CN106027179A - Wideband frequency spectrum sensing method based on comprehensive co-prime analysis and device thereof - Google Patents

Abstract

The invention discloses a wideband frequency spectrum sensing method based on comprehensive co-prime analysis and a device thereof. The sensing method comprises the following steps: processing input signals to acquire original co-prime spectrum output; processing signals obtained after twice down-sampling to acquire complementary co-prime spectrum output; acquiring a calculation result of a first indicator and a calculation result of a second indicator respectively by using energy detection results of the original co-prime spectrum output and the complementary co-prime spectrum output which are put through energy detectors; and acquiring an indicative power spectrum through the calculation result of the first indicator and the calculation result of the second indicator. The sensing device comprises a microcontroller, an output driver and a display circuit. The sensing method can be used to comprehensively detect the statistic distribution characteristics of energy output of an original co-prime spectrum analyzer and a complementary co-prime spectrum analyzer under the precondition of a rate far lower than the Nyquist rate and blind estimation.

Description

A kind of broader frequency spectrum cognitive method based on comprehensive coprime analysis and device thereof

Technical field

The present invention relates to digital processing field, particularly relate to a kind of broader frequency spectrum perception based on comprehensive coprime analysis Method and device thereof, be specifically related to signal lack sampling, coprime perception, multiphase filtering, spectrum estimation, wave filter design.

Background technology

Radio-frequency spectrum is the strategic resource of country, if effective exploitation and utilize this resource, so that it may create big quantity of material Wealth and initiation great social profit.The distribution of traditional radio spectrum resources is with the form licensed by government department Distribute to primary user (Primary Users, PU), but along with mobile communication, radio and television, WLAN, Internet of Things etc. Technical field is growing to Radio Spectrum Resource demand, and the scarcity of frequency spectrum resource has become as outstanding problem.

By the reply ever-increasing bandwidth demand that nowadays a large amount of portable radios and system are brought, cognition wireless (CR) power technology arises at the historic moment.As a promising wireless technology, by dynamically positioning blank frequency spectrum, cognition wireless Power technology can realize opportunistic transmission, thus utilize Radio Spectrum Resource efficiently and solve that the frequency spectrum that nowadays exists is congested asks Topic [1] [2].And the key technology in the unlimited power technology of cognition is exactly frequency spectrum perception, result based on frequency spectrum perception, cognition wireless Electricity system just can search and not utilize frequency band, thus carries out chance on the premise of not disturbing the user communicating transmission Property transmission.Due in typical cognitive radio scene, prior art there is no priori for the occupied information of current spectral and knows Knowing, it is the most necessary for therefore whole broadband being carried out perception.

Traditional frequency spectrum sensing method based on nyquist sampling has energy measuring, bank of filters frequency spectrum perception and many windows The methods such as Power estimation [3].But for broader frequency spectrum perception, if using traditional nyquist sampling method to whole broadband Sample, it will the sampling rate that demand is the highest, and analog-digital converter (Analog-Digital Converter) now Performance and cost requirement be also difficult to meet demand (being limited by high mode switching rate).

For this problem, the most many scholars start with compressed sensing [4] method and carry out composing perception, and such method is first Utilization is far below sub-Nyquist sample obtained by Nyquist rate sampling and carrys out reconstruction signal, the most again to the signal reconstructed Carrying out frequency spectrum perception, common signal reconfiguring method has base to follow the trail of (BP) [5] and greedy tracing algorithm (such as orthogonal matching pursuit OMP)[6].Compression sensing method can be divided into blind recovery and non-blind to recover according to the presence or absence of priori again, and non-blind recovers required Sampling rate be generally below the desired rate of blind recovery.With sub-nyquist sampling method [7] based on analog-digital converter group As a example by, blind restoration method [8] can recover M subband with at least 2M analog-digital converter in the case of being not required to sub-band information Signal.

But for spectrum aware application, enliven the position of frequency band owing to its objective is to determine, therefore there is no need first profit Original signal is recovered by methods such as compressed sensing.Save the process of signal reconstruction, sub-nyquist sampling just can be greatly improved Under frequency spectrum perception efficiency.According to this thinking, scholar's Wang Xiaodong utilizes the low rate sample of analog-digital converter group sampling gained to estimate Count out the power spectrum [9] of broadband signal.But its method remains a need for some prioris carrys out the one-tenth work recovery of guaranteed output spectrum.Cause This, find the sparse spectrum blind estimating method that can apply in practice, thoroughly break away from the constraint of high-speed sampler, be one and urgently solve A difficult problem certainly.

For solving sparse Power estimation problem, in recent years, a kind of novel Power estimation method coprime perception (co-prime Sensing) theory [10-13] receives more and more attention, and it is parallel that first single analog input signal is made two-way by the method Sparse sampling (requiring that two-way down-sampling factor M, N numerical value meet coprime relation), then sample sparse to the two-way obtained divides Do not make multiphase filtering, single channel signal is converted into multi-path subsignal, more simultaneously parallel to two-way multiphase filtering output letter Number being IDFT (Inverse DFT), finally the multi channel signals to IDFT output does cross-correlation and scans to estimate signal spectrum position Put, thus obtain High-Resolution Spectral truly.The method is because experienced by two-stage down-sampling (coprime down-sampling and heterogeneous filter Ripple down-sampling), therefore it is very suitable for low data-rate applications occasion, and simple in construction, algorithm is clear, solves dilute to a certain extent Dredge Power estimation problem.This lack sampling spectrum cognitive method shows the highest value [10-13] when analyzing narrow band signal.

Inventor, during realizing the present invention, finds at least to suffer from the drawback that in prior art and not enough:

When classical coprime cognitive method is used for analyzing broadband signal, the most serious cross term interference can be produced, The pseudo-peak that many places amplitude is bigger is shown as, so that the mid frequency of each wideband spectrum thickens not really on spectrogram Clearly.

List of references

[1]J.Mitola and G.Q.Maguire,“Cognitive radio:Making software radios more personal,”IEEE Pers.Commun.Mag.,vol.6,no.4,pp.13-18,1999.

[2]S.Haykin,“Cognitive radio:Brain empowered wireless communications,”IEEE Trans.Commun.,vol.23,no.2,pp.201–220,Feb.2005.

[3]D.D.Ariananda,M.K.Lakshmanan,and H.Nikoo,“A survey on spectrum sensing techniques for cognitive radio,”in Proc.2nd Int.Workshop Cognit.Radio Adv.Spectrum Manag.(CogART),2009,pp.74–79.

[4]D.Donoho,“Compressed sensing,”IEEE Trans.Inf.Theory,vol.52,no.4, pp.1289–1306,2006.

[5]S.Chen,D.donoho,and M.Saunders,“Atomic decomposition by basis pursuit”,SIAM Rev,vol.43,no.1,pp.129-159,2001.

[6]J.Tropp and A.C.Gilbert,“Signal recovery from random measurements via orthogonal matching pursuit”,IEEE Trans.Inform.Theory,vol.53,Issue:12, 2007.

[7]R.Venkataramani and Y.Bresler,“Perfect reconstruction formulas and bounds on aliasing error in sub-Nyquist nonuniform sampling of multiband signalss,”IEEE Trans.Inf.Theory,vol.46,no.6,pp.2173–2183,Sep.2000.

[8]M.Mishali and Y.C.Eldar,“Blind multiband signal reconstruction: Compressed sensing for analog signals,”IEEE Trans.Signal Process.,vol.57, no.3,pp.993–1009,Mar.2009.

[9]C.P.Yen,Y.Tsai,and X.Wang,“Wideband spectrum sensing based on sub- Nyquist sampling,”IEEE Trans.Signal Process.,vol.61,pp.3028–3040,Jun.2013.

[10]Vaidyanathan P P,Pal P.Sparse sensing with co-pprime samplers and arrays[J].IEEE Transactions on Signal Processing,2011,59(2):573-586.

[11]Vaidyanathan P P,Pal P.Theory of sparse coprime sensing in multiple dimensions[J].Signal Processing,IEEE Transactions on,2011,59(8): 3592-3608.

[12]Vaidyanathan P P,Pal P.Sparse coprime sensing with multidimensional lattice arrays[C]//Digital Signal Processing Workshop and IEEE Signal Processing Education Workshop(DSP/SPE),2011IEEE.IEEE,2011:425- 430.

[13]Vaidyanathan P P,Pal P.Coprime Sampling and Arrays in One and Multiple Dimensions[M]//Multiscale Signal Analysis and Modeling.Springer New York,2013:105-137.

Summary of the invention

The invention provides a kind of broader frequency spectrum cognitive method based on comprehensive coprime analysis and device thereof, the present invention refers to It is described below:

A kind of broader frequency spectrum cognitive method based on comprehensive coprime analysis, described cognitive method comprises the following steps:

Input signal is processed, obtains former coprime spectrum output；

Signal after twice down-sampling is processed, obtains complementary coprime spectrum output；

Utilize the former coprime energy detection results composing output by energy detector and the energy of complementation coprime spectrum output Testing result tries to achieve result of calculation and the result of calculation of the second indicator of the first indicator respectively；

Indicative power spectrum is got by the result of calculation of the first indicator and the result of calculation of the second indicator.

Described input signal is processed, obtain the step of former coprime spectrum output particularly as follows:

Input signal is carried out two-way down-sampling, obtains two-way sparse signal；

Two-way sparse signal is carried out respectively multiphase filtering, exports two multichannel sparse signals, dilute by two multichannels Thin signal can respectively obtain the multichannel output sequence of correspondence；

In each moment, 2 road multichannel output sequences are carried out IDFT respectively, respectively obtains output sequence；

To output sequence cross-correlation, obtain former coprime spectrum output.

Described signal after twice down-sampling is processed, obtain the step of complementary coprime spectrum output particularly as follows:

To the signal Zhong Mei road signal after twice down-sampling respectively plus phase shift and carry out IDFT and respectively obtain output Sequence；

Ask for the cross-correlation of output sequence, obtain complementary coprime spectrum output.

Described utilization by energy detector former coprime spectrum output energy detection results and complementation coprime spectrum output Energy detection results try to achieve respectively the result of calculation of the first indicator and the result of calculation of the second indicator step particularly as follows:

Given false-alarm probability, sets threshold value, if frequency spectrum output will judge at spectral position i more than threshold value, energy detector For the most occupied；

If frequency spectrum output is less than or equal to threshold value, energy detector is unoccupied by being judged at spectral position i.

A kind of device of broader frequency spectrum perception based on comprehensive coprime analysis, described sensing device includes:

Microcontroller, for processing input signal, obtains former coprime spectrum output；To the letter after twice down-sampling Number process, obtain complementary coprime spectrum output；Utilize by energy detector former coprime spectrum output energy detection results, The energy detection results of spectrum output coprime with complementation tries to achieve result of calculation and the calculating of the second indicator of the first indicator respectively Result；Indicative power spectrum is got by the result of calculation of the first indicator and the result of calculation of the second indicator；

Output drives and display circuit, is used for showing indicative power spectrum.

Described microcontroller includes:

Former coprime spectralyzer, for processing input signal, obtains former coprime spectrum output；

Complementary coprime spectralyzer, for processing the signal after twice down-sampling, obtains complementary coprime spectrum defeated Go out.

The technical scheme that the present invention provides provides the benefit that: broadband based on the comprehensive coprime analysis frequency that the present invention proposes Spectrum cognitive method and device thereof, if for wideband spectrum perception and Practical Project field, can produce following beneficial effect:

The first, greatly the sampling rate that wideband spectrum perception needs is reduced；

Relative to traditional broader frequency spectrum cognitive method, this method not only eliminates unnecessary signal reconstruction step, moreover it is possible to The spectrum reconstruct of certain resolution is realized in the case of the sample using lack sampling.Owing to traditional cognitive method needs the Kui with how Signal in whole broadband is sampled by this special speed, then by signal reconstruction out, finally calculates its power spectrum.When frequency When bandwidth reaches the GHz order of magnitude, the Nyquist sampling rate of heritage perception method demand will exceed existing analog digital conversion The performance boundary that device can reach.As long as and ensure the snap of a number of feedthrough system simultaneously, this method just can will be sampled Rate reduction is to max (f_s/M,f_s/ N), wherein M, N can be with arbitrary sizes.Obviously this method can be greatly reduced required sampling Speed, thus reduce the cost of wideband spectrum perception.

Such as in experiment 1, only it is achieved that this needs 1.5GHz sampling rate to enter by the highest sampling rate of 45.45MHz The desired result that row sampling could realize.

The second, can accurately estimate to compose really the situation of taking；

It is f that the spectrum cognitive method that this method proposes can realize resolution_sThe spectrum of/MN takies estimation, and frequency number is MN.Visible Keeping certain f_sIn the case of, choose the biggest relatively prime integers to ensuring that the spectrum not missing some narrow band signal becomes Point.

By experiment 1 it can be seen that choose relatively prime integers just can be by all 12 sons with a width of 5MHz to M=37, N=33 Band signal gives location out.

3rd, there is higher noiseproof feature.

Coprime perception Power estimation method needs to carry out a certain amount of sample mean, therefore can cause certain time delay (each snap of signal can be overlapped, and time delay just can be greatly reduced), but the strongest suppression noise can be reached Effect；

Such as in experiment 2, at SNR=-50dB under such circumstances, analyzer remains to draw spectrum distribution clearly.

Accompanying drawing explanation

Fig. 1 is the flow chart of a kind of broader frequency spectrum cognitive method based on comprehensive coprime analysis；

Fig. 2 is the flow chart of former coprime spectralyzer；

Fig. 3 is the flow chart of the coprime spectralyzer of complementary type；

Fig. 4 is the low pass filter transmission curve figure of Remez algorithm design；

Fig. 5 is the result figure of coprime broader frequency spectrum perception；

Fig. 6 is the result figure of original coprime perception output；

Fig. 7 is the perception situation schematic diagram under SNR=0dB；

Fig. 8 is the perception situation schematic diagram under SNR=-50dB；

Fig. 9 is the hardware enforcement figure of the present invention；

Figure 10 is DSP internal processes flow graph.

Detailed description of the invention

For making the object, technical solutions and advantages of the present invention clearer, below embodiment of the present invention is made further Ground describes in detail.

In order to remove pseudo-effect present in art methods, the embodiment of the present invention proposes based on comprehensive coprime analysis Broader frequency spectrum cognitive method, former coprime spectrum can be divided on the premise of far below Nyquist rate and blind estimate by the method The statistical distribution characteristic of the energy output of parser and the coprime spectralyzer of complementation carries out comprehensive detection.

Embodiment 1

The flow process of the broader frequency spectrum cognitive method that the embodiment of the present invention proposes is as it is shown in figure 1, the method mainly includes signal Coprime perception process, energy-probe and frequency spectrum combine and calculate, described below:

101: input signal is processed, obtain former coprime spectrum output；

102: the signal after twice down-sampling is processed, obtain complementary coprime spectrum output；

103: utilize the former coprime energy detection results composing output by energy detector and the coprime spectrum of complementation to export Energy detection results tries to achieve result of calculation and the result of calculation of the second indicator of the first indicator respectively；

104: get indicative power spectrum by the result of calculation of the first indicator and the result of calculation of the second indicator.

Wherein, input signal is processed in step 101, obtain former coprime spectrum output step particularly as follows:

Input signal is carried out two-way down-sampling, obtains two-way sparse signal；

Two-way sparse signal is carried out respectively multiphase filtering, exports two multichannel sparse signals, dilute by two multichannels Thin signal can respectively obtain the multichannel output sequence of correspondence；

In each moment, 2 road multichannel output sequences are carried out IDFT respectively, respectively obtains output sequence；

To output sequence cross-correlation, obtain former coprime spectrum output.

Wherein, the signal after twice down-sampling is processed in step 102, obtain the step of complementary coprime spectrum output Rapid particularly as follows:

To the signal Zhong Mei road signal after twice down-sampling respectively plus phase shift and carry out IDFT and respectively obtain output Sequence；

Ask for the cross-correlation of output sequence, obtain complementary coprime spectrum output.

Wherein, utilizing in step 103 composes, by the former coprime of energy detector, the energy detection results and complementation exported The energy detection results of coprime spectrum output tries to achieve the result of calculation of the first indicator and the result of calculation of the second indicator respectively Step particularly as follows:

Given false-alarm probability, sets threshold value, if frequency spectrum output will judge at spectral position i more than threshold value, energy detector For the most occupied；

If frequency spectrum output is less than or equal to threshold value, energy detector is unoccupied by being judged at spectral position i.

In sum, the embodiment of the present invention proposes broader frequency spectrum cognitive method based on comprehensive coprime analysis, the method Can be on the premise of far below Nyquist rate and blind estimate, to former coprime spectralyzer and the coprime spectralyzer of complementation Energy output statistical distribution characteristic carry out comprehensive detection.

Embodiment 2

Below in conjunction with concrete computing formula, example, the scheme in embodiment 1 is described in detail, as detailed below retouches State:

201: the coprime perception of signal processes；

Step 1 carries out two-way down-sampling to input signal x (n), and the down-sampling factor is respectively M and N, and (M and N is coprime Integer), obtain two-way sparse signal x_u(n) and x_v(n)；

Step 2 is to x_u(n) and x_vN () carries out multiphase filtering respectively, export two multichannel sparse signals:WithWherein 0≤p≤M-1,0≤q≤N-1；

Wherein, two multichannel sparse signals the multichannel output sequence of correspondence can be respectively obtained

Step 3 is at each moment n, to multichannel output sequenceCarry out M point IDFT, obtain defeated Go out sequence { u_k(n), k=0 ..., M-1}.

Similarly, to multichannel output sequenceCarry out N point IDFT, obtain output sequence { v_l (n), l=0 ..., N-1}.

Step 4 is to the signal { x after twice down-sampling_u,p(n), p=0 ..., M-1} Zhong Mei road signal adds respectively Phase shift { (-1)ⁿe^jpπ/M, p=0 ..., M-1}, exportedRepeat Step 3 to be exported

In like manner can get the output of another road

Step 5 asks for the output sequence { u after IDFT_k(n), k=0 ..., M-1} and { v_l(n), l=0 ..., N-1}'s Cross-correlation, obtains former coprime spectrum outputAsk for output sequenceWithCross-correlation, obtain complementary coprime spectrum output

202: energy-probe；

Given false-alarm probability P_fa, set threshold value T_hAs shown in formula (1)

T_h=v Q^-1(P_fa)+σ² (1)

Wherein, v is the output of former coprime power spectrumSpectral position i adds up the standard deviation obtained；σ²It it is noise merit Rate；Q^-1X () is the inverse function of Q (x), then by P_faSubstitute into Q^-1In (x).

If frequency spectrum output P (i) ＞ T_h, energy detector is the most occupied by being judged at spectral position i；If frequency spectrum exports P(i)≤T_h, energy detector is unoccupied by being judged at spectral position i.Former coprime spectralyzer and the coprime analysis of spectrum of complementary type Device respectively obtains energy detection results by after energy-probeWith

203: frequency spectrum combines and calculates.

Utilize the energy detection results that the former coprime spectrum by energy detector exportsWith The energy detection results of complementary coprime spectrum outputIndicator ρ is tried to achieve respectively by formula (2)₁(i) and ρ₂(i)。

$\begin{array}{c}{\mathrm{\ρ}}_1\left(i\right)=Y_{xx}\left(e^{{\mathrm{j\ω}}_i}\right){\tilde{Y}}_{xx}\left(e^{{\mathrm{j\ω}}_{i+1}}\right)\\ {\mathrm{\ρ}}_2\left(i\right)=Y_{xx}\left(e^{{\mathrm{j\ω}}_i}\right){\tilde{Y}}_{xx}\left(e^{{\mathrm{j\ω}}_i}\right)\end{array}---\left(2\right)$

Finally combine indicator computing formula (3) and try to achieve frequency spectrum.

$P\left(i\right)=\sqrt{\frac{{\mathrm{\ρ}}_1\left(i\right)+{\mathrm{\ρ}}_2\left(i\right)}{2}}---\left(3\right)$

The overall signal of the present invention processes framework

Assume that broadband signal s (t) to be detected is made up of many subbands, i.e.

$s\left(t\right)=\underset{i=1}{\overset{m}{\Σ}}{s}_i\left(t\right)---\left(4\right)$

Then the observation signal in the perception cycle can be expressed as shown in formula (5), and wherein n (t) represents additive white Gaussian noise.

$x\left(t\right)=\underset{i=1}{\overset{m}{\Σ}}{s}_i\left(t\right)+n\left(t\right)---\left(5\right)$

Again observation signal is separately input to Fig. 2, in based on bank of filters the coprime spectralyzer shown in 3, and passes through Energy-probe, the output sequence followed by above-mentioned two-way constructs ρ₁(i), ρ₂(i), then obtain indicative power spectrum by formula (3) P(i)。

In sum, the embodiment of the present invention proposes broader frequency spectrum cognitive method based on comprehensive coprime analysis, the method Can be on the premise of far below Nyquist rate and blind estimate, to former coprime spectralyzer and the coprime spectralyzer of complementation Energy output statistical distribution characteristic carry out comprehensive detection.

Embodiment 3

Below in conjunction with concrete computing formula, the principle of scheme in embodiment 1,2 is described in detail, as detailed below Describe:

301: former coprime spectralyzer

In Fig. 2, input signal x (n) is imaginary Nyquist sampling, its sampling interval T=1/f_Nyq, respectively with down-sampled Factor N, M (requiring that both meet coprime relation) parallel down-sampling form two-way low rate stream x_u(n) and x_vN () (this is equivalent to use Two sampling rates are f_Nyq/N、f_NyqInput analogue signal is directly done discretization by the ADC of/M parallel)；Then to x_u(n) (or x_v (n)) do M road (or N road) poly phase, and respectively by heterogeneous subfilter E_p(z) and R_qZ () is filtered, more respectively to respectively The parallel output filtered mutually is the IDFT (or IDFT of N point) of M point, and finally to two-way IDFT, (i.e. correspondence is multiplied by IDFT matrix With) each branch road output carry out cross-correlation scan i.e. obtain coprime spectrum export

302: the coprime spectralyzer of complementary type

In order to obtain more intuitive broadband signal power spectrum, need to design a coprime spectralyzer of complementary type, in Fig. 3 Complementary coprime spectralyzer is compared to Fig. 2 situation, and it is identical that great majority process, and difference has only made following process:

1) alternately change x sample-by-sample_u,p(n) (or x_v,p(n)) symbol；

2) bank of filters E is being entered_p(z) (or R_q(z)) front, dephased for first step signal is multiplied by e^-jpπ/M(or e^-jpπ/N)。

May certify that, above process is equal to do whole input signal x (n) after frequency displacement size is 0.5 Δ f=π/MN, Obtain sequence complementary therewithOriginal coprime analysis of spectrum result

And then, in step below, to original coprime spectrum outputSpectrum output coprime with complementationDo comprehensive Close, cross term interference and pseudo-peak effect can be eliminated.

303: energy-probe

Respectively obtain from the output of former coprime spectralyzer and the coprime spectralyzer of complementary type WithOwing to, in wideband spectrum perception, number of sub-bands is numerous,WithCan be approximately Gaussian random sequence (from central limit theorem), therefore frequency spectrum detection problem can be expressed as the dualism hypothesis of each frequency Check problem:

$\begin{array}{c}{H}_0:S_{xx}\left(e^{{\mathrm{j\ω}}_i}\right),{\tilde{S}}_{xx}\left(e^{{\mathrm{j\ω}}_i}\right)~N({\mathrm{\σ}}^2,{\mathrm{\υ}}^2)\\ H_1:S_{xx}\left(e^{{\mathrm{j\ω}}_i}\right),{\tilde{S}}_{xx}\left(e^{{\mathrm{j\ω}}_i}\right)~N\left(P\right(i)+{\mathrm{\σ}}^2,{\mathrm{\υ}}^2)\end{array}---\left(6\right)$

Wherein, v is former coprime spectrum output(or) on spectral position i, add up the standard deviation calculated, σ²Be to Fixed noise power, P (i) is unknown signal power.Therefore by false-alarm probability P_faIt is set to:

$P_{fa}=Q\left(\frac{T_h-{\mathrm{\σ}}^2}{v}\right)---\left(7\right)$

Once it is determined that normal false-alarm probability, then threshold value T_hCan be calculated by formula (1).If(or) ＞ T_h, sound Bright for H₁, energy detector is the most occupied by being judged at frequency i；If(or)≤T_h, it is claimed as H₀, energy Amount detector is unoccupied by being judged at frequency i.

In sum, the embodiment of the present invention proposes broader frequency spectrum cognitive method based on comprehensive coprime analysis, the method Can be on the premise of far below Nyquist rate and blind estimate, to former coprime spectralyzer and the coprime spectralyzer of complementation Energy output statistical distribution characteristic carry out comprehensive detection.

Embodiment 4

Below in conjunction with concrete test data the scheme in embodiment 1,2 and 3 carried out feasibility checking, as detailed below retouch State:

(1) broader frequency spectrum perception checking

Experiment 1

Assuming in a cognitive radio system, the wideband spectrum scope that monitor is F=[f_min,f_max].In perception week Interim, there is m movable incoherent subband signal, i-th subband signal is given as following form:

$s_i\left(t\right)=\underset{n\∈Z}{\Σ}{d}_i\[n\]g_i(t-{\mathrm{nT}}_i)e^{j2{\mathrm{\πf}}_{i}t},i=1,2,\mathrm{...},m---\left(8\right)$

Wherein, { d_i[n] } it is modulation symbol sequence, g_iT () is pulse shaping function (T_iIt is symbol interval), g_i(t-nT_i) For the filter function added to each transmitted symbol, f_iIt is s_iThe carrier frequency of (t), it is assumed that each subband signal s₁(t),s₂ (t),...,s_mT () is separate and zero-mean.

In this experiment, if f_min=0, Nyquist rate f_s=f_max=1/T=1.5GHz.Subband bandwidth is all set to 5MHz (by symbol intervals τ_iDetermine), the normalized frequency of 12 road signal carriers is set to f₁=10 Δ f, f₂=130 Δ f, f₃=250 Δs f,f₄=370 Δ f, f₅=490 Δ f, f₆=610 Δ f, f₇=730 Δ f, f₈=850 Δ f, f₉=970 Δ f, f₁₀=1030 Δs f,f₁₁=1090 Δ f, f₁₂=1210 Δ f (frequency resolution Δ f=f_max/MN).{ the d of all subband signals_i[n] } it is QPSK symbol, pulse shaping function g_iT root raised cosine function that () uses rolloff-factor to be 0.1, the fast umber of beats L of consumption_s= 1000, relatively prime integers to being taken as M=37, N=33.Target false-alarm probability P in energy-probe_faIt is set to 0.01.H (z) and G Z () uses the design of Remez algorithm, transmission curve is as shown in Figure 4.

The signal finally observed is the result that formula (8) substitutes into formula (5), noise is set to 0 simultaneously, noise immunity in this example Can will describe in experiment later, the power spectrum that this method is estimated is as shown in Figure 5.

P (i) from Fig. 5 is it can be seen that through this method, can be accurately positioned the carrier wave position of all 12 subband signals Put.Only by the output of energy-probe in figureWithOwing to pseudo-peak is too much, all cannot directly tell subband The true distribution situation of signal, it is necessary to get up to react by both reasonable combination and compose occupied information really.

And for not by the original coprime spectrum output of energy-probeWithAs shown in Figure 6, although dimly The position of visible carrier wave main peak, but the pseudo-peak that existence is a lot of and main peak amplitude is more or less the same so that it is not suitable for practical situation.

This method indicated above can be greatly lowered sampling rate, and (in this experiment, two-way sampling rate is only f_s1= f_max/ M ≈ 40.54MHz, f_s2=f_max/ N ≈ 45.45MHz), the spectrum distribution feelings in whole broadband can be shown real surface again Condition.If needing to improve the frequency resolution of analyzer, bigger relatively prime integers pair can be used instead while increasing fast umber of beats.

It addition, the distribution from the testing result of Fig. 5 can be seen that, the spectrum simply entering signal can separate, it is possible to detection The position of all spectral peaks.Thus the method for the present invention is a kind of truly " blind " detection method.

(2) add make an uproar in the case of robustness experiment

Experiment 2

(sub band number only changes into m=6, and carrier frequency changes f into experiment 1 similar signal in selection₁=100 Δ f, f₂= 350Δf,f₃=600 Δ f, f₄=610 Δ f, f₅=715 Δ f, f₆=1000 Δ f), use zero mean Gaussian white noise structure The excitation of different signal to noise ratios (SNR), fast umber of beats L_s=300.This experiment is the identification under 0dB and-50dB by being given at signal to noise ratio Situation, as shown in Figure 7 and Figure 8.

By Fig. 7 with Fig. 8 it can be seen that f all can be accurately positioned under two kinds of different signal to noise ratios₁=100 Δ f, f₂=350 Δ f, f₃=600 Δ f, f₄=610 Δ f, f₅=715 Δ f, f₆The main peak of=1000 Δ f and the other peak of minority.

In the case of signal to noise ratio is for-50dB, still can successfully react real spectrum distribution situation, and successfully suppresses Live all of noise contribution, this effect being averaged mainly due to algorithm snap a number of to signal, although snap Number increases and also brings along the prolongation of detecting period, but enhance coprime spectralyzer to robustness of making an uproar.For radio communication Present in channel for much noise, the preferable noiseproof feature of the present invention makes it to be better adapted to broader frequency spectrum perception In.

In sum, this method only need to give the false-alarm detection probability of each subband center frequency, just can calculate the energy of correspondence Amount detection detection threshold value, it is thus achieved that high-precision, completely inhibit the mid frequency testing result of the broadband signal of pseudo-effect.Therefore it is our The occasion that method relates to lack sampling wireless spectrum perception at cognitive radio, industry Internet of Things etc. has broader application prospect.

Embodiment 5

Below in conjunction with concrete hardware, the method in embodiment 1-3 is designed corresponding hardware unit, as detailed below retouch State:

A kind of device of broader frequency spectrum perception based on comprehensive coprime analysis, this sensing device includes:

Microcontroller, for processing input signal, obtains former coprime spectrum output；To the letter after twice down-sampling Number process, obtain complementary coprime spectrum output；Utilize by energy detector former coprime spectrum output energy detection results, The energy detection results of spectrum output coprime with complementation tries to achieve result of calculation and the calculating of the second indicator of the first indicator respectively Result；Indicative power spectrum is got by the result of calculation of the first indicator and the result of calculation of the second indicator；

Output drives and display circuit, is used for showing indicative power spectrum.

Wherein, microcontroller includes:

Former coprime spectralyzer, for processing input signal, obtains former coprime spectrum output；

Complementary coprime spectralyzer, for processing the signal after twice down-sampling, obtains complementary coprime spectrum defeated Go out.

In sum, this device only need to give the false-alarm detection probability of each subband center frequency, just can calculate the energy of correspondence Amount detection detection threshold value, it is thus achieved that high-precision, completely inhibit the mid frequency testing result of the broadband signal of pseudo-effect.Therefore this dress Put and relate to the occasion of lack sampling wireless spectrum perception at cognitive radio, industry Internet of Things etc. and have broader application prospect.

Embodiment 6

Below in conjunction with Fig. 9 and Figure 10, the sensing device in embodiment 5 is described further, described below:

In fig .9, first by actual observation signal, filter coefficient, relatively prime integers to M, N and snap quantity L_sOutside being stored in In portion RAM, then they are input in DSP in real time, through DSP internal core algorithm, signal are carried out down-sampling, heterogeneous filter The judgement of ripple, IDFT, energy-probe, indicator calculating and frequency spectrum reconfiguration etc., finally drive display and display mould thereof by output Block demonstrates that on monoblock frequency band, spectrum takies situation.

Wherein, the DSP (Digital Signal Processor, digital signal processor) of Fig. 9 is core devices, at frequency During spectrum perception, complete following major function:

1) call internal core algorithm, complete the down-sampling of actual acquisition signal, multiphase filtering, IDFT, energy-probe The processes such as judgement, indicator calculating and frequency spectrum reconfiguration；

2) control down-sampled values and sample of signal, in real time it is adjusted so that it is correspond to actual needs；

3) spectrum sensing results is exported in real time to driving and display module.

It may be noted that owing to have employed digitized method of estimation, thus determine the complexity of Fig. 9 system, correctness and steady The peripheral connection of DSP device during principal element is not Fig. 9 qualitatively, but the core that DSP internal program memory is stored Algorithm.

The internal processes flow process of DSP device is as shown in Figure 10.

The core of " coprime broader frequency spectrum cognitive method based on bank of filters and the device thereof " that proposed is calculated by the present invention Method is implanted in DSP device, completes high accuracy, low sampling rate, efficient wideband spectrum perception based on this.

Figure 10 flow process is divided into following several step:

1) the most according to actual needs, the down-sampled values (M and N, coprime integer) of signal, and the filter of design are set Ripple device coefficient, for multiphase filtering process, gives snap quantity L simultaneously_s；

2) then, CPU main controller reads, from I/O port, the parameter that ginseng sets, and enters internal RAM；

3) present invention carries out the design of frequency spectrum perception by the processing procedure of Fig. 1 is the most crucial part of DSP algorithm, and running should After algorithm, i.e. can get institute's observation simulation takies situation；

4) judging whether this method meets actual demand, if being unsatisfactory for, program returns, and sets signal ginseng the most as requested Number；

5) until design result meets actual requirement, then exported by the output bus of DSP and show that driving sets to outside Standby, frequency spectrum perception result is carried out digital-scroll technique.

It may be noted that realize owing to have employed DSP so that the design of whole frequency spectrum perception device becomes the most flexibly and fast, can root According to being actually needed in frequency spectrum perception device design process, flexible transformation desired parameters, it is allowed to finally meet requirement of engineering.

The embodiment of the present invention is to the model of each device in addition to doing specified otherwise, and the model of other devices does not limits, As long as the device of above-mentioned functions can be completed.

It will be appreciated by those skilled in the art that accompanying drawing is the schematic diagram of a preferred embodiment, the invention described above embodiment Sequence number, just to describing, does not represent the quality of embodiment.

The foregoing is only presently preferred embodiments of the present invention, not in order to limit the present invention, all spirit in the present invention and Within principle, any modification, equivalent substitution and improvement etc. made, should be included within the scope of the present invention.

Claims (6)

1. a broader frequency spectrum cognitive method based on comprehensive coprime analysis, it is characterised in that described cognitive method includes following Step:

Input signal is processed, obtains former coprime spectrum output；

Signal after twice down-sampling is processed, obtains complementary coprime spectrum output；

Utilize the former coprime energy detection results composing output by energy detector and the energy measuring of complementation coprime spectrum output Result tries to achieve result of calculation and the result of calculation of the second indicator of the first indicator respectively；

Indicative power spectrum is got by the result of calculation of the first indicator and the result of calculation of the second indicator.

A kind of broader frequency spectrum cognitive method based on comprehensive coprime analysis the most according to claim 1, it is characterised in that institute State and input signal processed, obtain former coprime spectrum output step particularly as follows:

Input signal is carried out two-way down-sampling, obtains two-way sparse signal；

Two-way sparse signal is carried out respectively multiphase filtering, exports two multichannel sparse signals, by two sparse letters of multichannel Number can respectively obtain the multichannel output sequence of correspondence；

In each moment, 2 road multichannel output sequences are carried out IDFT respectively, respectively obtains output sequence；

To output sequence cross-correlation, obtain former coprime spectrum output.

A kind of broader frequency spectrum cognitive method based on comprehensive coprime analysis the most according to claim 1, it is characterised in that institute State and the signal after twice down-sampling processed, obtain the step of complementary coprime spectrum output particularly as follows:

To the signal Zhong Mei road signal after twice down-sampling respectively plus phase shift and carry out IDFT and respectively obtain output sequence Row；

Ask for the cross-correlation of output sequence, obtain complementary coprime spectrum output.

A kind of broader frequency spectrum cognitive method based on comprehensive coprime analysis the most according to claim 1, it is characterised in that institute State and utilize the energy detection results of the former coprime spectrum output by energy detector and the energy measuring of complementation coprime spectrum output to tie Fruit try to achieve respectively the result of calculation of the first indicator and the result of calculation of the second indicator step particularly as follows:

Given false-alarm probability, sets threshold value, if frequency spectrum output is more than threshold value, energy detector is judged at spectral position i Occupied；

If frequency spectrum output is less than or equal to threshold value, energy detector is unoccupied by being judged at spectral position i.

5. the device of a broader frequency spectrum perception based on comprehensive coprime analysis, it is characterised in that described sensing device includes:

Microcontroller, for processing input signal, obtains former coprime spectrum output；Signal after twice down-sampling is entered Row processes, and obtains complementary coprime spectrum output；Utilize the energy detection results and mutually of former coprime spectrum output by energy detector The energy detection results mending coprime spectrum output tries to achieve result of calculation and the result of calculation of the second indicator of the first indicator respectively； Indicative power spectrum is got by the result of calculation of the first indicator and the result of calculation of the second indicator；

Output drives and display circuit, is used for showing indicative power spectrum.

The device of a kind of broader frequency spectrum perception based on comprehensive coprime analysis the most according to claim 5, it is characterised in that Described microcontroller includes:

Former coprime spectralyzer, for processing input signal, obtains former coprime spectrum output；

Complementary coprime spectralyzer, for processing the signal after twice down-sampling, obtains complementary coprime spectrum output.