CN103873111A - Narrow-band interference suppression system and method adopting compressed sensing technology and used for pulse ultra wideband receiver - Google Patents

Abstract

The invention discloses a narrow-band interference suppression system and method adopting the compressed sensing technology and used for a pulse ultra wideband receiver. The method includes the six steps of out-band noise filtration, pilot frequency part narrow-band interference estimation, pilot frequency part narrow-band interference suppression, signal correlation template reconstruction, load part narrow-band interference estimation and suppression and related demodulation. The method is characterized in that narrow-band interference of pilot symbol waveforms is estimated one by one by utilizing the compressed sensing technology, a corresponding narrow-band interference observation sequence is subtracted from a pilot frequency observation sequence through a subtractor, and the estimation accuracy of a signal correlation template is improved; narrow-band interference of load symbol waveforms is estimated one by one by utilizing the compressed sensing technology, the narrow-band interference is eliminated from load signal waveforms through the subtractor, and the signal to interference ratio of load signals is improved. According to the system and method, the narrow-band interference estimation and suppression are achieved by utilizing the compressed sensing technology, the sampling rate is greatly reduced, good narrow-band interference suppression effects are achieved, and the performance of the pulse ultra wideband receiver is improved.

Description

The Suppression of narrow band interference system and method for the pulse ultra wideband receiver of compressed sensing

Technical field

The Suppression of narrow band interference system and method that the present invention relates to a kind of pulse ultra wideband receiver of compressed sensing, belongs to broadband wireless communications field.

Background technology

Nyquist sampling theorem is pointed out: from discrete signal, recover original signal, its sampling rate is at least 2 times of signal bandwidth undistortedly.Compressed sensing technology (CS:Compressed Sensing) is that a kind of emerging signal is processed framework.Be different from traditional nyquist sampling theorem, compressive sensing theory is while compressed signal under the sampling rate far below nyquist sampling theorem.This theory is pointed out: for any compressible or sparse signal, by an observing matrix, signal is projected on a lower dimensional space, then by a series of restructing algorithm reconstruct or approach out primary signal.Corresponding restructing algorithm comprises base tracing algorithm (BP:Basic Pursuit Algorithm), matching pursuit algorithm (MP:Matching Pursuit Algorithm), orthogonal matching pursuit algorithm (OMP:Orthogonal Matching Pursuit Algorithm), subspace tracing algorithm (SP:Subspace Pursuit Algorithm) etc.

Impulse radio ultra wide band system (IR UWB) is the emerging short-distance wireless communication technology of a kind of two-forty, low cost, low-power consumption, good confidentiality, can be used for the near radio data network such as Wireless Personal Network (WPAN:Wireless Personal Area Network), wireless body area network (WBAN:Wireless Body Area Network), also can be used for the system such as radar range finding, radar imagery.Different from traditional wireless communication technology, pulse ultra-broad band adopts spectrum overlapping mode to remove to share the frequency spectrum resource using at present, and it is simultaneous with existing narrow-band wireless systems.Although FCC (FCC) has limited radio ultra wide band system radiant power to ensure the normal operation of existing narrow band width wireless communication systems, but, because other narrow band width wireless communication systems are very high with respect to radio ultra wide band system radiant power, its inevitable paired pulses radio ultra wide band system produces and disturbs.These disturb and in ultra broadband bandwidth, show as the form that multiple arrowbands disturb.For ensureing the reliable communication of radio ultra wide band system, need to suppress the interference of other communication systems.On the one hand, classical Narrow Band Interference Suppression Technique is the trap interference mitigation technology based on nyquist sampling theorem.This method is converted at frequency domain and is realized estimation and the inhibition disturbed by FFT, then changes back to time domain by IFFT and then completes the process of inhibition of whole interference.But the time-frequency conversion based on fft algorithm need to be sampled with Nyquist rate to Radio Frequency Interfere, due to high bandwidth (Ghz) characteristic of impulse radio ultra wide band system, therefore need high sample rate simultaneously.This has increased difficulty and cost that system hardware is realized undoubtedly.On the other hand, in recent years, only there is one section of research compressive sensing theory to be applied to article (Xu Zhan, Zhao Yifei, Zhou Chunhui, Wang Jing, the peace Jianping of impulse radio ultra wide band system arrowband Interference Estimation both at home and abroad." the impulse radio ultra wide band system arrowband interference estimation algorithm based on compressed sensing ", Chinese journal of scientific instrument, the 32nd the 3rd phase of volume of March in 2011) occur, article emphasis is to disturb the improvement of OMP algorithm for arrowband, only simply introduce the estimation procedure that arrowband disturbs, but the whole flow process of arrowband Interference Estimation and inhibition has not been carried out the introduction of system.

Summary of the invention

Object of the present invention is exactly in order to address the above problem, a kind of Suppression of narrow band interference system and method for pulse ultra wideband receiver of compressed sensing is provided, the method is estimated by compressed sensing technology and is suppressed arrowband and disturb, reduce the sample rate of receiver, and there is good Suppression of narrow band interference effect.

To achieve these goals, the method comprises the following steps:

The Suppression of narrow band interference system of the pulse ultra wideband receiver of compressed sensing, comprises

Broadband filter module, utilizes broadband filter, the out-of-band noise in filtering impulse ultra-wideband signal;

Compressed sensing arrowband interference estimation block, with compressive sensing theory, in the pilot portion from impulse ultra-wideband signal, obtains arrowband and disturbs template;

Compressed sensing ultra broadband correlate template estimation module, utilizes the arrowband that compressed sensing arrowband interference estimation block produces to disturb template, eliminates the impact that in pilot tone, arrowband disturbs, and then utilizes compressive sensing theory, obtains ultra-broadband signal correlate template;

Signal delay module: load signal is suitably postponed, and be sent to load Suppression of narrow band interference module;

Load observation memory module: the temporary load observation sequence being obtained by observation module, and the observation sequence of each load signal is sent to arrowband Interference Estimation template successively;

Load Suppression of narrow band interference module: utilize subtracter to deduct the arrowband comprising in this load signal of being estimated by arrowband Interference Estimation template from each load signal waveform and disturb, and the load signal suppressing after arrowband disturbs is sent to correlation demodulation module;

Correlation demodulation module: the signal templates producing taking signal correction template generation module, as correlate template, utilizes the load signal after correlator and decision device disturb the inhibition arrowband of load Suppression of narrow band interference module output to carry out correlation demodulation.

Described broadband filter module comprises:

Ultra-wideband antenna module: receive the impulse ultra-wideband signal from wireless channel, and this signal is sent into receiving filter module;

Receiving filter module: paired pulses ultra-broadband signal filtering out-of-band noise and interference.

Described compressed sensing arrowband interference estimation block comprises:

Observation module: frequency pilot sign waveform and load symbol waveform are observed respectively, and pilot tone observation sequence is sent to pilot tone observation memory module, load observation sequence is sent to load observation memory module;

Pilot tone observation memory module: the observation sequence of temporary each frequency pilot sign waveform, and be sent to arrowband interference estimation block and subtracter block;

Arrowband interference estimation block: utilize OMP algorithm to estimate that arrowband disturbs, the arrowband containing in the frequency pilot sign waveform of estimation is disturbed and is sent to pilot tone arrowband disturbance-observer module, the arrowband containing in the load signal waveform of estimation is disturbed and is sent to load Suppression of narrow band interference module.

Described compressed sensing ultra broadband correlate template estimation module comprises:

Pilot tone arrowband disturbance-observer module: the frequency pilot sign arrowband that arrowband interference estimation block is estimated disturbs and observes respectively, and the arrowband disturbance-observer sequence that observation is obtained is sent to arrowband disturbance-observer memory module;

Arrowband disturbance-observer memory module: temporary arrowband disturbance-observer sequence, and be sent to subtracter block;

Subtracter block: utilize subtracter, deduct its corresponding arrowband disturbance-observer sequence from each pilot tone observation sequence, obtain and suppress the pilot tone observation sequence that arrowband disturbs;

The average module of vector: the pilot tone observation sequence that the inhibition arrowband obtaining in subtracter block is disturbed is sued for peace and is averaged operation acquisition average pilot observation sequence, to reduce the impact of additive white Gaussian noise, and is sent to signal correction template generation module;

Signal correction template generation module: the average pilot observation sequence obtaining according to the average module of vector, utilizes OMP algorithm reconstruction signal correlate template, and is sent to correlation demodulation module.

The narrow-band interference rejection method of the pulse ultra wideband receiver based on compressed sensing, comprises the steps:

Step (1): receiving-transmitting sides is set up after communication, receiving terminal receives the impulse ultra-wideband signal from wireless channel, described impulse ultra-wideband signal comprises two parts: pilot portion and data payload part, and filtering is superimposed upon the out-of-band noise on impulse ultra-wideband signal;

Step (2): receiving terminal is based on compressive sensing theory, utilize the gaussian random matrix generating in advance as observing matrix, each frequency pilot sign waveform in impulse ultra-wideband signal pilot portion in step (1) is observed, obtain some pilot tone observation sequences, afterwards the arrowband containing in each frequency pilot sign waveform is disturbed and estimated;

Step (3):

With the observing matrix in step (2), the arrowband interference of estimating in step (2) is observed respectively, obtained some arrowbands disturbance-observer sequence;

The some pilot tone observation sequences that obtain from step (2) deduct corresponding arrowband disturbance-observer sequence, obtain and remove some pilot tone observation sequences that arrowband disturbs;

Step (4): some pilot tone observation sequences that the removal arrowband obtaining in step (3) is disturbed are averaged operation and obtain average pilot observation sequence, to reduce the impact of additive white Gaussian noise, and according to the average pilot observation sequence obtaining, reconstruct signal correction template; Described signal correction template refers to: for the local reference signal of correlator in correlation receiver;

Step (5): utilize the observing matrix in step (2), each data symbol waveform in the data payload part of impulse ultra-wideband signal in step (1) is observed, obtain data payload observation sequence, afterwards the arrowband containing in each data symbol waveform is disturbed and estimated, obtain arrowband interference waveform, then utilize subtracter, from load signal waveform, cut arrowband interference waveform, eliminate the impact that arrowband disturbs;

Step (6): the load signal waveform after the elimination arrowband obtaining in step (5) is disturbed, utilize the signal correction template producing in step (4), carry out correlation demodulation by correlation receiver and obtain data symbol sequence.

The concrete steps of described step (1) are:

Suppose that the ultra-broadband signal waveform that receiving terminal obtains is:

$\begin{array}{c}s\left(t\right)=\underset{i=1}{\overset{N_p}{\mathrm{\Σ}}}g(t-{\mathrm{iT}}_s)+\underset{j=1}{\overset{N_s}{\mathrm{\Σ}}}{b}_{j}g(t-N_p{T}_s-j{T}_s)+n\left(t\right)+f_{\mathrm{nb}}\left(t\right)\\ =\underset{i=1}{\overset{N_p}{\mathrm{\Σ}}}{\stackrel{\·}{g}}_i\left(t\right)+\underset{j=1}{\overset{N_s}{\mathrm{\Σ}}}{\stackrel{\·}{g}}_j\left(t\right)\end{array}$

Wherein, N _prepresent frequency pilot sign number, all frequency pilot signs are 1, N _srepresent load information symbol numbers, T _sfor symbol period, b _jfor binary data modulation symbol and b _j∈ { 1,1}, f _nb(t) and n (t) represent respectively arrowband disturb and white Gaussian noise,

t ∈ ((i-1) T _s, iT _s), i=1,2 ..., N _prepresent the frequency pilot sign waveform that i band disturbs,

t ∈ ((j-1) T _s+ N _pt _s, jT+N _pt _s), j=1,2 ..., N _srepresent the load signal waveform that j band disturbs.

The concrete steps of described step (2) are:

Receiving terminal, based on compressive sensing theory, utilizes the gaussian random observing matrix generating in advance

each frequency pilot sign waveform in impulse ultra-wideband signal pilot portion in step (1) is observed, obtained N _pthe observation sequence of individual frequency pilot sign waveform

wherein y _i, i=1,2 ..., N _pfor M × 1 dimensional vector, represent the observation sequence of i frequency pilot sign waveform, and its k element is:

Wherein represent i the pilot signal waveform being disturbed, for k observation waveform in observing matrix, N _pfor pilot tone number, M is observation waveform number;

When the arrowband interference containing in each frequency pilot sign waveform is reconstructed, the sparse dictionary that adopts inverse discrete Fourier transform (IDFT) matrix to disturb as arrowband:

${\mathrm{\&psi;}}_{\mathrm{nb}}=[{\mathrm{\&psi;}}_1\left(t\right),{\mathrm{\&psi;}}_2\left(t\right),\&CenterDot;\&CenterDot;\&CenterDot;,{\mathrm{\&psi;}}_N\left(t\right)]=\frac{1}{\sqrt{N}}{\left[\begin{array}{ccccc}{W}^0& W^0& {\mathrm{<figure-callout\; id="0"\; label="W"\; filenames="HDA0000481925690000012.png"\; state="\{\{state\}\}">W</figure-callout>}}^0& \&CenterDot;\&CenterDot;\&CenterDot;& W^0\\ W^0& W^1& {\mathrm{<figure-callout\; id="2"\; label="W"\; filenames="HDA0000481925690000011.png,HDA0000481925690000012.png"\; state="\{\{state\}\}">W</figure-callout>}}^2& \&CenterDot;\&CenterDot;\&CenterDot;& W^{N-1}\\ W^0& W^2& {\mathrm{<figure-callout\; id="4"\; label="W"\; filenames="HDA0000481925690000011.png,HDA0000481925690000012.png"\; state="\{\{state\}\}">W</figure-callout>}}^4& \&CenterDot;\&CenterDot;\&CenterDot;& W^{2(N-1)}\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ W^0& W^{N-1}& W^{2(N-1)}& \&CenterDot;\&CenterDot;\&CenterDot;& W^{(N-1)(N-1)}\end{array}\right]}^{*}$

Wherein,

n is the sampling number of each frequency pilot sign waveform while carrying out nyquist sampling, the computing of * representing matrix conjugate transpose.Because arrowband disturbs as frequency-domain sparse, so the arrowband in each frequency pilot sign cycle disturbs

t ∈ ((i-1) T _s, iT _s), i=1,2 ..., N _pbe expressed as:

$f_{\mathrm{nb}}^i\left(t\right)=\underset{k=1}{\overset{N}{\mathrm{\&Sigma;}}}{\mathrm{\&psi;}}_k\left(t\right){\mathrm{\&theta;}}_k={\mathrm{\&psi;}}_{\mathrm{nb}}{\mathrm{\&theta;}}_{\mathrm{nb}}^i,$ t∈((i-1)T _s,iT _s),i＝1,2,...,N _p。

Wherein, Ψ _nbfor arrowband disturbs sparse dictionary,

be the projection coefficient vector that in i frequency pilot sign waveform, arrowband disturbs, and because arrowband interference is sparse, in only have the active element of minority.

Disturb sparse dictionary Ψ by gaussian random observing matrix Φ and arrowband _nbcan obtain restructuring matrix:

V _nb＝ΦΨ _nb

Wherein, Φ is observing matrix, Ψ _nbfor disturbing sparse dictionary in arrowband.

According to restructuring matrix V _nbwith the N obtaining _pindividual observation sequence

then estimating just to estimate by OMP algorithm the arrowband containing in each frequency pilot sign waveform disturbs

wherein

t ∈ ((i-1) T _s, iT _s), i=1,2 ..., N _prepresent the estimation of the arrowband interference containing in i frequency pilot sign waveform.

The step of described OMP algorithm is: suppose that OMP algorithm maximum iteration time is defined as K;

The first step: initialization: residual values r ₀=y, indexed set

(empty set), Increment Matrix

iterations t=1;

Second step: find out residual values r _t-1with V _nbin row V _jthe corresponding row of inner product maximum

The 3rd step: upgrade indexed set

upgrade Increment Matrix

The 4th step: obtained by least square method

The 5th step: upgrade residual values

The 6th step: judge whether to meet t >=K, if meet, stop iteration; If do not meet, carry out second step.

Residual values y in OMP algorithm is used respectively to each pilot tone measured value y _i, i=1,2 ..., N _preplace, just obtain the projection coefficient vector that arrowband that each frequency pilot sign waveform contains disturbs

i=1,2 ..., N _pestimation

i=1,2 ..., N _p, therefore, the arrowband of estimation disturbs and is:

${\hat{f}}_{\mathrm{nb}}^i\left(t\right)={\mathrm{\&psi;}}_{\mathrm{nb}}{\widehat{\mathrm{\&theta;}}}_{\mathrm{nb}}^i,$ t∈((i-1)T _s,iT _s),i＝1,2,...,N _p。

Wherein, Ψ _nbfor arrowband disturbs sparse dictionary, be that the arrowband containing in i frequency pilot sign waveform disturbs projection coefficient vector

i=1,2 ..., N _pestimation.

The concrete steps of described step (3) are:

With the gaussian random observing matrix in step (2)

the arrowband of estimating in step (2) is disturbed

observe respectively, obtain corresponding arrowband disturbance-observer sequence wherein

i=1,2 ..., N _pfor M × 1 dimensional vector, represent the estimation that the arrowband to containing in i frequency pilot sign waveform disturbs

observes and obtain observation sequence, and its k element is:

Wherein

represent the estimation of the arrowband interference containing in i frequency pilot sign waveform, for k observation waveform in observing matrix, N _pfor pilot tone number, M is observation waveform number;

The some pilot tone observation sequences that obtain from step (2)

in deduct corresponding arrowband disturbance-observer sequence

impact, obtain remove arrowband disturb pilot tone observation sequence;

$\begin{array}{c}[y_{1-f},y_{2-f},\&CenterDot;\&CenterDot;\&CenterDot;,y_{N_p-f}]=[{\mathrm{<figure-callout\; id="1"\; label="y"\; filenames="HDA0000481925690000011.png"\; state="\{\{state\}\}">y</figure-callout>}}_1,{\mathrm{<figure-callout\; id="2"\; label="y"\; filenames="HDA0000481925690000011.png,HDA0000481925690000012.png"\; state="\{\{state\}\}">y</figure-callout>}}_2,\&CenterDot;\&CenterDot;\&CenterDot;y_{N_p}]-[{\mathrm{<figure-callout\; id="1"\; label="y\; f"\; filenames="HDA0000481925690000011.png"\; state="\{\{state\}\}">y</figure-callout>}}_f^{},{\mathrm{<figure-callout\; id="2"\; label="y\; f"\; filenames="HDA0000481925690000011.png,HDA0000481925690000012.png"\; state="\{\{state\}\}">y</figure-callout>}}_f^{},\&CenterDot;\&CenterDot;\&CenterDot;,y_f^{N_p}]\\ =[y_1-{\mathrm{<figure-callout\; id="1"\; label="y\; f"\; filenames="HDA0000481925690000011.png"\; state="\{\{state\}\}">y</figure-callout>}}_f^{},y_2-{\mathrm{<figure-callout\; id="2"\; label="y\; f"\; filenames="HDA0000481925690000011.png,HDA0000481925690000012.png"\; state="\{\{state\}\}">y</figure-callout>}}_f^{},\&CenterDot;\&CenterDot;\&CenterDot;,y_{N_p}-y_f^{N_p}]\end{array}$

Wherein,

represent that i frequency pilot sign waveform suppresses the observation sequence after arrowband disturbs.

The concrete steps of described step (4) are:

Pilot tone observation sequence after the some inhibition arrowband obtaining in step (3) is disturbed

be averaged operation and obtain average pilot observation sequence

to reduce the impact of additive white Gaussian noise,

$\stackrel{\&OverBar;}{y}=\frac{1}{N_p}\underset{i=1}{\overset{N_p}{\mathrm{\&Sigma;}}}{y}_{i-f}=\frac{1}{N_p}\underset{i=1}{\overset{N_p}{\mathrm{\&Sigma;}}}(y_i-y_f^i)$

Wherein, y _i-f, i=1,2 ..., N _prepresent that i pilot waveform suppresses the observation sequence after arrowband disturbs.

For the reconstruct of signal correction template, the sparse dictionary of employing is the sparse dictionary Ψ of characteristic vector _g, its production process is as follows:

Due to the time-varying characteristics of pulse ultra-broad band channel, the impulse ultra-wideband signal g (t) receiving in step (1) is a random process in essence.Therefore, obtain its covariance function R (t-τ) by a large amount of channel samples,

R(t-τ)＝E[g(t)g(τ+t)]

Suppose λ ₁> λ ₂> λ ₃> ... > λ _nrepresent the characteristic value of Fredholm integral operator, wherein, N is the sampling number of each frequency pilot sign waveform while carrying out nyquist sampling.Have for R (t-τ):

$\&Integral;R(t-\mathrm{\&tau;})u_j\left(\mathrm{\&tau;}\right)\mathrm{d\&tau;}={\mathrm{\&lambda;}}_j{u}_j\left(t\right)$

Wherein, u _j(t) be λ _jcharacteristic of correspondence vector, and { u _j(t) } be the complete set of one group of orthogonal basis function, meet:

$\&Integral;u_i\left(t\right)u_j\left(t\right)=\begin{array}{c}\left\{\begin{array}{cc}0& i\&NotEqual;j\\ 1& i=j\end{array}\right.\end{array},i,j=\mathrm{1,2,3},\&CenterDot;\&CenterDot;\&CenterDot;,N$

Therefore, [u ₁(t), u ₂(t), u ₃(t) ..., u _n(t)] formed the orthogonal basis of one group of g (t), this group orthogonal basis is the sparse dictionary of characteristic vector.

By step (2) gaussian random observing matrix

with the sparse dictionary Ψ of characteristic vector _g, can obtain correlate template restructuring matrix,

V _g＝ΦΨ _g

In conjunction with the average pilot observation sequence obtaining in this step

just can utilize the reconstruct of OMP algorithm to obtain signal correction template

oMP algorithmic procedure is described in step (2).

The concrete steps of described step (5) are:

First, utilize the gaussian random observing matrix in step (2)

each load symbol waveform in impulse ultra-wideband signal loading section in step (1) is observed, obtained the observation sequence of each load symbol waveform

wherein

for M × 1 dimensional vector, represent the observation sequence of j load symbol waveform, and its k element is:

Wherein

represent j the load signal waveform being disturbed,

for k observation waveform in observing matrix, N _sfor load symbol numbers, M is observation waveform number;

Then, according to the method for estimation of in step (2), the arrowband that contains in pilot signal being disturbed, the estimation that the arrowband that utilizes OMP algorithm to obtain to contain in each load symbol waveform disturbs

represent the estimation of the arrowband interference containing in j load symbol waveform, and

be expressed as:

${\hat{f}}_{\mathrm{nb}}^j\left(t\right)={\mathrm{\&psi;}}_{\mathrm{nb}}{\widehat{\mathrm{\&theta;}}}_{\mathrm{nb}}^j,$ t∈((j-1)T _s+N _pT _s,jT _s+N _pT _s),j＝1,2,...,N _s。

Wherein, Ψ _nbfor arrowband disturbs sparse dictionary, for utilizing the projection coefficient vector of arrowband interference in j the load symbol waveform that OMP algorithm obtains

j=1,2 ..., N _sestimation.

Finally, utilize subtracter, from load signal waveform, eliminate the impact that arrowband disturbs, obtain and suppress the load signal waveform s that arrowband disturbs _load(t):

$s_{\mathrm{load}}\left(t\right)=\underset{j=1}{\overset{N_s}{\mathrm{\&Sigma;}}}[{\stackrel{\&CenterDot;}{g}}_j\left(t\right)-{\hat{f}}_{\mathrm{nb}}^j\left(t\right)],$ t∈((j-1)T _s+N _pT _s,jT _s+N _pT _s),j＝1,2,...,N _s

The concrete steps of described step (6) are:

Load signal waveform s after the elimination arrowband obtaining in step (5) is disturbed _load(t), utilize the signal correction template producing in step (4)

carry out correlation demodulation by correlation receiver and obtain symbol sebolic addressing

j=1,2 ..., N _smeet

Wherein,

represent the load signal waveform that j band disturbs, represent the estimation of the arrowband interference containing in j load symbol waveform,

for correlate template, N _srepresent load information symbol numbers.

Beneficial effect of the present invention is:

The inventive method, realizes estimation and inhibition that arrowband is disturbed by compressed sensing technology, have following beneficial effect:

1. greatly reduce sampling rate by compressed sensing, broken through the technical bottleneck that traditional sampling theory faces, be conducive to control receiver cost and power consumption;

2. utilize compressed sensing and restructing algorithm detect adaptively random narrow-band interference and suppress, improved receiver correlate template estimated accuracy;

3. the high power arrowband being superimposed upon in ultra-broadband signal by accurate estimation inhibition disturbs, and has improved the detecting reliability of receiver.

Brief description of the drawings

Fig. 1 is the inventive method embodiment structured flowchart;

The error rate change curve that Fig. 2 produces for the inventive method embodiment;

Wherein, 1. ultra-wideband antenna module, 2. receiving filter module, 3 observation modules, 4 pilot tone observation memory module 5. arrowband interference estimation block, 6. pilot tone arrowband disturbance-observer module, 7. arrowband disturbance-observer memory module, 8. subtracter block, the 9. average module of vector, 10. signal correction template generation module, 11. signal delay modules, 12. load observation memory modules, 13. load Suppression of narrow band interference modules, 14. correlation demodulation modules.

Embodiment

Below in conjunction with accompanying drawing and embodiment, the invention will be further described.

As shown in Figure 1, a kind of inventive method embodiment structured flowchart, each module effect is as follows:

Ultra-wideband antenna module 1: receive the impulse ultra-wideband signal from wireless channel, and this signal is sent into receiving filter module 2;

Receiving filter module 2: paired pulses ultra-broadband signal filtering out-of-band noise and interference;

Observation module 3: frequency pilot sign waveform and load symbol waveform are observed respectively, and pilot tone observation sequence is sent to pilot tone observation memory module 4, load observation sequence is sent to load observation memory module 12;

Pilot tone observation memory module 4: the observation sequence of temporary each frequency pilot sign waveform, and be sent to arrowband interference estimation block 5 and subtracter block 8;

Arrowband interference estimation block 5: utilize OMP algorithm to estimate that arrowband disturbs, the arrowband containing in the frequency pilot sign waveform of estimation is disturbed and is sent to pilot tone arrowband disturbance-observer module 6, the arrowband containing in the load signal waveform of estimation is disturbed and is sent to load Suppression of narrow band interference module 13;

Pilot tone arrowband disturbance-observer module 6: the frequency pilot sign arrowband that arrowband interference estimation block 5 is estimated disturbs and observes respectively, and the arrowband disturbance-observer sequence that observation is obtained is sent to arrowband disturbance-observer memory module 7;

Arrowband disturbance-observer memory module 7: temporary arrowband disturbance-observer sequence, and be sent to subtracter block 8;

Subtracter block 8: utilize subtracter, deduct its corresponding arrowband disturbance-observer sequence from each pilot tone observation sequence, obtain and suppress the pilot tone observation sequence that arrowband disturbs;

The average module 9 of vector: the pilot tone observation sequence that the inhibition arrowband obtaining in subtracter block 8 is disturbed is sued for peace to be averaged to operate and obtained average pilot observation sequence, to reduce the impact of additive white Gaussian noise, and be sent to signal correction template generation module 10;

Signal correction template generation module 10: the average pilot observation sequence obtaining according to the average module 9 of vector, utilizes OMP algorithm reconstruction signal correlate template, and is sent to correlation demodulation module 14;

Signal delay module 11: load signal is suitably postponed, and be sent to load Suppression of narrow band interference module 13;

Load observation memory module 12: the temporary load observation sequence being obtained by observation module 3, and the observation sequence of each load signal is sent to arrowband Interference Estimation template 5 successively;

Load Suppression of narrow band interference module 13: utilize subtracter to deduct the arrowband comprising in this load signal of being estimated by arrowband Interference Estimation template 5 from each load signal waveform and disturb, and the load signal suppressing after arrowband disturbs is sent to correlation demodulation module 14;

Correlation demodulation module 14: the signal templates producing taking signal correction template generation module 10 is as correlate template, the load signal after utilizing inhibition arrowband that correlator and decision device are exported load Suppression of narrow band interference module 13 to disturb carries out correlation demodulation.

The simulation parameter of this inventive method embodiment: simulated environment: Matlab7.13; Transmitter basic pulse waveform: Gauss's second dervative impulse waveform; Channel model: IEEE802.15.3a CM1 channel model; UWB modulation system is BPSK; Observing matrix: gaussian random matrix; Arrowband disturbs sparse dictionary: inverse discrete Fourier transform matrix; UWB receives the sparse dictionary of signal: the sparse dictionary of characteristic vector model; Impulse ultra-wideband signal power :-30dBm; Arrowband interference power is :-10dBm; CS restructing algorithm: OMP algorithm; Channel number: 100; Each packet frequency pilot sign number 10; Total load data symbol number 1000000.

The narrow-band interference rejection method of the pulse ultra wideband receiver based on compressed sensing, comprises the steps:

Step (1): receiving-transmitting sides is set up after communication, receiving terminal receives the impulse ultra-wideband signal from wireless channel, described impulse ultra-wideband signal comprises two parts: pilot portion and data payload part, and filtering is superimposed upon the out-of-band noise on impulse ultra-wideband signal;

Step (2): receiving terminal is based on compressive sensing theory, utilize the gaussian random matrix generating in advance as observing matrix, each frequency pilot sign waveform in impulse ultra-wideband signal pilot portion in step (1) is observed, obtain some pilot tone observation sequences, utilize afterwards OMP algorithm that the arrowband containing in each frequency pilot sign waveform is disturbed and estimated;

Step (3):

With the observing matrix in step (2), the arrowband interference of estimating in step (2) is observed respectively, obtained some arrowbands disturbance-observer sequence;

The some pilot tone observation sequences that obtain from step (2) deduct corresponding arrowband disturbance-observer sequence, obtain and remove some pilot tone observation sequences that arrowband disturbs;

Step (4): the some pilot tone observation sequences that obtain in step (3) are averaged to operation and obtain average pilot observation sequence, to reduce the impact of additive white Gaussian noise, and according to the average pilot observation sequence obtaining, utilize OMP algorithm to reconstruct signal correction template;

Step (5): utilize the observing matrix in step (2), each data symbol waveform in the data payload part of impulse ultra-wideband signal in step (1) is observed, obtain data payload observation sequence, utilize afterwards OMP algorithm that the arrowband containing in each data symbol waveform is disturbed and estimated, obtain arrowband interference waveform, then utilize subtracter, from load signal waveform, cut arrowband interference waveform, eliminate the impact that arrowband disturbs;

Step (6): the load signal waveform after the elimination arrowband obtaining in step (5) is disturbed, utilize the signal correction template producing in step (4), carry out correlation demodulation by correlation receiver and obtain data symbol sequence.

The concrete steps of described step (1) are:

Suppose that the ultra-broadband signal waveform that receiving terminal obtains is:

$\begin{array}{c}s\left(t\right)=\underset{i=1}{\overset{N_p}{\mathrm{\&Sigma;}}}g(t-{\mathrm{iT}}_s)+\underset{j=1}{\overset{N_s}{\mathrm{\&Sigma;}}}{b}_{j}g(t-N_p{T}_s-j{T}_s)+n\left(t\right)+f_{\mathrm{nb}}\left(t\right)\\ =\underset{i=1}{\overset{N_p}{\mathrm{\&Sigma;}}}{\stackrel{\&CenterDot;}{g}}_i\left(t\right)+\underset{j=1}{\overset{N_s}{\mathrm{\&Sigma;}}}{\stackrel{\&CenterDot;}{g}}_j\left(t\right)\end{array}$

Wherein, N _prepresent frequency pilot sign number, N _srepresent load information symbol numbers, T _sfor symbol period, b _jfor binary data modulation symbol and b _j∈ { 1, } 1, f _nb(t) and n (t) represent respectively arrowband disturb and white Gaussian noise,

t ∈ ((i-1) T _s, iT _s), i=1,2 ..., N _prepresent the frequency pilot sign waveform that i band disturbs,

t ∈ ((j-1) T _s+ N _pt _s, jT+N _pt _s), j=1,2 ..., N _srepresent the load signal waveform that j band disturbs.

The concrete steps of described step (2) are:

Receiving terminal, based on compressive sensing theory, utilizes the gaussian random observing matrix generating in advance

each frequency pilot sign waveform in impulse ultra-wideband signal pilot portion in step (1) is observed, obtained N _pthe observation sequence of individual frequency pilot sign waveform

wherein y _i, i=1,2 ..., N _pfor M × 1 dimensional vector, represent the observation sequence of i frequency pilot sign waveform, and its k element is:

Wherein

represent i the pilot signal waveform being disturbed,

for k observation waveform in observing matrix, N _pfor pilot tone number, M is observation waveform number;

When the arrowband interference containing in each frequency pilot sign waveform is reconstructed, the sparse dictionary that adopts inverse discrete Fourier transform (IDFT) matrix to disturb as arrowband:

${\mathrm{\&psi;}}_{\mathrm{nb}}=[{\mathrm{\&psi;}}_1\left(t\right),{\mathrm{\&psi;}}_2\left(t\right),\&CenterDot;\&CenterDot;\&CenterDot;,{\mathrm{\&psi;}}_N\left(t\right)]=\frac{1}{\sqrt{N}}{\left[\begin{array}{ccccc}{W}^0& W^0& {\mathrm{<figure-callout\; id="0"\; label="W"\; filenames="HDA0000481925690000012.png"\; state="\{\{state\}\}">W</figure-callout>}}^0& \&CenterDot;\&CenterDot;\&CenterDot;& W^0\\ W^0& W^1& {\mathrm{<figure-callout\; id="2"\; label="W"\; filenames="HDA0000481925690000011.png,HDA0000481925690000012.png"\; state="\{\{state\}\}">W</figure-callout>}}^2& \&CenterDot;\&CenterDot;\&CenterDot;& W^{N-1}\\ W^0& W^2& {\mathrm{<figure-callout\; id="4"\; label="W"\; filenames="HDA0000481925690000011.png,HDA0000481925690000012.png"\; state="\{\{state\}\}">W</figure-callout>}}^4& \&CenterDot;\&CenterDot;\&CenterDot;& W^{2(N-1)}\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ W^0& W^{N-1}& W^{2(N-1)}& \&CenterDot;\&CenterDot;\&CenterDot;& W^{(N-1)(N-1)}\end{array}\right]}^{*}$

Wherein,

n is the sampling number of each frequency pilot sign waveform while carrying out nyquist sampling, the computing of * representing matrix conjugate transpose.Because arrowband disturbs as frequency-domain sparse, so the arrowband in each frequency pilot sign cycle disturbs

t ∈ ((i-1) T _s, iT _s), i=1,2 ..., N _pbe expressed as:

$f_{\mathrm{nb}}^i\left(t\right)=\underset{k=1}{\overset{N}{\mathrm{\&Sigma;}}}{\mathrm{\&psi;}}_k\left(t\right){\mathrm{\&theta;}}_k={\mathrm{\&psi;}}_{\mathrm{nb}}{\mathrm{\&theta;}}_{\mathrm{nb}}^i,$ t∈((i-1)T _s,iT _s),i＝1,2,...,N _p。

Wherein, Ψ _nbfor arrowband disturbs sparse dictionary,

be the projection coefficient vector that in i frequency pilot sign waveform, arrowband disturbs, and because arrowband interference is sparse,

in only have the active element of minority.

Disturb sparse dictionary Ψ by gaussian random observing matrix Φ and arrowband _nbcan obtain restructuring matrix:

V _nb＝ΦΨ _nb

Wherein, Φ is observing matrix, Ψ _nbfor disturbing sparse dictionary in arrowband.

According to restructuring matrix V _nbwith the N obtaining _pindividual observation sequence then estimating just to estimate by OMP algorithm the arrowband containing in each frequency pilot sign waveform disturbs

wherein

t ∈ ((i-1) T _s, iT _s), i=1,2 ..., N _prepresent the estimation of the arrowband interference containing in i frequency pilot sign waveform.

The step of described OMP algorithm is: suppose that OMP algorithm maximum iteration time is defined as K;

The first step: initialization: residual values r ₀=y, indexed set

(empty set), Increment Matrix iterations t=1;

Second step: find out residual values r _t-1with V _nbin row V _jthe corresponding row of inner product maximum

The 3rd step: upgrade indexed set upgrade Increment Matrix

The 4th step: obtained by least square method

The 5th step: upgrade residual values

The 6th step: judge whether to meet t >=K, if meet, stop iteration; If do not meet, carry out second step.

Residual values y in OMP algorithm is used respectively to each pilot tone measured value y _i, i=1,2 ..., N _preplace, just obtain the projection coefficient vector that arrowband that each frequency pilot sign waveform contains disturbs

i=1,2 ..., N _pestimation

i=1,2 ..., N _p, therefore, the arrowband of estimation disturbs and is:

${\hat{f}}_{\mathrm{nb}}^j\left(t\right)={\mathrm{\&psi;}}_{\mathrm{nb}}{\widehat{\mathrm{\&theta;}}}_{\mathrm{nb}}^j,$ t∈((i-1)T _s,iT _s),i＝1,2,...,N _p。

Wherein, Ψ _nbfor arrowband disturbs sparse dictionary,

be that the arrowband containing in i frequency pilot sign waveform disturbs projection coefficient vector

i=1,2 ..., N _pestimation.

The concrete steps of described step (3) are:

With the gaussian random observing matrix in step (2) the arrowband of estimating in step (2) is disturbed

observe respectively, obtain corresponding arrowband disturbance-observer sequence

wherein

i=1,2 ..., N _pfor M × 1 dimensional vector, represent the estimation that the arrowband to containing in i frequency pilot sign waveform disturbs

observes and obtain observation sequence, and its k element is:

Wherein represent the estimation of the arrowband interference containing in i frequency pilot sign waveform, for k observation waveform in observing matrix, N _pfor pilot tone number, M is observation waveform number;

The some pilot tone observation sequences that obtain from step (2)

in deduct corresponding arrowband disturbance-observer sequence impact, obtain remove arrowband disturb pilot tone observation sequence;

$\begin{array}{c}[y_{1-f},y_{2-f},\&CenterDot;\&CenterDot;\&CenterDot;,y_{N_p-f}]=[{\mathrm{<figure-callout\; id="1"\; label="y"\; filenames="HDA0000481925690000011.png"\; state="\{\{state\}\}">y</figure-callout>}}_1,{\mathrm{<figure-callout\; id="2"\; label="y"\; filenames="HDA0000481925690000011.png,HDA0000481925690000012.png"\; state="\{\{state\}\}">y</figure-callout>}}_2,\&CenterDot;\&CenterDot;\&CenterDot;y_{N_p}]-[{\mathrm{<figure-callout\; id="1"\; label="y\; f"\; filenames="HDA0000481925690000011.png"\; state="\{\{state\}\}">y</figure-callout>}}_f^{},{\mathrm{<figure-callout\; id="2"\; label="y\; f"\; filenames="HDA0000481925690000011.png,HDA0000481925690000012.png"\; state="\{\{state\}\}">y</figure-callout>}}_f^{},\&CenterDot;\&CenterDot;\&CenterDot;,y_f^{N_p}]\\ =[y_1-{\mathrm{<figure-callout\; id="1"\; label="y\; f"\; filenames="HDA0000481925690000011.png"\; state="\{\{state\}\}">y</figure-callout>}}_f^{},y_2-{\mathrm{<figure-callout\; id="2"\; label="y\; f"\; filenames="HDA0000481925690000011.png,HDA0000481925690000012.png"\; state="\{\{state\}\}">y</figure-callout>}}_f^{},\&CenterDot;\&CenterDot;\&CenterDot;,y_{N_p}-y_f^{N_p}]\end{array}$

Wherein,

i=1,2 ..., N _prepresent that i frequency pilot sign waveform suppresses the observation sequence after arrowband disturbs.

The concrete steps of described step (4) are:

Pilot tone observation sequence after the some inhibition arrowband obtaining in step (3) is disturbed

be averaged operation and obtain average pilot observation sequence

to reduce the impact of additive white Gaussian noise,

$\stackrel{\&OverBar;}{y}=\frac{1}{N_p}\underset{i=1}{\overset{N_p}{\mathrm{\&Sigma;}}}{y}_{i-f}=\frac{1}{N_p}\underset{i=1}{\overset{N_p}{\mathrm{\&Sigma;}}}(y_i-y_f^i)$

Wherein, y _i-f, i=1,2 ..., N _prepresent that i pilot waveform suppresses the observation sequence after arrowband disturbs.

For the reconstruct of signal correction template, the sparse dictionary of employing is the sparse dictionary Ψ of characteristic vector _g, its production process is as follows:

Due to the time-varying characteristics of pulse ultra-broad band channel, the impulse ultra-wideband signal g (t) receiving in step (1) is a random process in essence.Therefore, obtain its covariance function R (t-τ) by a large amount of channel samples,

R(t-τ)＝E[g(t)g(τ+t)]

Suppose λ ₁> λ ₂> λ ₃> ... > λ _nrepresent the characteristic value of Fredholm integral operator, wherein, N is the sampling number of each frequency pilot sign waveform while carrying out nyquist sampling.Have for R (t-τ):

$\&Integral;R(t-\mathrm{\&tau;})u_j\left(\mathrm{\&tau;}\right)\mathrm{d\&tau;}={\mathrm{\&lambda;}}_j{u}_j\left(t\right)$

Wherein, u _j(t) be λ _jcharacteristic of correspondence vector, and { u _j(t) } be the complete set of one group of orthogonal basis function, meet:

$\&Integral;u_i\left(t\right)u_j\left(t\right)=\begin{array}{c}\left\{\begin{array}{cc}0& i\&NotEqual;j\\ 1& i=j\end{array}\right.\end{array},i,j=\mathrm{1,2,3},\&CenterDot;\&CenterDot;\&CenterDot;,N$

Therefore, [u ₁(t), u ₂(t), u ₃(t) ..., u _n(t)] formed the orthogonal basis of one group of g (t), this group orthogonal basis is the sparse dictionary of characteristic vector.

By step (2) gaussian random observing matrix with the sparse dictionary Ψ of characteristic vector _g, can obtain correlate template restructuring matrix,

V _g＝ΦΨ _g

In conjunction with the average pilot observation sequence obtaining in this step

just can utilize the reconstruct of OMP algorithm to obtain signal correction template oMP algorithmic procedure is described in step (2).

The concrete steps of described step (5) are:

First, utilize the gaussian random observing matrix in step (2)

each load symbol waveform in impulse ultra-wideband signal loading section in step (1) is observed, obtained the observation sequence of each load symbol waveform

wherein

j=1,2 ..., N _sfor M × 1 dimensional vector, represent the observation sequence of j load symbol waveform, and its k element is:

Wherein represent j the load signal waveform being disturbed,

for k observation waveform in observing matrix, N _sfor load symbol numbers, M is observation waveform number;

Then, according to the method for estimation of in step (2), the arrowband that contains in pilot signal being disturbed, the estimation that the arrowband that utilizes OMP algorithm to obtain to contain in each load symbol waveform disturbs

represent the estimation of the arrowband interference containing in j load symbol waveform, and

be expressed as:

${\hat{f}}_{\mathrm{nb}}^j\left(t\right)={\mathrm{\&psi;}}_{\mathrm{nb}}{\widehat{\mathrm{\&theta;}}}_{\mathrm{nb}}^j,$ t∈((j-1)T _s+N _pT _s,jT _s+N _pT _s),j＝1,2,...,N _s。

Wherein, Ψ _nbfor arrowband disturbs sparse dictionary,

for utilizing the projection coefficient vector of arrowband interference in j the load symbol waveform that OMP algorithm obtains

j=1,2 ..., N _sestimation.

Finally, utilize subtracter, from load signal waveform, eliminate the impact that arrowband disturbs, obtain and suppress the load signal waveform s that arrowband disturbs _load(t):

$s_{\mathrm{load}}\left(t\right)=\underset{j=1}{\overset{N_s}{\mathrm{\&Sigma;}}}[{\stackrel{\&CenterDot;}{g}}_j\left(t\right)-{\hat{f}}_{\mathrm{nb}}^j\left(t\right)],$ t∈((j-1)T _s+N _pT _s,jT _s+N _pT _s),j＝1,2,...,N _s

The concrete steps of described step (6) are:

Load signal waveform s after the elimination arrowband obtaining in step (5) is disturbed _load(t), utilize the signal correction template producing in step (4)

carry out correlation demodulation by correlation receiver and obtain symbol sebolic addressing

j=1,2 ..., N _smeet

Wherein,

j=1,2 ..., N _srepresent the load signal waveform that j band disturbs,

represent the estimation of the arrowband interference containing in j load symbol waveform,

for correlate template, N _srepresent load information symbol numbers.

The error rate that Fig. 2 produces for the inventive method emulation embodiment and the relation curve (" inhibition ") of signal to noise ratio have provided the performance curve (" unrestraint ") while not carrying out Suppression of narrow band interference simultaneously.Relatively two curves are visible, and the arrowband that the inventive method exists in paired pulses ultra-broadband signal effectively disturbs estimates and suppresses to have good bit error rate performance.

By reference to the accompanying drawings the specific embodiment of the present invention is described although above-mentioned; but not limiting the scope of the invention; one of ordinary skill in the art should be understood that; on the basis of technical scheme of the present invention, those skilled in the art do not need to pay various amendments that creative work can make or distortion still in protection scope of the present invention.

Claims (10)

1. the Suppression of narrow band interference system of the pulse ultra wideband receiver of compressed sensing, is characterized in that, comprises

Broadband filter module, utilizes broadband filter, the out-of-band noise in filtering impulse ultra-wideband signal;

Compressed sensing arrowband interference estimation block, with compressive sensing theory, in the pilot portion from impulse ultra-wideband signal, obtains arrowband and disturbs template;

Compressed sensing ultra broadband correlate template estimation module, utilizes the arrowband that compressed sensing arrowband interference estimation block produces to disturb template, eliminates the impact that in pilot tone, arrowband disturbs, and then utilizes compressive sensing theory, obtains ultra-broadband signal correlate template;

Signal delay module: load signal is suitably postponed, and be sent to load Suppression of narrow band interference module;

Load observation memory module: the temporary load observation sequence being obtained by observation module, and the observation sequence of each load signal is sent to arrowband Interference Estimation template successively;

Load Suppression of narrow band interference module: utilize subtracter to deduct the arrowband comprising in this load signal of being estimated by arrowband Interference Estimation template from each load signal waveform and disturb, and the load signal suppressing after arrowband disturbs is sent to correlation demodulation module;

Correlation demodulation module: the signal templates producing taking signal correction template generation module, as correlate template, utilizes the load signal after correlator and decision device disturb the inhibition arrowband of load Suppression of narrow band interference module output to carry out correlation demodulation.

2. the Suppression of narrow band interference system of the pulse ultra wideband receiver of compressed sensing as claimed in claim 1, is characterized in that,

Described broadband filter module comprises:

Ultra-wideband antenna module: receive the impulse ultra-wideband signal from wireless channel, and this signal is sent into receiving filter module;

Receiving filter module: paired pulses ultra-broadband signal filtering out-of-band noise and interference.

3. the Suppression of narrow band interference system of the pulse ultra wideband receiver of compressed sensing as claimed in claim 1, is characterized in that,

Described compressed sensing arrowband interference estimation block comprises:

Observation module: frequency pilot sign waveform and load symbol waveform are observed respectively, and pilot tone observation sequence is sent to pilot tone observation memory module, load observation sequence is sent to load observation memory module;

Pilot tone observation memory module: the observation sequence of temporary each frequency pilot sign waveform, and be sent to arrowband interference estimation block and subtracter block;

Arrowband interference estimation block: utilize OMP algorithm to estimate that arrowband disturbs, the arrowband containing in the frequency pilot sign waveform of estimation is disturbed and is sent to pilot tone arrowband disturbance-observer module, the arrowband containing in the load signal waveform of estimation is disturbed and is sent to load Suppression of narrow band interference module.

4. the Suppression of narrow band interference system of the pulse ultra wideband receiver of compressed sensing as claimed in claim 1, is characterized in that,

Described compressed sensing ultra broadband correlate template estimation module comprises:

Pilot tone arrowband disturbance-observer module: the frequency pilot sign arrowband that arrowband interference estimation block is estimated disturbs and observes respectively, and the arrowband disturbance-observer sequence that observation is obtained is sent to arrowband disturbance-observer memory module;

Arrowband disturbance-observer memory module: temporary arrowband disturbance-observer sequence, and be sent to subtracter block;

Subtracter block: utilize subtracter, deduct its corresponding arrowband disturbance-observer sequence from each pilot tone observation sequence, obtain and suppress the pilot tone observation sequence that arrowband disturbs;

The average module of vector: the pilot tone observation sequence that the inhibition arrowband obtaining in subtracter block is disturbed is sued for peace and is averaged operation acquisition average pilot observation sequence, to reduce the impact of additive white Gaussian noise, and is sent to signal correction template generation module;

Signal correction template generation module: the average pilot observation sequence obtaining according to the average module of vector, utilizes OMP algorithm reconstruction signal correlate template, and is sent to correlation demodulation module.

5. the narrow-band interference rejection method of the pulse ultra wideband receiver based on compressed sensing, is characterized in that, comprises the steps:

Step (1): receiving-transmitting sides is set up after communication, receiving terminal receives the impulse ultra-wideband signal from wireless channel, described impulse ultra-wideband signal comprises two parts: pilot portion and data payload part, and filtering is superimposed upon the out-of-band noise on impulse ultra-wideband signal;

Step (2): receiving terminal is based on compressive sensing theory, utilize the gaussian random matrix generating in advance as observing matrix, each frequency pilot sign waveform in impulse ultra-wideband signal pilot portion in step (1) is observed, obtain some pilot tone observation sequences, afterwards the arrowband containing in each frequency pilot sign waveform is disturbed and estimated;

Step (3):

With the observing matrix in step (2), the arrowband interference of estimating in step (2) is observed respectively, obtained some arrowbands disturbance-observer sequence;

The some pilot tone observation sequences that obtain from step (2) deduct corresponding arrowband disturbance-observer sequence, obtain and remove some pilot tone observation sequences that arrowband disturbs;

Step (4): the some pilot tone observation sequences that obtain in step (3) are averaged to operation and obtain average pilot observation sequence, to reduce the impact of additive white Gaussian noise, and according to the average pilot observation sequence obtaining, reconstruct signal correction template;

Step (5): utilize the observing matrix in step (2), each data symbol waveform in the data payload part of impulse ultra-wideband signal in step (1) is observed, obtain data payload observation sequence, afterwards the arrowband containing in each data symbol waveform is disturbed and estimated, obtain arrowband interference waveform, then utilize subtracter, from load signal waveform, cut arrowband interference waveform, eliminate the impact that arrowband disturbs;

Step (6): the load signal waveform after the elimination arrowband obtaining in step (5) is disturbed, utilize the signal correction template producing in step (4), carry out correlation demodulation by correlation receiver and obtain data symbol sequence.

6. method as claimed in claim 5, is characterized in that, the concrete steps of described step (1) are:

Suppose that the ultra-broadband signal waveform that receiving terminal obtains is:

$\begin{array}{c}s\left(t\right)=\underset{i=1}{\overset{N_p}{\mathrm{\&Sigma;}}}g(t-{\mathrm{iT}}_s)+\underset{j=1}{\overset{N_s}{\mathrm{\&Sigma;}}}{b}_{j}g(t-N_p{T}_s-j{T}_s)+n\left(t\right)+f_{\mathrm{nb}}\left(t\right)\\ =\underset{i=1}{\overset{N_p}{\mathrm{\&Sigma;}}}{\stackrel{\&CenterDot;}{g}}_i\left(t\right)+\underset{j=1}{\overset{N_s}{\mathrm{\&Sigma;}}}{\stackrel{\&CenterDot;}{g}}_j\left(t\right)\end{array}$

Wherein, N _prepresent frequency pilot sign number, all frequency pilot signs are 1, N _srepresent load information symbol numbers, T _sfor symbol period, b _jfor binary data modulation symbol and b _j∈ { 1,1}, f _nb(t) and n (t) represent respectively arrowband disturb and white Gaussian noise,

t ∈ ((i-1) T _s, iT _s), i=1,2 ..., N _prepresent the frequency pilot sign waveform that i band disturbs,

t ∈ ((j-1) T _s+ N _pt _s, jT+N _pt _s), j=1,2 ..., N _srepresent the load signal waveform that j band disturbs.

7. method as claimed in claim 5, is characterized in that, the concrete steps of described step (2) are:

Receiving terminal, based on compressive sensing theory, utilizes the gaussian random observing matrix generating in advance

each frequency pilot sign waveform in impulse ultra-wideband signal pilot portion in step (1) is observed, obtained N _pthe observation sequence of individual frequency pilot sign waveform

wherein y _i, i=1,2 ..., N _pfor M × 1 dimensional vector, represent the observation sequence of i frequency pilot sign waveform, and its k element is:

Wherein represent i the pilot signal waveform being disturbed,

for k observation waveform in observing matrix, N _pfor pilot tone number, M is observation waveform number;

When the arrowband interference containing in each frequency pilot sign waveform is reconstructed, the sparse dictionary that adopts inverse discrete Fourier transform IDFT matrix to disturb as arrowband:

${\mathrm{\&psi;}}_{\mathrm{nb}}=[{\mathrm{\&psi;}}_1\left(t\right),{\mathrm{\&psi;}}_2\left(t\right),\&CenterDot;\&CenterDot;\&CenterDot;,{\mathrm{\&psi;}}_N\left(t\right)]=\frac{1}{\sqrt{N}}{\left[\begin{array}{ccccc}{W}^0& W^0& W^0& \&CenterDot;\&CenterDot;\&CenterDot;& W^0\\ W^0& W^1& W^2& \&CenterDot;\&CenterDot;\&CenterDot;& W^{N-1}\\ W^0& W^2& W^4& \&CenterDot;\&CenterDot;\&CenterDot;& W^{2(N-1)}\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ W^0& W^{N-1}& W^{2(N-1)}& \&CenterDot;\&CenterDot;\&CenterDot;& W^{(N-1)(N-1)}\end{array}\right]}^{*}$

Wherein,

n is the sampling number of each frequency pilot sign waveform while carrying out nyquist sampling, the computing of * representing matrix conjugate transpose; Because arrowband disturbs as frequency-domain sparse, so the arrowband in each frequency pilot sign cycle disturbs

t ∈ ((i-1) T _s, iT _s), i=1,2 ..., N _pbe expressed as:

$f_{\mathrm{nb}}^i\left(t\right)=\underset{k=1}{\overset{N}{\mathrm{\&Sigma;}}}{\mathrm{\&psi;}}_k\left(t\right){\mathrm{\&theta;}}_k={\mathrm{\&psi;}}_{\mathrm{nb}}{\mathrm{\&theta;}}_{\mathrm{nb}}^i,$ t∈((i-1)T _s,iT _s),i＝1,2,...,N _p；

Wherein, Ψ _nbfor arrowband disturbs sparse dictionary,

be the projection coefficient vector that in i frequency pilot sign waveform, arrowband disturbs, and because arrowband interference is sparse,

in only have the active element of minority;

Disturb sparse dictionary Ψ by gaussian random observing matrix Φ and arrowband _nbobtain restructuring matrix:

V _nb＝ΦΨ _nb

Wherein, Φ is observing matrix, Ψ _nbfor disturbing sparse dictionary in arrowband;

According to restructuring matrix V _nbwith the N obtaining _pindividual observation sequence

then estimating just to estimate by OMP algorithm the arrowband containing in each frequency pilot sign waveform disturbs

wherein

t ∈ ((i-1) T _s, iT _s), i=1,2 ..., N _prepresent the estimation of the arrowband interference containing in i frequency pilot sign waveform.

8. method as claimed in claim 5, is characterized in that, the concrete steps of described step (3) are:

With the gaussian random observing matrix in step (2)

the arrowband of estimating in step (2) is disturbed

observe respectively, obtain corresponding arrowband disturbance-observer sequence

wherein

i=1,2 ..., N _pfor M × 1 dimensional vector, represent the estimation that the arrowband to containing in i frequency pilot sign waveform disturbs observes and obtain observation sequence, and its k element is:

Wherein

represent the estimation of the arrowband interference containing in i frequency pilot sign waveform,

for k observation waveform in observing matrix, N _pfor pilot tone number, M is observation waveform number;

The some pilot tone observation sequences that obtain from step (2)

in deduct corresponding arrowband disturbance-observer sequence

impact, obtain remove arrowband disturb pilot tone observation sequence;

$\begin{array}{c}[y_{1-f},y_{2-f},\&CenterDot;\&CenterDot;\&CenterDot;,y_{N_p-f}]=[y_1,y_2,\&CenterDot;\&CenterDot;\&CenterDot;y_{N_p}]-[y_f^1,y_f^2,\&CenterDot;\&CenterDot;\&CenterDot;,y_f^{N_p}]\\ =[y_1-y_f^1,y_2-y_f^2,\&CenterDot;\&CenterDot;\&CenterDot;,y_{N_p}-y_f^{N_p}]\end{array}$

Wherein,

i=1,2 ..., N _prepresent that i frequency pilot sign waveform suppresses the observation sequence after arrowband disturbs.

9. method as claimed in claim 5, is characterized in that, the concrete steps of described step (4) are:

Pilot tone observation sequence after the some inhibition arrowband obtaining in step (3) is disturbed

be averaged operation and obtain average pilot observation sequence

to reduce the impact of additive white Gaussian noise,

$\stackrel{\&OverBar;}{y}=\frac{1}{N_p}\underset{i=1}{\overset{N_p}{\mathrm{\&Sigma;}}}{y}_{i-f}=\frac{1}{N_p}\underset{i=1}{\overset{N_p}{\mathrm{\&Sigma;}}}(y_i-y_f^i)$

Wherein, y _i-f, i=1,2 ..., N _prepresent that i pilot waveform suppresses the observation sequence after arrowband disturbs;

For the reconstruct of signal correction template, the sparse dictionary of employing is the sparse dictionary Ψ of characteristic vector _g, its production process is as follows:

Due to the time-varying characteristics of pulse ultra-broad band channel, the impulse ultra-wideband signal g (t) receiving in step (1) is a random process in essence; Therefore, obtain its covariance function R (t-τ) by a large amount of channel samples,

R(t-τ)＝E[g(t)g(τ+t)]

Suppose λ ₁> λ ₂> λ ₃> ... > λ _nrepresent the characteristic value of Fredholm integral operator, wherein, N is the sampling number of each frequency pilot sign waveform while carrying out nyquist sampling; Have for R (t-τ):

$\&Integral;R(t-\mathrm{\&tau;})u_j\left(\mathrm{\&tau;}\right)\mathrm{d\&tau;}={\mathrm{\&lambda;}}_j{u}_j\left(t\right)$

Wherein, u _j(t) be λ _jcharacteristic of correspondence vector, and { u _j(t) } be the complete set of one group of orthogonal basis function, meet:

$\&Integral;u_i\left(t\right)u_j\left(t\right)=\begin{array}{c}\left\{\begin{array}{cc}0& i\&NotEqual;j\\ 1& i=j\end{array}\right.\end{array},i,j=\mathrm{1,2,3},\&CenterDot;\&CenterDot;\&CenterDot;,N$

Therefore, [u ₁(t), u ₂(t), u ₃(t) ..., u _n(t)] formed the orthogonal basis of one group of g (t), this group orthogonal basis is the sparse dictionary of characteristic vector;

By step (2) gaussian random observing matrix

with the sparse dictionary Ψ of characteristic vector _g, obtain correlate template restructuring matrix,

V _g＝ΦΨ _g

In conjunction with the average pilot observation sequence obtaining in this step

facility obtains signal correction template with the reconstruct of OMP algorithm

t ∈ (0, T _s).

10. method as claimed in claim 5, is characterized in that, the concrete steps of described step (5) are:

First, utilize the gaussian random observing matrix in step (2)

each load symbol waveform in impulse ultra-wideband signal loading section in step (1) is observed, obtained the observation sequence of each load symbol waveform

wherein for M × 1 dimensional vector, represent the observation sequence of j load symbol waveform, and its k element is:

Wherein

represent j the load signal waveform being disturbed,

for k observation waveform in observing matrix, N _sfor load symbol numbers, M is observation waveform number;

Then, according to the method for estimation of in step (2), the arrowband that contains in pilot signal being disturbed, the estimation that the arrowband that utilizes OMP algorithm to obtain to contain in each load symbol waveform disturbs

represent the estimation of the arrowband interference containing in j load symbol waveform, and

be expressed as:

${\hat{f}}_{\mathrm{nb}}^j\left(t\right)={\mathrm{\&psi;}}_{\mathrm{nb}}{\widehat{\mathrm{\&theta;}}}_{\mathrm{nb}}^j,$ t∈((j-1)T _s+N _pT _s,jT _s+N _pT _s),j＝1,2,...,N _s；

Wherein, Ψ _nbfor arrowband disturbs sparse dictionary,

for utilizing the projection coefficient vector of arrowband interference in j the load symbol waveform that OMP algorithm obtains

j=1,2 ..., N _sestimation;

Finally, utilize subtracter, from load signal waveform, eliminate the impact that arrowband disturbs, obtain and suppress the load signal waveform s that arrowband disturbs _load(t):

$s_{\mathrm{load}}\left(t\right)=\underset{j=1}{\overset{N_s}{\mathrm{\&Sigma;}}}[{\stackrel{\&CenterDot;}{g}}_j\left(t\right)-{\hat{f}}_{\mathrm{nb}}^j\left(t\right)],$ t∈((j-1)T _s+N _pT _s,jT _s+N _pT _s),j＝1,2,...,N _s。

Images