CN105207966A - Compressed sensing PIE (Pulse Interference Elimination) system based on space-frequency coding - Google Patents

Abstract

The invention relates to a compressed sensing PIE system based on space-frequency coding. In a transmitter, a channel encoder, a modulator, an orthogonal converter, a symbol interleaver, a serial-to-parallel converter and a space-frequency encoder are successively connected in series, the space-frequency encoder is connected with two subcarrier mapping devices of the same structure, and each subcarrier mapping device is successively connected with an up sampler, a reverse discrete Fourier transformer, a cyclic prefix adder, a parallel-to-serial converter, a D/A conversion and RF emitter and an emission antenna. A receiver comprises a reception antenna, an RF front-end converter, an A/D converter, a Doppler frequency offset estimator, a Doppler frequency offset compensator, a cyclic prefix remover, a discrete Fourier transformer, a compressed sensing interference eliminator, a frequency-domain down sampler, a channel estimator, a space-frequency decoder, a symbolic solution interleaver, a reverse orthogonal converter and a channel decoder successively in serial connection. The compressed sensing algorithm can be directly utilized to reconstruct pulse signals of a range finder and converts the pulse signals to the frequency domain for elimination, and thus, the pulse signals of the range finder are effectively inhibited.

Description

A kind of compressed sensing pulse interference suppression system based on space-frequency coding

Technical field

The present invention relates to a kind of disturbance restraining method.Particularly relate to a kind of compressed sensing pulse interference suppression system based on space-frequency coding.

Background technology

L frequency band digital aviation communication system is the following Air-Ground cell mobile communication systems disposed along air route, land of civil aviaton, this system is mainly air route, land, the airborne vehicle on termination environment and airport provides air traffic control, airline's operation management data communication service, and this system is one of technical way of the following aviation mobile communications system of civil aviaton.According to the planning of International Civil Aviation Organization, L frequency band digital aviation communication system by be deployed in L frequency range rangefinder in embedded mode radio frequency channel between, take transmission bandwidth 500KHz.Due to rangefinder and OFDM signal frequency spectrum, to there is part overlapping, and range-finding system transmitting power is higher, therefore unavoidably produces the problem of rangefinder signal disturbing L frequency band digital aviation communication system OFDM receiver.

At present, the research for L frequency band digital aviation communication system rangefinder AF panel mainly concentrates on single antenna interference cancellation and array antenna AF panel two aspects.In single antenna interference cancellation, document EPPLEU, HOFFMANNF, SCHELLM.ModelingdmeinterferenceimpactonL-DCAS1 [C] //IEEE.IntegratedCommunications, NavigationandSurveillanceConference (ICNS) .Herndon, VA:IEEE, 2012:G7-1-G7-13.And EPPLEU, BRANDESS, GLIGOREVICS, etal.ReceiveroptimizationforL-DCAS1 [C] //DASC'09.IEEE/AIAA28th.DigitalAvionicsSystemsConference, 2009.Orlando, FL:IEEE, 2009:4.B.1-1-4.B.1-12, utilize rangefinder interference signal to present the feature of impulse disturbances in time domain, propose pulse and extinguish AF panel method; Extinguish for solving pulse the problem causing orthogonal frequency-division multiplex singal to produce inter-sub-carrier interference, document [9-11] proposes pulse and extinguishes inter-sub-carrier interference compensation method.

Based in the L frequency band digital aviation communication system OFDM receiver rangefinder pulse interference suppression of array antenna, document Haitao Liu, Liu Yazhou, Cheng Wei. the disturbance restraining method [J] of united orthogonal projection and blind adaptive beamforming. (system engineering and electronic technology .2015, Network Priority is published) propose first to utilize orthogonal projection algorithm to eliminate the interference of high strength rangefinder, then adopt blind adaptive beamforming method to extract the method for OFDM direct projection footpath signal; Eliminate on impulse disturbances method basis in rectangular projection.Document Haitao Liu, Liu Yazhou, Zhang Xuejun. the rangefinder pulse interference suppression method [J] of united orthogonal projection and CLEAN. (signal transacting .2015.6, Network Priority is published) propose further to utilize CLEAN algorithm to estimate the method for OFDM direct projection footpath direction of signal, and extract OFDM direct projection footpath signal by conventional beamformer algorithm.

The method that OFDM receiver impulse disturbances is eliminated mainly comprises: pulse extinguishing method, associating pulse extinguishing and inter-sub-carrier interference compensation method, rectangular projection and blind adaptive beamforming method, rectangular projection and the direction of signal estimation technique, describes the shortcomings and deficiencies of these technology below respectively.

The basic thought of pulse extinguishing method: receiver, according to the amplitude of Received signal strength, first determines the position that pulse interference signal occurs, then, the signal sample that there is impulse disturbances is set to zero, to eliminate the impact of impulse disturbances.In actual applications, there is the problem of following two aspects in the method: the position that pulse interference signal occurs not easily is determined, pulse causes inter-sub-carrier interference after extinguishing.When determining the position that pulse interference signal occurs, the method that usual employing thresholding compares, when Received signal strength amplitude exceedes the thresholding of setting, receiver thinks that these signal samples comprise impulse disturbances, because orthogonal frequency-division multiplex singal self peak-to-average force ratio (PAPR) is higher, and propagate through multipath channel, receiver not easily determines that pulse clipping is exterminated a family limit usually, and this will cause receiver bit error loss of energy.Another aspect, after receiver adopts pulse to extinguish, make Received signal strength produce inter-sub-carrier interference, inter-sub-carrier interference finally also will cause the deterioration of receiver bit error performance.

The basic thought of associating pulse extinguishing and inter-sub-carrier interference compensation method: first utilize pulse extinguishing method to eliminate impulse disturbances, then inter-sub-carrier interference is reconstructed further, last paired pulses extinguishes the inter-sub-carrier interference produced and compensates, thus reduce the impact that method generation inter-sub-carrier interference is extinguished in pulse, improve the reliability of link transmission.There is the problem of two aspects in the method, the first pulse clipping is exterminated a family to limit and not easily determined, the computational complexity of the reconstruct of secondary iteration inter-sub-carrier interference and compensation is higher.

The basic thought of the disturbance restraining method of united orthogonal projection and blind adaptive beamforming: receiver is first by eliminating high strength rangefinder impulse disturbances by Received signal strength vector project to the method for the interference signal orthogonal complement space, then the symmetry characteristic of orthogonal frequency-division multiplex singal Cyclic Prefix is utilized, maximize criterion based on product metric in desired signal and reference signal vector and obtain Wave beam forming weights, and extract OFDM direct projection footpath signal by Beamforming Method.The method Shortcomings in practical application: when rangefinder interference is less, rectangular projection AF panel poor-performing.

The basic thought of the rangefinder pulse interference suppression method of united orthogonal projection and CLEAN: receiver is first by eliminating high strength rangefinder impulse disturbances by Received signal strength vector project to the method for the interference signal orthogonal complement space, then the symmetry characteristic of orthogonal frequency-division multiplex singal Cyclic Prefix is utilized, adopt CLEAN algorithm estimated signal come to, then by conventional beam forming extraction OFDM direct projection footpath signal.In practical application, the method exists main not enough: when receiving orthogonal frequency-division multiplex singal power and be lower and data snap being less, the direction of signal estimated performance of put forward the methods is poor.

Summary of the invention

Technical problem to be solved by this invention is, provides a kind of interference suppressioning effect better based on the compressed sensing pulse interference suppression system of space-frequency coding.

The technical solution adopted in the present invention is: a kind of compressed sensing pulse interference suppression system based on space-frequency coding, includes based on the OFDM transmitter of space-frequency coding and the compressed sensing pulse interference suppression receiver based on sky decoding frequently.

The described OFDM transmitter based on space-frequency coding includes: the channel encoder for docking collection of letters source signal coding be connected in series successively, for the source signal after coding being modulated into the modulator of orthogonal phase modulation symbol, orthogonal phase modulation symbolic is carried out to the quadrature transformer of orthogonal transform, the symbol of orthogonal transform is carried out to the symbol interleaver of symbol interleaving, signal after symbol interleaving is converted to the deserializer of parallel orthogonal phase modulation symbol and the space-frequency encoder of generation two-way coded signal that parallel each OFDM symbol is encoded successively, described space-frequency encoder connects the identical subcarrier mapper for carrying out subcarrier mapping respectively to two-way coded signal of two structures respectively, the output of each described subcarrier mapper be connected with in turn sub-carrier map after signal carry out the up-sampler of 4 times of up-samplings, signal after 4 times of up-samplings is carried out to the inverse discrete Fourier transform device of Fourier transform, for the formation of the Cyclic Prefix adder of parallel complete OFDM symbol, parallel OFDM symbol is carried out to the parallel-to-serial converter of parallel-serial conversion, digital-to-analogue conversion is carried out to the signal after parallel-serial conversion and the D/A conversion of launching and radiofrequency launcher and transmitting antenna,

The described compressed sensing pulse interference suppression receiver based on sky decoding frequently includes and is connected in series successively: the reception antenna of the signal that the OFDM transmitter for receiving based on space-frequency coding is launched, received signal is converted into successively radio-frequency front-end transducer and the A/D converter of analog baseband signal and digital baseband signal, for compensating the Doppler shift compensator of the Doppler shift of Received signal strength, for removing the cyclic prefix remover of the Cyclic Prefix of orthogonal frequency-division multiplex singal, effective orthogonal frequency-division multiplex singal is carried out to the first discrete Fourier transform device of Fourier transform, for reconstructing interference signal and eliminating the compressed sensing interference eliminator of interference signal, signal after elimination interference is carried out to the frequency domain down-sampler of frequency domain sample, for completing the empty decoder frequently of the empty process of decoding frequently, for carrying out the symbol deinterleaver of symbolic solution intertexture to signal, signal after deinterleaving is carried out to the inverse orthogonal transformation device of inverse orthogonal transformation, for carrying out the demodulator of PM signal PM of data demodulates and the channel decoder for going out to launch bit sequence valuation to the signal recuperation after demodulation, wherein, described radio-frequency front-end transducer also will be connected Doppler shift compensator by the Doppler shift estimator of the Doppler shift for estimating received signal with the signal output part of A/D converter, the signal output part of described frequency domain down-sampler also will pass through for estimating that two transmitting antennas divide the channel estimator of two channel frequency domain response being clipped to reception antenna to connect empty decoder frequently.

Described radio-frequency front-end transducer and A/D converter have employed four times of over-samplings to analog baseband signal in modulus sampling process.

The described course of work based on the space-frequency encoder in the OFDM transmitter of space-frequency coding is: first by the OFDM symbol vector representation of the kth after deserializer serioparallel exchange be:

X(k)＝[X(k,0),X(k,1),...,X(k,l),...,X(k,L-1)] ^T(1)

Wherein, k=1,2 ..., K, l=0,1 ..., L-1, T represent transpose operation, and X (k, l) represents l subcarrier data in a kth OFDM symbol; Data vector X is obtained to after kth OFDM symbol vector X (k) coding ₁(k) and X ₂(k), wherein X ₁k () is expressed as:

X ₁(k)＝[X(k,0),-X ^*(k,1),...,X(k,L-2),-X ^*(k,L-1)] ^T(2)

X ₂k () is expressed as:

X ₂(k)＝[X(k,1),X ^*(k,0),...,X(k,L-1),X ^*(n,L-2)] ^T(3)

Specific coding rule is wherein:

X ₁(k,i)＝X(k,i)

X ₁(k,i+1)＝-X ^*(k,i+1)

(4)

X ₂(k,i)＝X(k,i+1)

X ₂(k,i+1)＝X ^*(k,i)i＝0,2,4,6,...

Wherein, X _α(k, i), α=1,2 i-th subcarrier data representing a kth OFDM symbol in coded vector; Thus obtain corresponding relation:

X _1,e(k)＝X _e(k)(5-a)

$X_{1,o}\left(k\right)=-X_o^{*}\left(k\right)---(5-b)$

X _2，e(k)＝X _o(k)(5-c)

$X_{2,o}\left(k\right)=X_e^{*}\left(k\right)---(5-d).$

Described including based on the compressed sensing interference eliminator in the compressed sensing pulse interference suppression receiver of sky decoding frequently disturbs reconstructor and the second discrete Fourier transform device.

The course of work of described compressed sensing interference eliminator is:

Definition Ω represents frequency domain transmission signal phasor null symbol position number form set, () _Ωrepresent the subvector or submatrix that are made up of element corresponding to sequence number in set omega or row; For frequency-region signal vector Y _kthe signal extracting null symbol subchannel obtains:

$\begin{array}{c}{\left(Y_k\right)}_{\mathrm{\Ω}}={\left({\mathrm{\ΛX}}_k^{ov}\right)}_{\mathrm{\Ω}}+{\left({\mathrm{Fi}}_k\right)}_{\mathrm{\Ω}}+{\left(N_k\right)}_{\mathrm{\Ω}}\\ ={\left(F\right)}_{\mathrm{\Ω}}{i}_k+{\left(N_k\right)}_{\mathrm{\Ω}}\end{array}---\left(12\right)$

In formula, Y _kthe frequency-region signal vector obtained after representing vector Fourier transform to received signal, represent a kth OFDM symbol frequency domain transmission signal phasor, Λ represents frequency domain channel transmission matrix, and F represents Discrete Fourier transform and is unitary matrice, N _krepresent Frequency domain noise signal phasor, i _krepresent the rangefinder signal phasor of channel input;

Utilize rangefinder pulse signal i _ktime domain sparse characteristic, use compression sensing method reconstruct rangefinder pulse signal, suppose that the kth rangefinder pulse interference signal that receiver reconstructs is designated as further by Fourier transform, Reconstruction Impulse signal is transformed into frequency domain then impulse disturbances elimination is carried out at frequency domain, the kth frequency-domain received signal vector be improved

$\begin{array}{c}{\tilde{Y}}_k=Y_k-F{\hat{i}}_k\\ ={\mathrm{\ΛX}}_k^{ov}+F\left(i_k-{\hat{i}}_k\right)+N_k\end{array}---\left(13\right)$

Suppose that receiver can Accurate Reconstruction rangefinder pulse signal, namely the kth frequency-domain received signal vector estimated value approximate representation then improved is:

${\tilde{Y}}_k\≈{\mathrm{\ΛX}}_k^{ov}+N_k---\left(14\right).$

The course of work of described empty decoder and channel estimator frequently for completing the empty process of decoding is frequently:

By the frequency domain orthogonal frequency-division multiplex singal vector valuation that frequency domain down-sampler exports be modeled as:

${\tilde{Z}}_k=H_1\left(k\right)\·X_1\left(k\right)+H_2\left(k\right)\·X_2\left(k\right)+N\left(k\right)---\left(17\right)$

Wherein, X ₁(k) and X ₂k () represents a kth OFDM space-frequency coding vector of the first antenna and the transmission of the second antenna respectively; H ₁(k) and H ₂k () correspondence represents channel frequency domain response, N (k) represents the frequency domain white complex gaussian noise vector of channel input, dot product between ". " representative vector element; By the valuation of frequency domain orthogonal frequency-division multiplex singal vector be not difficult to draw by even item form vector expression

${\tilde{Z}}_{k,e}=H_{1,e}\left(k\right)X_{1,e}\left(k\right)+H_{2,e}\left(k\right)X_{2,e}\left(k\right)+N_e\left(k\right)---\left(18\right)$

With odd term vector expression

${\tilde{Z}}_{k,o}=H_{1,o}\left(k\right)X_{1,o}\left(k\right)+H_{2,o}\left(k\right)X_{2,o}\left(k\right)+N_o\left(k\right)---\left(19\right)$

Suppose that receiver is known or accurately can estimate channel frequency domain response H ₁(k) and H ₂k (), then send the valuation of OFDM symbol vector can be expressed as:

${\hat{X}}_e\left(k\right)=H_{1,e}^{*}\left(k\right){\tilde{Z}}_{k,e}+H_{2,o}\left(k\right){\tilde{Z}}_{k,o}^{*}---\left(20\right)$

${\hat{X}}_o\left(k\right)=H_{2,e}^{*}\left(k\right){\tilde{Z}}_{k,e}-H_{1,o}\left(k\right){\tilde{Z}}_{k,o}^{*}---\left(21\right)$

Wherein, with respectively by representing the data vector that middle even item and odd term are formed; By corresponding relation:

X _1,e(k)＝X _e(k)(5-a)

$X_{1,o}\left(k\right)=-X_o^{*}\left(k\right)---(5-b)$

X _2,e(k)＝X _o(k)(5-c)

$X_{2,o}\left(k\right)=X_e^{*}\left(k\right)---(5-d)$

Bring into with expression formula, replace the X in formula successively _{1, e}(k), X _{2, e}(k), X _{1, o}(k), X _{2, o}(k):

$\begin{array}{c}{\tilde{Z}}_{k,e}=H_{1,e}\left(k\right)X_e\left(k\right)+H_{2,e}\left(k\right)X_o\left(k\right)+N_e\left(k\right)\\ {\tilde{Z}}_{k,o}=-H_{1,o}\left(k\right)X_o^{*}\left(k\right)+H_{2,o}\left(k\right)X_e^{*}\left(k\right)+N_o\left(k\right)\end{array}---\left(22\right)$

And bring into with expression formula after abbreviation, and consider adjacent sub-carrier channel frequency domain response approximately equal, i.e. H _{1, o}(k) ≈ H _{1, e}(k), H _{2, o}(k) ≈ H _{2, e}(k):

${\hat{X}}_e\left(k\right)=\left({\left|H_{1,e}\left(k\right)\right|}^2+{\left|H_{2,e}\left(k\right)\right|}^2\right)X_e\left(k\right)+N_1\left(k\right)---\left(23\right)$

${\hat{X}}_o\left(k\right)=\left({\left|H_{1,o}\left(k\right)\right|}^2+{\left|H_{2,o}\left(k\right)\right|}^2\right)X_o\left(k\right)+N_2\left(k\right)$

Thus recover original transmission OFDM symbol vector X (n):

$\begin{array}{c}{X}_e\left(k\right)={\hat{X}}_e\left(k\right)/({\left|H_{1,e}\left(k\right)\right|}^2+{\left|H_{2,e}\left(k\right)\right|}^2)+{\stackrel{\‾}{N}}_1\left(k\right)\\ X_o\left(k\right)={\hat{X}}_o\left(k\right)/({\left|H_{1,o}\left(k\right)\right|}^2+{\left|H_{2,o}\left(k\right)\right|}^2)+{\stackrel{\‾}{N}}_2\left(k\right)\end{array}---\left(24\right)$

Wherein, N ₁(k), N ₂(k), be the equivalent noise vector produced in abbreviation process.

A kind of compressed sensing pulse interference suppression system based on space-frequency coding of the present invention, compared with extinguishing disturbance restraining method with pulse, there is not pulse clipping and to exterminate a family limit offering question in method of the present invention, there is not the problem of inter-sub-carrier interference, the present invention directly can utilize compressed sensing algorithm to reconstruct rangefinder pulse signal and is transformed into frequency domain and eliminate, thus effectively suppresses rangefinder pulse signal.

Compared with pulse amplitude limit method disturbance restraining method, there is not pulse amplitude limit thresholding offering question in method of the present invention, there is not residual impulse interference problem, and the present invention has better link error transmission performance.

Project compared with the disturbance restraining method of blind adaptive beamforming with united orthogonal, method of the present invention does not need multiple antennas to receive, do not need known or estimated disturbance signal number in advance, and interference suppression algorithm of the present invention is suitable for the interference signal of different capacity, therefore the scope of application of the present invention is more wide.

Project compared with the rangefinder pulse interference suppression method of CLEAN with united orthogonal, method of the present invention do not need by the theoretical estimated disturbance signal of Estimation of Spatial Spectrum come to, and not needing conventional beamformer algorithm to extract desired signal, the present invention simultaneously does not need multiple antennas and a large amount of data snaps.Therefore, method of the present invention is more practical in the future, and interference suppressioning effect is better.

Accompanying drawing explanation

Fig. 1 is the OFDM transmitter theory diagram based on space-frequency coding in the present invention;

Fig. 2 is the compressed sensing pulse interference suppression receiver principle block diagram based on sky decoding frequently in the present invention;

Fig. 3 a is time domain signal waveform figure before rangefinder pulse interference signal of the present invention reconstruct;

Fig. 3 b is time domain signal waveform figure after rangefinder pulse interference signal of the present invention reconstruct;

Fig. 4 is the normalized mean squared error figure of rangefinder impulse reconstruction after filtering;

Fig. 5 a is the power spectrum (transmit signal power is 1) that orthogonal frequency division multiplex ransmitting penetrates signal;

Fig. 5 b is the power spectrum (rangefinder signal carrier is biased 500KHz, signal interference ratio SIR=-10dB) of rangefinder pulse signal residual interference signal after equivalent frequency overlapped-resistable filter;

Fig. 5 c is the power spectrum chart of receiver Received signal strength;

Fig. 5 d is the power spectrum chart of signal after pulse interference suppression;

Fig. 6 exports Signal to Interference plus Noise Ratio change curve after suppressor pulse interference;

Fig. 7 is system errors performance curve (orthogonal phase modulation) figure.

Embodiment

Below in conjunction with embodiment and accompanying drawing, a kind of compressed sensing pulse interference suppression system based on space-frequency coding of the present invention is described in detail.

A kind of compressed sensing pulse interference suppression system based on space-frequency coding of the present invention, includes based on the OFDM transmitter of space-frequency coding and the compressed sensing pulse interference suppression receiver based on sky decoding frequently.

As shown in Figure 1, OFDM transmitter based on space-frequency coding of the present invention includes: the channel encoder 11 for docking collection of letters source signal coding be connected in series successively, for the source signal after coding being modulated into the modulator 12 of orthogonal phase modulation symbol, orthogonal phase modulation symbolic is carried out to the quadrature transformer 13 of orthogonal transform, the symbol of orthogonal transform is carried out to the symbol interleaver 14 of symbol interleaving, signal after symbol interleaving is converted to the deserializer 15 of parallel orthogonal phase modulation symbol and the space-frequency encoder 16 of generation two-way coded signal that parallel each OFDM symbol is encoded successively, described space-frequency encoder 16 connects the identical subcarrier mapper 17a/17b for carrying out subcarrier mapping respectively to two-way coded signal of two structures respectively, the output of each described subcarrier mapper 17a/17b be connected with in turn sub-carrier map after signal carry out the up-sampler 18a/18b of 4 times of up-samplings, signal after 4 times of up-samplings is carried out to the inverse discrete Fourier transform device 19a/19b of Fourier transform, for the formation of the Cyclic Prefix adder 10a/10b of parallel complete OFDM symbol, parallel OFDM symbol is carried out to the parallel-to-serial converter 11a/11b of parallel-serial conversion, digital-to-analogue conversion is carried out to the signal after parallel-serial conversion and the D/A conversion of launching and radiofrequency launcher 12a/12b and transmitting antenna 13a/13b.

Source signal to be orthogonal phase modulation symbolic by sending into modulators modulate after channel encoder, and orthogonal phase modulation symbol, after orthogonal transform and symbol interleaving, is sent into deserializer and converted parallel orthogonal phase modulation symbol to.Space frequency block coding device is encoded successively to parallel each modulation symbol, generates the signal after two-way coding.Generating orthogonal frequency division multiplexing symbol subcarrier data after two-way coded signal is all mapped by subcarrier.OFDM symbol sends into inverse discrete fourier (inverse discrete fourier) converter after 4 times of up-samplings, forms parallel complete OFDM symbol after the signal that inverse discrete fourier (inverse discrete fourier) converter exports adds Cyclic Prefix.Parallel OFDM symbol by parallel-serial conversion become serial OFDM symbol, be analog baseband signal by digital-to-analogue conversion, convert radiofrequency signal to by radiofrequency emitting module again, last radiofrequency signal is burst into channel by antenna.

The described course of work based on the space-frequency encoder 16 in the OFDM transmitter of space-frequency coding is: for convenience of describing space-frequency coding principle, first by the OFDM symbol vector representation of the kth after deserializer 15 serioparallel exchange be:

X(k)＝[X(k,0),X(k,1),...,X(k,l),...,X(k,L-1)] ^T(1)

Wherein, k=1,2 ..., K, l=0,1 ..., L-1, T represent transpose operation, and X (k, l) represents l subcarrier data in a kth OFDM symbol; Data vector X is obtained to after kth OFDM symbol vector X (k) coding ₁(k) and X ₂(k), wherein X ₁k () is expressed as:

X ₁(k)＝[X(k,0),-X ^*(k,1),...,X(k,L-2),-X ^*(k,L-1)] ^T(2)

X ₂k () is expressed as:

X ₂(k)＝[X(k,1),X ^*(k,0),...,X(k,L-1),X ^*(n,L-2)] ^T(3)

Specific coding rule is wherein:

X ₁(k,i)＝X(k,i)

X ₁(k,i+1)＝-X ^*(k,i+1)

(4)

X ₂(k,i)＝X(k,i+1)

X ₂(k,i+1)＝X ^*(k,i)i＝0,2,4,6,...

Wherein, X _α(k, i), α=1,2 i-th subcarrier data representing a kth OFDM symbol in coded vector; Composite type (1), (2), (3) formula in conjunction with above-mentioned coding rule, thus obtain corresponding relation:

X _1,e(k)＝X _e(k)(5-a)

$X_{1,o}\left(k\right)=-X_o^{*}\left(k\right)---(5-b)$

X _2,e(k)＝X _o(k)(5-c)

$X_{2,o}\left(k\right)=X_e^{*}\left(k\right)---(5-d)$

By data vector X ₁(k) and X ₂k (), through subcarrier mapping, up-sampling and inverse discrete Fourier transform, is burst into channel respectively by transmitting antenna 1 and transmitting antenna 2 after increasing Cyclic Prefix.

As shown in Figure 2, compressed sensing pulse interference suppression receiver based on sky frequently decoding of the present invention includes and is connected in series successively: the reception antenna 21 of the signal that the OFDM transmitter for receiving based on space-frequency coding is launched, received signal is converted into successively radio-frequency front-end transducer and the A/D converter 22 of analog baseband signal and digital baseband signal, for compensating the Doppler shift compensator 24 of the Doppler shift of Received signal strength, for removing the cyclic prefix remover 25 of the Cyclic Prefix of orthogonal frequency-division multiplex singal, effective orthogonal frequency-division multiplex singal is carried out to the first discrete Fourier transform device 26 of Fourier transform, for reconstructing interference signal and eliminating the compressed sensing interference eliminator 27 of interference signal, to the frequency domain down-sampler 28 that the signal after eliminating interference is sampled, for completing the empty decoder 29 frequently of the empty process of decoding frequently, for carrying out the symbol deinterleaver 211 of symbolic solution intertexture to signal, signal after deinterleaving is carried out to the inverse orthogonal transformation device 212 of inverse orthogonal transformation, for carrying out the demodulator of PM signal PM 213 of data demodulates and the channel decoder 214 for going out to launch bit sequence valuation to the signal recuperation after demodulation, wherein, described radio-frequency front-end transducer also will be connected Doppler shift compensator 24 by the Doppler shift estimator 23 for the Doppler shift of estimating received signal with the signal output part of A/D converter 22, the signal output part of described frequency domain down-sampler 28 also will pass through for estimating that two transmitting antenna 13a/13b divide the channel estimator 210 of two channel frequency domain response being clipped to reception antenna 21 to connect empty decoder 29 frequently.Described including based on the compressed sensing interference eliminator 27 in the compressed sensing pulse interference suppression receiver of sky decoding frequently disturbs reconstructor 271 and the second discrete Fourier transform device 272.

From the radiofrequency signal of antenna, analog baseband signal is converted to through radio-frequency front-end, analog baseband signal is digital baseband signal by analog-to-digital conversion subsequently, for the demodulation avoiding rangefinder pulse signal sampling to produce spectral aliasing interference orthogonal frequency-division multiplex singal, in modulus sampling process, four times of over-samplings be have employed to analog baseband signal, and the digital baseband signal after analog-to-digital conversion is estimated and sends into cyclic prefix remover after compensation to remove Cyclic Prefix through Doppler shift.

Suppose the built vertical timing synchronization of receiver, then sampled output signal is after removing Cyclic Prefix, and a kth OFDM symbol Received signal strength vector representation is y _k=[y _{k, 1}..., y _k,n..., y _{k, VN}] ^t, be modeled as:

y _k＝H _1kx _1k+H _2kx _2k+i _k+n _k(7)

Wherein, x _1kfor orthogonal frequency-division multiplex singal vector launched by transmitting antenna 1, H _1kfor transmitting antenna 1 is to reception antenna channel time domain transmission matrix; x _2kfor orthogonal frequency-division multiplex singal vector launched by transmitting antenna 2, H _2kfor transmitting antenna 2 is to reception antenna channel time domain transmission matrix; i _kfor the rangefinder signal phasor i of channel input _k=[i _{k, 1}..., i _k,n... i _{k, VN}] ^t, n _kfor white complex gaussian noise vector n _k=[n _{k, 1}..., n _k,n... n _{k, VN}] ^t, wherein, n _kthe average of each component is 0, and variance is k=1,2 ... K.For convenience of deriving, two-way antenna transmission signal phasor is merged, order:

H _kx _k＝H _1kx _1k+H _2kx _2k(8)

Then (8) formula is rewritten as

y _k＝H _kx _k+i _k+n _k(9)

Suppose that channel is in a kth OFDM symbol transmission period, channel transfer characteristic keeps constant, i.e. H _k=H, k=1 ..., K, utilizes the cycle characteristics of channel transfer matrices H further, channel transfer matrices is decomposed into H=F ^hΛ F, formula (9) is expressed as further:

$y_k=F^H{\mathrm{\ΛX}}_k^{ov}+i_k+n_k---\left(10\right)$

In formula, F ^hfor inverse discrete Fourier transform matrix, Λ is frequency domain channel transmission matrix, for a kth OFDM symbol frequency domain transmission signal phasor.By discrete Fourier transform (DFT) by Received signal strength vector y _kbe transformed into frequency domain:

$\begin{array}{c}{Y}_k={\mathrm{Fy}}_k\\ ={\mathrm{\ΛX}}_k^{ov}+{\mathrm{Fi}}_k+N_k\end{array}---\left(11\right)$

In formula, N _k=Fn _kfor Frequency domain noise signal phasor, consider that F is unitary matrice, therefore Frequency domain noise signal phasor N _kstatistical property remains unchanged, and is still the random vector of multiple Gaussian Profile.

The course of work of described compressed sensing interference eliminator 27 is: first define Ω and represent frequency domain transmission signal phasor null symbol position number form set, () _Ωrepresent the subvector (submatrix) be made up of the element (OK) that sequence number in set omega is corresponding; For frequency-region signal vector Y _kthe signal extracting null symbol subchannel obtains:

$\begin{array}{c}{\left(Y_k\right)}_{\mathrm{\Ω}}={\left({\mathrm{\ΛX}}_k^{ov}\right)}_{\mathrm{\Ω}}+{\left({\mathrm{Fi}}_k\right)}_{\mathrm{\Ω}}+{\left(N_k\right)}_{\mathrm{\Ω}}\\ ={\left(F\right)}_{\mathrm{\Ω}}{i}_k+{\left(N_k\right)}_{\mathrm{\Ω}}\end{array}---\left(12\right)$

In formula, Y _kthe frequency-region signal vector obtained after representing vector Fourier transform to received signal, represent a kth OFDM symbol frequency domain transmission signal phasor, Λ represents frequency domain channel transmission matrix, and F represents Discrete Fourier transform and is unitary matrice, N _krepresent Frequency domain noise signal phasor, i _krepresent the rangefinder signal phasor of channel input; Because this formula is a underdetermined equation, cannot directly by measured value (Y _k) _Ωcalculate rangefinder pulse signal i _k.

According to the rangefinder pulse signal reconstructing method based on convex optimization, document DonohoDL.CompressedSensing [J] .IEEETransactionsonInformationTheory, 2006,52 (4): 1289-1306, and document TroppJA.Justrelax:Convexprogrammingmethodsforidentifying sparsesignalsinnoise [J] .IEEETransactionsInformationTheory, 2006,52 (3): 1030-1051 give for standard compression perception observation signal model:

r＝Φc+v(25)

In formula, r is the observation vector of P × 1, and Φ is P × Q observing matrix (P≤Q), c is the sparse vector of Q × 1, and v is the unknown noise vector of P × 1.Document FuchsJJ.Recoveryofexactsparserepresentationsinthepresenc eofnoise [C] //IEEEInternationalConferenceonAcoustics, Speech, andSignalProcessing, Montreal, QC, the minimum 0-Norm Solution that Canada:IEEE, 2004,5:533-536 provide this underdetermined equation is:

$\underset{c}{\min }{\left|\right|c\left|\right|}_{0}s.t.{\left|\right|r-\mathrm{\Φc}\left|\right|}_2\≤\mathrm{\ϵ}---\left(26\right)$

Wherein, ε is non-negative error term, is determined by random noise || v|| ₂≤ ε, formula (26) is a NP-hard (nondeterministic polynomial time trouble) problem.For solving the minimum 0-norm Solve problems of underdetermined equation, document DonohoDL, EladM, TemlyakovVN.Stablerecoveryofsparseovercompleterepresenta tionsinthepresenceofnoise [J] .IEEETransactionsonInformationTheory, 2006,52 (1): 6-18 prove 0-norm minimum problem to be converted into 1-norm minimum problem:

$\begin{array}{ccc}\underset{c}{min}\left|\right|c|{|}_1& s.t.& \left|\right|r-\mathrm{\Φ}c|{|}_2\≤\mathrm{\ϵ}\end{array}---\left(27\right)$

According to document EwoutVD, MichaelPF.Probingtheparetofrontierforbasispursuitsolutio n [J] .SocietyforIndustrialAppliedMathematics, 2008,31 (2): 890-912 researchs, formula (27) can be expressed equivalently as:

$\underset{c}{min}\frac{1}{2}\left|\right|r-\mathrm{\Φ}c|{|}_2^2+\mathrm{\γ}|\left|c\right|{|}_1---\left(28\right)$

In formula, for Lagrange factor, formula (28) is lasso trick (LASSO) problem, and its last solution is:

$\tilde{c}=\arg \underset{c}{min}\frac{1}{2}\left|\right|r-\mathrm{\Φ}c|{|}_2^2+\mathrm{\γ}|\left|c\right|{|}_1---\left(29\right)$

With reference to the solution procedure of (25)-(29) formula, then formula (12) sparse signal i _kcomputational methods be expressed as:

${\hat{i}}_k=\arg \underset{i_k}{min}\frac{1}{2}\left|\right|{\left(Y_k\right)}_{\mathrm{\Ω}}-{\left(F\right)}_{\mathrm{\Ω}}{i}_k|{|}_2^2+\mathrm{\γ}|\left|i_k\right|{|}_1---\left(30\right)$

In simulation study, CVX tool box is adopted to try to achieve optimum sparse solution (BoydS.CVX:Matlabsoftwarefordisciplinedconvexprogramming [OL] .California:StanfordUniversity, 2012 [2014-09-05] .http: //cvx.com/cvx/).

So the present invention utilizes rangefinder pulse signal i _ktime domain sparse characteristic, use compression sensing method reconstruct rangefinder pulse signal, suppose that the kth rangefinder pulse interference signal that receiver reconstructs is designated as further by Fourier transform, Reconstruction Impulse signal is transformed into frequency domain then impulse disturbances elimination is carried out at frequency domain, the kth frequency-domain received signal vector be improved

$\begin{array}{c}{\tilde{Y}}_k=Y_k-F{\hat{i}}_k\\ ={\mathrm{\ΛX}}_k^{ov}+F\left(i_k-{\hat{i}}_k\right)+N_k\end{array}---\left(13\right)$

Suppose that receiver can Accurate Reconstruction rangefinder pulse signal, namely the kth frequency-domain received signal vector estimated value approximate representation then improved is:

${\tilde{Y}}_k\≈{\mathrm{\ΛX}}_k^{ov}+N_k---\left(14\right)$

Eliminate the signal phasor after interference frequency domain down-sampled signal after ideal low-pass filter

$\begin{array}{c}{\tilde{X}}_k\[n\]=\frac{1}{\sqrt{V}}{F}^{LP}\[n+N\left(V-1\right)/2\]\·{\tilde{Y}}_k\[n+N\left(V-1\right)/2\]\\ =\frac{1}{\sqrt{V}}{\tilde{X}}_k\[n+N\left(V-1\right)/2\],n=0,\mathrm{...},N-1;k=1,\mathrm{...},K\end{array}---\left(15\right)$

In formula, F ^lPthe frequency response that [n] is ideal low-pass filter:

$F^{LP}\[n\]=\left\{\begin{array}{cc}1,& n=N\left(V-1\right)/2,\mathrm{...},N\left(V-1\right)/2+N-1\\ 0,& otherwise\end{array}\right.---\left(16\right)$

the estimated value that subchannel data sends symbolic vector is extracted by demapping

The course of work of described empty decoder 29 and channel estimator 210 frequently for completing the empty process of decoding is frequently:

The frequency domain orthogonal frequency-division multiplex singal vector valuation that frequency domain down-sampler 28 is exported be modeled as:

${\tilde{Z}}_k=H_1\left(k\right)\·X_1\left(k\right)+H_2\left(k\right)\·X_2\left(k\right)+N\left(k\right)---\left(17\right)$

Wherein, X ₁(k) and X ₂k () represents the kth OFDM space-frequency coding vector that the first antenna (13a) and the second antenna (13b) send respectively; H ₁(k) and H ₂k () correspondence represents channel frequency domain response, N (k) represents the frequency domain white complex gaussian noise vector of channel input, dot product between ". " representative vector element; By the valuation of frequency domain orthogonal frequency-division multiplex singal vector be not difficult to draw by even item form vector expression

${\tilde{Z}}_{k,e}=H_{1,e}\left(k\right)X_{1,e}\left(k\right)+H_{2,e}\left(k\right)X_{2,e}\left(k\right)+N_e\left(k\right)---\left(18\right)$

With odd term vector expression

${\tilde{Z}}_{k,o}=H_{1,o}\left(k\right)X_{1,o}\left(k\right)+H_{2,o}\left(k\right)X_{2,o}\left(k\right)+N_o\left(k\right)---\left(19\right)$

Suppose that receiver is known or accurately can estimate channel frequency domain response H ₁(k) and H ₂k (), then send the valuation of OFDM symbol vector can be expressed as:

${\hat{X}}_e\left(k\right)=H_{1,e}^{*}\left(k\right){\tilde{Z}}_{k,e}+H_{2,o}\left(k\right){\tilde{Z}}_{k,o}^{*}---\left(20\right)$

${\hat{X}}_o\left(k\right)=H_{2,e}^{*}\left(k\right){\tilde{Z}}_{k,e}-H_{1,o}\left(k\right){\tilde{Z}}_{k,o}^{*}---\left(21\right)$

Wherein, with respectively by representing the data vector that middle even item and odd term are formed; The corresponding relation of (5) formula is brought into (18), (19) two formulas, replace the X in (18), (19) formula successively _{1, e}(k), X _{2, e}(k), X _{1, o}(k), X _{2, o}(k):

$\begin{array}{c}{\tilde{Z}}_{k,e}=H_{1,e}\left(k\right)X_e\left(k\right)+H_{2,e}\left(k\right)X_o\left(k\right)+N_e\left(k\right)\\ {\tilde{Z}}_{k,o}=-H_{1,o}\left(k\right)X_o^{*}\left(k\right)+H_{2,o}\left(k\right)X_e^{*}\left(k\right)+N_o\left(k\right)\end{array}---\left(22\right)$

(22) formula is brought into abbreviation after (20), (21) two formulas, and consider adjacent sub-carrier channel frequency domain response approximately equal, i.e. H _{1, o}(k) ≈ H _{1, e}(k), H _{2, o}(k) ≈ H _{2, e}(k):

$\begin{array}{c}{\hat{X}}_e\left(k\right)=\left({\left|H_{1,e}\left(k\right)\right|}^2+{\left|H_{2,e}\left(k\right)\right|}^2\right)X_e\left(k\right)+N_1\left(k\right)\\ {\hat{X}}_o\left(k\right)=\left({\left|H_{1,o}\left(k\right)\right|}^2+{\left|H_{2,o}\left(k\right)\right|}^2\right)X_o\left(k\right)+N_2\left(k\right)\end{array}---\left(23\right)$

Thus recover original transmission OFDM symbol vector X (n):

$\begin{array}{c}{X}_e\left(k\right)={\hat{X}}_e\left(k\right)/\left({\left|H_{1,e}\left(k\right)\right|}^2+{\left|H_{2,e}\left(k\right)\right|}^2\right)+{\stackrel{\‾}{N}}_1\left(k\right)\\ X_o\left(k\right)={\hat{X}}_o\left(k\right)/\left({\left|H_{1,o}\left(k\right)\right|}^2+{\left|H_{2,o}\left(k\right)\right|}^2\right)+{\stackrel{\‾}{N}}_2\left(k\right)\end{array}---\left(24\right)$

Wherein, N ₁(k), N ₂(k), be the equivalent noise vector produced in abbreviation process.

Before and after Fig. 3 a, Fig. 3 b display compressed sensing reconstruct rangefinder impulse disturbances, time domain waveform compares (single rangefinder interference, signal interference ratio=-10dB, noise power=0), in figure, abscissa represents the time, ordinate represents signal amplitude (real part), rangefinder impulse disturbances time domain signal waveform before the curve representative reconstruct of Fig. 3 a, the curve of Fig. 3 b represents the rangefinder impulse disturbances time domain signal waveform that compressed sensing reconstructs.Curve compares and shows: utilize the reconstruct rangefinder pulse interference signal that compression sensing method can be good.

The normalized mean squared error curve that after Fig. 4 display gives filtering, rangefinder pulse signal reconstructs, abscissa represents signal to noise ratio, the normalized mean squared error (NMSE) that after ordinate represents filtering, rangefinder pulse signal reconstructs.Comprise 3 curves in Fig. 4, wherein indicate "○", " ◇ ", " " curve represent NMSE curve when signal interference ratio is-3dB ,-7dB ,-10dB respectively.Curve compares and shows: (1), along with the increase of signal to noise ratio, after filtering, the normalized mean squared error of rangefinder pulse signal reconstruct linearly reduces, and illustrates that the increase of receiver signal to noise ratio contributes to the filtered rangefinder pulse interference signal of accurate reconstruction; (2) under signal to noise ratio same case, along with the increase of rangefinder pulse signal power after filtering, the precision of compressed sensing Reconstruction Impulse increases.

Fig. 5 a, Fig. 5 b, Fig. 5 c, Fig. 5 d display gives the comparison curves (orthogonal frequency-division multiplex singal power is 1, single rangefinder interference source, signal interference ratio=-10dB, and noise power is 0) of power spectrum signal before and after pulse interference suppression.The abscissa of each figure represents frequency (MHz), ordinate representation signal power (dBw).Wherein, Fig. 5 a shows the power spectrum (transmit signal power is 1) that orthogonal frequency division multiplex ransmitting penetrates signal, be can observe by Fig. 5 a: orthogonal frequency-division multiplex singal dominant frequency component is positioned at-250KHz ~+250KHz, in OFDM passband, power spectrum signal value is-30dBw.(rangefinder signal carrier is biased 500KHz to the power spectrum of Fig. 5 b display rangefinder pulse signal residual interference signal after equivalent frequency overlapped-resistable filter, signal interference ratio SIR=-10dB), be can observe by Fig. 5 b: after receiver frequency overlapped-resistable filter, the major frequency components of residual interference signal is positioned at about+250KHz, and residual interference signal still has stronger signal power.Fig. 5 c display provides the power spectrum of receiver Received signal strength, be can observe by Fig. 5 c: in-250KHz ~+150KHz frequency range, orthogonal frequency-division multiplex singal power spectrum value is-30dBw, and in+150KHz ~+250KHz frequency band range, obviously can observe the frequency component of residual pulse interference signal, and interference signal maximum intensity value is-10dBw.Fig. 5 d display provides the power spectrum of signal after pulse interference suppression, be can observe by Fig. 5 d: in the frequency range of-250KHz ~+250KHz, orthogonal frequency-division multiplex singal power spectrum value is-30dBw, and in 150KHz ~ 250KHz frequency range, do not observe rangefinder pulse interference signal component.Fig. 5-a ~ Fig. 5-d compares and shows: the reconstruct rangefinder interference signal that compressed sensing algorithm can be good, and is effectively eliminated.

Fig. 6 display exports Signal to Interference plus Noise Ratio change curve after giving suppressor pulse interference, and in figure, abscissa represents input signal-to-noise ratio, exports Signal to Interference plus Noise Ratio after ordinate representative suppresses interference.Wherein indicate "○", " ◇ ", " " curve represent signal interference ratio respectively for exporting Signal to Interference plus Noise Ratio curve when-3dB ,-7dB ,-10dB.Curve compares and shows: (1), along with the increase of input signal-to-noise ratio, exports dry the making an uproar of letter and increases gradually, illustrates that the increase of receiver signal to noise ratio exports Signal to Interference plus Noise Ratio after contributing to promoting AF panel; (2) under signal to noise ratio same case, along with the increase of rangefinder pulse signal power after filtering, suppress the output Signal to Interference plus Noise Ratio after interference to reduce gradually, when illustrating that interfering signal power is larger, interference power residual after suppressing is also larger.

Fig. 7 display provides the bit error performance curve of L frequency band digital aviation communication system 1 system, and in figure, abscissa represents signal to noise ratio, and ordinate represents bit error rate.Rangefinder interference signal carrier offset 500KHz is set, signal interference ratio SIR=-3db in emulation.The curve indicating " ■ " in figure represents the bit error performance curve of system when disturbing without rangefinder; Indicate systematic bits error performance curve when the representative of " ▲ " curve exists the interference of single rangefinder but do not eliminate; In the curve generation indicating "●", adopts empty coding and decoding and compressed sensing Reconstruction Impulse frequently disturb and eliminate rear systematic bits error performance curve; The curve representative indicating " ◆ " only adopts compressed sensing reconstruct to disturb and eliminates rear systematic bits error performance; The curve indicating " ∣ " represents that adopting ideal pulse to extinguish eliminates the rear systematic bits error performance curve of interference, and this curve as a comparison curve provides; Indicate the key technology means such as curve representative employing empty coding and decoding, chnnel coding, the compressed sensing frequently of " " and eliminate the rear systematic bits error performance of interference; The curve indicating " | " represents that adopting ideal pulse to extinguish eliminates the rear systematic bits error performance curve of interference; Note, the curve "-" not marking any legend in figure represents the system errors performance curve obtained after the key technology such as the employing orthogonal transform of patent proposition, symbol interleaving, empty coding and decoding, chnnel coding, compressed sensing frequently suppresses interference.Curve comparison shows: (1), when there is rangefinder pulse interference signal, will significantly worsen the reliability of system link transmission; 2) compare other AF panel schemes proposed in figure, the method that patent proposes can significantly improve the bit error performance of system, improves the reliability of L frequency band digital aviation communication system 1 system link transmission.

Claims (6)

1., based on a compressed sensing pulse interference suppression system for space-frequency coding, include based on the OFDM transmitter of space-frequency coding and the compressed sensing pulse interference suppression receiver based on sky decoding frequently, it is characterized in that:

The described OFDM transmitter based on space-frequency coding includes: the channel encoder (11) for docking collection of letters source signal coding be connected in series successively, for the source signal after coding being modulated into the modulator (12) of orthogonal phase modulation symbol, orthogonal phase modulation symbolic is carried out to the quadrature transformer (13) of orthogonal transform, the symbol of orthogonal transform is carried out to the symbol interleaver (14) of symbol interleaving, signal after symbol interleaving is converted to the deserializer (15) of parallel orthogonal phase modulation symbol and the space-frequency encoder (16) of generation two-way coded signal that parallel each OFDM symbol is encoded successively, described space-frequency encoder (16) connects the identical subcarrier mapper (17a/17b) for carrying out subcarrier mapping respectively to two-way coded signal of two structures respectively, the output of described each subcarrier mapper (17a/17b) be connected with in turn sub-carrier map after signal carry out the up-sampler (18a/18b) of 4 times of up-samplings, signal after 4 times of up-samplings is carried out to the inverse discrete Fourier transform device (19a/19b) of Fourier transform, for the formation of the Cyclic Prefix adder (10a/10b) of parallel complete OFDM symbol, parallel OFDM symbol is carried out to the parallel-to-serial converter (11a/11b) of parallel-serial conversion, digital-to-analogue conversion is carried out to the signal after parallel-serial conversion and the D/A conversion of launching and radiofrequency launcher (12a/12b) and transmitting antenna (13a/13b),

The described compressed sensing pulse interference suppression receiver based on sky decoding frequently includes and is connected in series successively: the reception antenna (21) of the signal that the OFDM transmitter for receiving based on space-frequency coding is launched, received signal is converted into successively radio-frequency front-end transducer and the A/D converter (22) of analog baseband signal and digital baseband signal, for compensating the Doppler shift compensator (24) of the Doppler shift of Received signal strength, for removing the cyclic prefix remover (25) of the Cyclic Prefix of orthogonal frequency-division multiplex singal, effective orthogonal frequency-division multiplex singal is carried out to the first discrete Fourier transform device (26) of Fourier transform, for reconstructing interference signal and eliminating the compressed sensing interference eliminator (27) of interference signal, signal after elimination interference is carried out to the frequency domain down-sampler (28) of frequency domain sample, for completing the empty decoder (29) frequently of the empty process of decoding frequently, for carrying out the symbol deinterleaver (211) of symbolic solution intertexture to signal, signal after deinterleaving is carried out to the inverse orthogonal transformation device (212) of inverse orthogonal transformation, for carrying out the demodulator of PM signal PM (213) of data demodulates and the channel decoder (214) for going out to launch bit sequence valuation to the signal recuperation after demodulation, wherein, described radio-frequency front-end transducer also will be connected Doppler shift compensator (24) by the Doppler shift estimator (23) for the Doppler shift of estimating received signal with the signal output part of A/D converter (22), the signal output part of described frequency domain down-sampler (28) also will connect empty decoder (29) frequently by the channel estimator (210) for estimating to be clipped to two transmitting antennas (13a/13b) dividing two channel frequency domain response of reception antenna (21).

2. a kind of compressed sensing pulse interference suppression system based on space-frequency coding according to claim 1, it is characterized in that, described radio-frequency front-end transducer and A/D converter (22) have employed four times of over-samplings to analog baseband signal in modulus sampling process.

3. a kind of compressed sensing pulse interference suppression system based on space-frequency coding according to claim 1, it is characterized in that, the described course of work based on the space-frequency encoder (16) in the OFDM transmitter of space-frequency coding is: first by the OFDM symbol vector representation of the kth after deserializer (15) serioparallel exchange be:

X(k)＝[X(k,0),X(k,1),...,X(k,l),...,X(k,L-1)] ^T(1)

Wherein, k=1,2 ..., K, l=0,1 ..., L-1, T represent transpose operation, and X (k, l) represents l subcarrier data in a kth OFDM symbol; Data vector X is obtained to after kth OFDM symbol vector X (k) coding ₁(k) and X ₂(k), wherein

X ₁k () is expressed as:

X ₁(k)＝[X(k,0),-X ^*(k,1),...,X(k,L-2),-X ^*(k,L-1)] ^T(2)

X ₂k () is expressed as:

X ₂(k)＝[X(k,1),X ^*(k,0),...,X(k,L-1),X ^*(n,L-2)] ^T(3)

Specific coding rule is wherein:

X ₁(k,i)＝X(k,i)

X ₁(k,i+1)＝-X ^*(k,i+1)

(4)

X ₂(k,i)＝X(k,i+1)

X ₂(k,i+1)＝X ^*(k,i)i＝0,2,4,6,...

Wherein, X _α(k, i), α=1,2 i-th subcarrier data representing a kth OFDM symbol in coded vector; Thus obtain corresponding relation:

X _1,e(k)＝X _e(k)(5-a)

$X_{1,o}\left(k\right)=-X_o^{*}\left(k\right)---(5-b)$

X _2,e(k)＝X _o(k)(5-c)

$X_{2,o}\left(k\right)=X_e^{*}\left(k\right)---(5-d).$

4. a kind of compressed sensing pulse interference suppression system based on space-frequency coding according to claim 1, it is characterized in that, described including based on the compressed sensing interference eliminator (27) in the compressed sensing pulse interference suppression receiver of sky decoding frequently disturbs reconstructor (271) and the second discrete Fourier transform device (272).

5. a kind of compressed sensing pulse interference suppression system based on space-frequency coding according to claim 1 or 4, it is characterized in that, the course of work of described compressed sensing interference eliminator (27) is:

Definition Ω represents frequency domain transmission signal phasor null symbol position number form set, () _Ωrepresent the subvector or submatrix that are made up of element corresponding to sequence number in set omega or row; For frequency-region signal vector Y _kthe signal extracting null symbol subchannel obtains:

$\begin{array}{c}{\left(Y_k\right)}_{\mathrm{\Ω}}={\left({\mathrm{\ΛX}}_k^{ov}\right)}_{\mathrm{\Ω}}+{\left({\mathrm{Fi}}_k\right)}_{\mathrm{\Ω}}+{\left(N_k\right)}_{\mathrm{\Ω}}\\ ={\left(F\right)}_{\mathrm{\Ω}}{i}_k+{\left(N_k\right)}_{\mathrm{\Ω}}\end{array}---\left(12\right)$

In formula, Y _kthe frequency-region signal vector obtained after representing vector Fourier transform to received signal, represent a kth OFDM symbol frequency domain transmission signal phasor, Λ represents frequency domain channel transmission matrix, and F represents Discrete Fourier transform and is unitary matrice, N _krepresent Frequency domain noise signal phasor, i _krepresent the rangefinder signal phasor of channel input;

Utilize rangefinder pulse signal i _ktime domain sparse characteristic, use compression sensing method reconstruct rangefinder pulse signal, suppose that the kth rangefinder pulse interference signal that receiver reconstructs is designated as further by Fourier transform, Reconstruction Impulse signal is transformed into frequency domain then impulse disturbances elimination is carried out at frequency domain, the kth frequency-domain received signal vector be improved

$\begin{array}{c}{\tilde{Y}}_k=Y_k-F{\hat{i}}_k\\ ={\mathrm{\ΛX}}_k^{ov}+F\left(i_k-{\hat{i}}_k\right)+N_k\end{array}---\left(13\right)$

Suppose that receiver can Accurate Reconstruction rangefinder pulse signal, namely the kth frequency-domain received signal vector estimated value approximate representation then improved is:

${\tilde{Y}}_k\≈{\mathrm{\ΛX}}_k^{ov}+N_k---\left(14\right).$

6. a kind of compressed sensing pulse interference suppression system based on space-frequency coding according to claim 1, it is characterized in that, the course of work of described empty decoder (29) and channel estimator (210) frequently for completing the empty process of decoding is frequently:

The frequency domain orthogonal frequency-division multiplex singal vector valuation that frequency domain down-sampler (28) is exported be modeled as:

${\tilde{Z}}_k=H_1\left(k\right)\·X_1\left(k\right)+H_2\left(k\right)\·X_2\left(k\right)+N\left(k\right)---\left(17\right)$

Wherein, X ₁(k) and X ₂k () represents the kth OFDM space-frequency coding vector that the first antenna (13a) and the second antenna (13b) send respectively; H ₁(k) and H ₂k () correspondence represents channel frequency domain response, N (k) represents the frequency domain white complex gaussian noise vector of channel input, dot product between ". " representative vector element; By the valuation of frequency domain orthogonal frequency-division multiplex singal vector be not difficult to draw by even item form vector expression

${\tilde{Z}}_{k,e}=H_{1,e}\left(k\right)X_{1,e}\left(k\right)+H_{2,e}\left(k\right)X_{2,e}\left(k\right)+N_e\left(k\right)---\left(18\right)$

With odd term vector expression

${\tilde{Z}}_{k,o}=H_{1,o}\left(k\right)X_{1,o}\left(k\right)+H_{2,o}\left(k\right)X_{2,o}\left(k\right)+N_o\left(k\right)---\left(19\right)$

Suppose that receiver is known or accurately can estimate channel frequency domain response H ₁(k) and H ₂k (), then send the valuation of OFDM symbol vector can be expressed as:

${\hat{X}}_e\left(k\right)=H_{1,e}^{*}\left(k\right){\tilde{Z}}_{k,e}+H_{2,o}\left(k\right){\tilde{Z}}_{k,o}^{*}---\left(20\right)$

${\hat{X}}_o\left(k\right)=H_{2,e}^{*}\left(k\right){\tilde{Z}}_{k,e}-H_{1,o}\left(k\right){\tilde{Z}}_{k,o}^{*}---\left(21\right)$

Wherein, with respectively by representing the data vector that middle even item and odd term are formed; By corresponding relation:

X _1,e(k)＝X _e(k)(5-a)

$X_{1,o}\left(k\right)=-X_o^{*}\left(k\right)---(5-b)$

X _2,e(k)＝X _o(k)(5-c)

$X_{2,o}\left(k\right)=X_e^{*}\left(k\right)---(5-d)$

Bring into with expression formula, replace the X in formula successively _{1, e}(k), X _{2, e}(k), X _{1, o}(k), X _{2, o}(k):

$\begin{array}{c}{\tilde{Z}}_{k,e}=H_{1,e}\left(k\right)X_e\left(k\right)+H_{2,e}\left(k\right)X_o\left(k\right)+N_e\left(k\right)\\ {\tilde{Z}}_{k,o}=-H_{1,o}\left(k\right)X_o^{*}\left(k\right)+H_{2,o}\left(k\right)X_e^{*}\left(k\right)+N_o\left(k\right)\end{array}---\left(22\right)$

And bring into with expression formula after abbreviation, and consider adjacent sub-carrier channel frequency domain response approximately equal, i.e. H _{1, o}(k) ≈ H _{1, e}(k), H _{2, o}(k) ≈ H _{2, e}(k):

$\begin{array}{c}{\hat{X}}_e\left(k\right)=\left({\left|H_{1,e}\left(k\right)\right|}^2+{\left|H_{2,e}\left(k\right)\right|}^2\right)X_e\left(k\right)+N_1\left(k\right)\\ {\hat{X}}_o\left(k\right)=\left({\left|H_{1,o}\left(k\right)\right|}^2+{\left|H_{2,o}\left(k\right)\right|}^2\right)X_o\left(k\right)+N_2\left(k\right)\end{array}---\left(23\right)$

Thus recover original transmission OFDM symbol vector X (n):

$\begin{array}{c}{X}_e\left(k\right)={\hat{X}}_e\left(k\right)/\left({\left|H_{1,e}\left(k\right)\right|}^2+{\left|H_{2,e}\left(k\right)\right|}^2\right)+{\stackrel{\‾}{N}}_1\left(k\right)\\ X_o\left(k\right)={\hat{X}}_o\left(k\right)/\left({\left|H_{1,o}\left(k\right)\right|}^2+{\left|H_{2,o}\left(k\right)\right|}^2\right)+{\stackrel{\‾}{N}}_2\left(k\right)\end{array}---\left(24\right)$

Wherein, N ₁(k), N ₂(k), be the equivalent noise vector produced in abbreviation process.