention;

Fig. 2 is the judgment variables computational methods schematic diagram of the embodiment of the invention;

Fig. 3 is the frequency spectrum sensing device block diagram of the embodiment of the invention;

Fig. 4 is the judgment variables computing module schematic diagram of the embodiment of the invention;

Fig. 5 is at the wireless microphone signal, and the performance of embodiment of the invention method and background technology is schematic diagram relatively;

Fig. 6 is at the Stochastic Modulation signal, and the performance of embodiment of the invention method and background technology is schematic diagram relatively.

Embodiment

Below in conjunction with specific embodiment, further illustrate the present invention, should understand these embodiment only is used for explanation the present invention and is not used in and limits the scope of the invention, after having read the present invention, those skilled in the art all fall within the application's claims institute restricted portion to the modification of the various equivalent form of values of the present invention.

A kind of blind frequency spectrum sensing method may further comprise the steps:

1) receives the wireless signal for the treatment of on the perception frequency range;

2) carry out sampling filter to received signal after, calculate the auto-correlation coefficient of signal;

N behind a sampling filter sample signal is expressed as, x (0), and x (1) ..., x (N-1).When reality realized, covariance matrix smoothly obtained usually by the following method.Choosing calculation window length is L, and the auto-correlation coefficient of signal can be expressed as,

${\mathrm{\&lambda;}}_l=\frac{1}{N}\underset{n=0}{\overset{N-1}{\mathrm{\&Sigma;}}}x\left(n\right)x^{*}(n-l)$ [formula 1]

Wherein, l=1,2 ..., L, corresponding covariance matrix is designated as R,

$R=\left[\begin{array}{cccc}{\mathrm{\&lambda;}}_0& {\mathrm{\&lambda;}}_1& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{\&lambda;}}_{L-1}\\ {\mathrm{\&lambda;}}_1^{*}& {\mathrm{\&lambda;}}_0& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{\&lambda;}}_{L-2}\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;\&CenterDot;\&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ {\mathrm{\&lambda;}}_{L-1}^{*}& {\mathrm{\&lambda;}}_{L-2}^{*}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{\&lambda;}}_0\end{array}\right]$ [formula 2]

Wherein, number of samples N is the positive integer greater than 1, and length of window L is the positive integer more than or equal to 1.R is the conjugation symmetrical matrix as can be seen.Use r _{I, j}The capable j column element of i of expression R, wherein, i=1 ..., L, j=1 ..., L.

We know that the ratio of minimax characteristic value is also referred to as two norm condition numbers.Decompose for fear of characteristic value, we propose to utilize two norm condition numbers of approximate matrix as judgment variables.Its basic thought is, utilizes the covariance matrix that receives signal, directly calculates its approximate conditional number under two norms, as judgment variables, and then judges whether there is signal according to this approximation.This method can be avoided the characteristic value decomposition operation of matrix fully.The approximation method of the conditional number of matrix under two norms has a lot, below we provide a certain embodiments.

According to linear algebra theory [Henry Wolkowicz, " Bounds for Eigenvalues Using Traces ", Linear Algebra and Its Applications, vol.29, pp.471-506,1980.], the conditional number of matrix R under two norms can be approximated to be T, is expressed as

$T=\frac{A+B/{(L-1)}^{1/2}}{A-B{(L-1)}^{1/2}}$ [formula 3]

Wherein, A=Tr (R)/L,

Matrix trace is asked in arithmetic operation Tr () expression.Be judgment variables with T, can avoid the matrix exgenvalue decomposition operation, complexity reduces greatly.According to the character of matrix trace,

$\mathrm{Tr}\left(R\right)=\underset{i=1}{\overset{L}{\mathrm{\&Sigma;}}}{r}_{i,i}$

$\mathrm{Tr}\left(R^2\right)=\underset{i=1}{\overset{L}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{L}{\mathrm{\&Sigma;}}}{r}_{i,\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="HDA00003244302800012.png"\; state="\{\{state\}\}">j</figure-callout>}}^2$

Utilize above-mentioned thought, we are 3) – 9 as follows) obtain judgment variables.

3) frequency spectrum sensing device remembers that to the summation of all diagonal entries of matrix R and divided by L this value is A;

4) square sum of frequency spectrum sensing device compute matrix R all elements, then divided by L, deduct again A square, ask evolution again, its result is designated as B;

5) frequency spectrum sensing device calculate B divided by

Its result is designated as C;

6) frequency spectrum sensing device calculating B multiply by

Its result is designated as D;

7) frequency spectrum sensing device calculates judgment variables T ₁: T ₁=A+C

8) frequency spectrum sensing device calculates judgment variables T ₂: T ₂=A-D

9) last, calculate decision threshold T=T ₁/ T ₂When T greater than predefined thresholding, then described frequency spectrum sensing device is judged has authorization signal to exist on this frequency spectrum, when T less than predefined thresholding, then described frequency spectrum sensing device is judged does not have authorization signal, i.e. this frequency spectrum free time.

As preferred version, according to the covariance matrix R shown in [formula 2], the calculating of A and B can obtain by the following method fast:

A=λ ₀[formula 4]

$B=\sqrt{\frac{2}{L}\underset{i=1}{\overset{L-1}{\mathrm{\&Sigma;}}}(L-i){{\mathrm{\&lambda;}}_i}^2}$ [formula 5]

As can be seen from the above, the calculating of numerical value B needs 2 (L-1)+1 time multiplyings and 1 extracting operation, and the calculating of judgment variables T needs 2 (L-1)+4 time multiplication and twice extracting operation altogether so, and (complexity is O (L to compare the method that characteristic value decomposes ³)), the complexity of the inventive method is very low.

Said method can also be in conjunction with many antennas sensory perceptual system (or multi-user cooperate perception).Only the calculating of coefficient correlation need be generalized to many antennas perception (or multi-user cooperate perception) system gets final product.System has K root antenna (or K cooperation perception user) to receive.Suppose that k root antenna (collaboration user) is expressed as y n sampled signal constantly _k(n), they can be arranged the signal phasor that is constructed as follows, y ₁(0), y ₂(0) ..., y _K(0), y ₁(1), y ₂(1) ..., y _K(1) ..., y ₁(N-1), y ₂(N-1) ..., y _K(N-1), this vector length is N * K.With above-mentioned y _k(n) to become length be the vector x (0) of N * K in the vector representation of Gou Chenging, x (1) ..., x (NK-1) also can calculate coefficient correlation according to this vector, and then obtains corresponding judgment variables.

For the distributed collaborative perception, also can adopt this method to carry out perception each perception user respectively, then sensing results is sent to the data fusion center, obtain more accurate judgement.

Below in conjunction with accompanying drawing, the job step of frequency spectrum sensing method of the present invention is further described.

As shown in Figure 1, at first, frequency spectrum sensing device receives the wireless signal for the treatment of on the perception frequency range, after carrying out sampling filter to received signal, calculates its coefficient correlation.Be that judgment variables is adjudicated according to the approximate condition number of covariance matrix under two norms then.When judgment variables greater than predefined thresholding, then judging has authorization signal to exist on this frequency spectrum, when judgment variables less than predefined thresholding, then judging does not have authorization signal, i.e. this frequency spectrum free time.

Embodiment as shown in Figure 2 can obtain A and B respectively according to [formula 4] and [formula 5], and then obtain the numerical value of C and D, thereby obtain judgment variables T.

As shown in Figure 3, frequency spectrum sensing device of the present invention comprises: wireless signal samples and filtration module, coefficient correlation computing module, judgment variables computing module and judging module.Wireless signal samples and filtration module are used for obtaining the wireless signal of institute's perception frequency range; The auto-correlation coefficient computing module is used for calculating the auto-correlation coefficient of signal; The judgment variables computing module is used for calculating the conditional number that covariance matrix R is similar under two norms, the structure judgment variables; Judging module comprises comparator, is used for relatively judgment variables and thresholding.Wherein, the realization block diagram of judgment variables computing module as shown in Figure 4.

Below by the actual emulation checking, provided the performance comparison of present embodiment and background technology by Fig. 5 and Fig. 6.Energy measuring, the background technology that we have contrasted under the noiseless uncertainty is the detection of minimax characteristic value ratio, the energy measuring under the noise uncertainty and the art of this patent.The art of this patent and minimax characteristic value ratio detect all insensitive to the noise uncertainty, just, exist the noise uncertainty whether not influence their performance.What provide among the figure is to exist under the noise uncertainty, the performance that the art of this patent and minimax characteristic value ratio detect.The noise coefficient of uncertainty is 0.5dB.

The simulation result of Fig. 5 is that the frequency spectrum detection with the wireless microphone signal is example, and sampling number is 4096 points.As can be seen from the figure, comparing the energy measuring under the noiseless uncertainty, is 0.9 o'clock in detection probability, and this method has the gain of 2dB, and in addition, this method is better than minimax characteristic value ratio and detects about 0.2dB.It should be noted that energy measuring has serious " signal to noise ratio wall " phenomenon when existing noise uncertain, worsen more than the 7dB than this method.This explanation, this method not only computation complexity are lower than the detection of minimax characteristic value ratio, and the ratio that performance also is better than the minimax characteristic value detects.

The simulation result of Fig. 6 is that the frequency spectrum detection with random signal is example.Adopt the QPSK signal, frequency selectivity Rayleigh fading channel, constant power 5 footpath channels, single transmit antenna, single reception antenna, sampling number are 4096 points.For random signal, this method slightly is better than the minimax characteristic value and detects, and the performance gain of 0.2dB is nearly arranged, and they all are inferior to energy measuring, the energy measuring when still still being much better than to exist noise uncertain.

One of ordinary skill in the art will appreciate that all or part of step in the said method can instruct related hardware to finish by program, described program can be stored in the computer-readable recording medium, as read-only memory, disk or CD etc.Alternatively, all or part of step of above-described embodiment also can use one or more integrated circuits to realize.Correspondingly, each the module/unit in above-described embodiment can adopt the form of hardware to realize, also can adopt the form of software function module to realize.The present invention is not restricted to the combination of the hardware and software of any particular form.

Claims (9)

1. a blind frequency spectrum sensing method is characterized in that, comprises the steps:

(1) receives the wireless signal for the treatment of on the perception frequency range;

(2) carry out sampling filter to received signal, calculate the auto-correlation coefficient of signal, and the structure covariance matrix, being designated as R, its dimension is L;

(3) calculate the conditional number that covariance matrix R is similar under two norms, the structure judgment variables is designated as T;

(4) judge according to judgment variables T whether frequency spectrum idle: when T greater than predefined decision threshold, then judging has authorization signal to exist on this frequency spectrum, when T less than predefined decision threshold, then judging does not have authorization signal, i.e. this frequency spectrum free time.

2. a kind of blind frequency spectrum sensing method according to claim 1, it is characterized in that: the sample signal of the N behind the sampling filter is expressed as, x (0), x (1) ... x (N-1), described coefficient correlation obtains by the following method: choosing calculation window length is L, from l=0 ... L-1, the coefficient correlation of calculating sample signal

${\mathrm{\&lambda;}}_l=\frac{1}{N}\underset{n=0}{\overset{N-1}{\mathrm{\&Sigma;}}}x\left(n\right)x^{*}(n-l)$

Wherein, when n-l＜0, x (0)=0, oeprator * represents to ask conjugation, and number of samples N is the positive integer greater than 1, and length of window L is the positive integer greater than 1; Corresponding covariance matrix R is expressed as,

$R=\left[\begin{array}{cccc}{\mathrm{\&lambda;}}_0& {\mathrm{\&lambda;}}_1& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{\&lambda;}}_{L-1}\\ {\mathrm{\&lambda;}}_1^{*}& {\mathrm{\&lambda;}}_0& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{\&lambda;}}_{L-2}\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;\&CenterDot;\&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& & \&CenterDot;\\ {\mathrm{\&lambda;}}_{L-1}^{*}& {\mathrm{\&lambda;}}_{L-2}^{*}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{\&lambda;}}_0\end{array}\right].$

3. a kind of blind frequency spectrum sensing method according to claim 2 is characterized in that: number of samples N determines according to the cycle of frequency spectrum perception, the precision of frequency spectrum perception.

4. a kind of blind frequency spectrum sensing method according to claim 2 is characterized in that: length of window L, determine according to computing capability, frequency spectrum perception precision and the complexity of awareness apparatus.

5. a kind of blind frequency spectrum sensing method according to claim 2 is characterized in that the building method of described T is:

Step I: to the summation of all diagonal entries of matrix R and divided by L, remember that this value is A;

Step I i: square sum of compute matrix R all elements, then divided by L, deduct again A square, ask evolution again, its result is designated as B;

Step I ii: calculate B divided by

Its result is designated as C;

Step I v: calculate B and multiply by

Its result is designated as D;

Step v: calculate T ₁=A+C calculates T ₂If=A-D is T ₂Smaller or equal to zero, then to T ₂Give a very little number, obtain:

$T=\frac{T_1}{T_2}=\frac{A+C}{A-D}.$

6. a kind of blind frequency spectrum sensing method according to claim 5, it is characterized in that: the calculating of A and B obtains by the following method:

A＝λ ₀， $B=\sqrt{\frac{2}{L}\underset{i=1}{\overset{L-1}{\mathrm{\&Sigma;}}}(L-i){{\mathrm{\&lambda;}}_i}^2}.$

7. a kind of blind frequency spectrum sensing method according to claim 1, it is characterized in that: described decision threshold obtains by theory or emulation according to desired false alarm probability or detection probability.

8. a kind of blind frequency spectrum sensing method according to claim 1 is characterized in that: be applicable to the frequency spectrum perception of multiaerial system or the cooperation perception of multinode.

9. a blind frequency spectrum sensing device comprises: wireless signal samples and filtration module, auto-correlation coefficient computing module, judgment variables computing module and judging module; It is characterized in that,

Described wireless signal samples and filtration module are used for obtaining the wireless signal of institute's perception frequency range;

Described auto-correlation coefficient computing module is used for calculating the auto-correlation coefficient of signal;

Described judgment variables computing module is used for calculating the conditional number that covariance matrix R is similar under two norms, the structure judgment variables;

Described judging module comprises comparator, is used for relatively judgment variables and thresholding.

Images