CN104954310A - Impulse noise eliminating method and device - Google Patents

Abstract

The invention provides an impulse noise eliminating method and device. The method comprises that S1) T transmitting antennas of an antenna system transmit frame headers and frame bodies of signal frames; S2) each reception antenna obtains the value of a reserved subcarrier in the frequency domain of the frame body to form the frequency-domain observation vector of impulse noise signals, and according to a multidimensional impulse noise combined sampling matrix composed of R frequency-domain observation vectors, a structural compressive sensing model is obtained; S3) estimation of time-domain impulse noise signals corresponding to the R reception antennas is obtained by utilizing the structural compressive sensing algorithm; and S4) the estimation of the time-domain impulse noise signals is subtracted from the corresponding time domain signals of the frame bodies of the R reception antennas to obtain frame body data of the signal frames of the R reception antennas of which impulse noise signals are eliminated. Thus, MIMO-OFDM impulse noises can be accurately estimated, and the spectral efficiency and robustness of the system are effectively improved.

Description

Impulse noise removing method and device

Technical field

The present invention relates to digital signal transmission technique field, particularly a kind of impulse noise removing method based on spatial coherence and structuring compressed sensing and device.

Background technology

In Broadband high-speed data transmission, there is the Complex Noise interference such as the interference of frequency selective fading, time selective fading and arrowband, impulse noise (Impulse Noise, IN), affect data transmission quality.In order to overcome above-mentioned severe channel conditions, OFDM (Orthogonal Frequency Division Multiplexing, OFDM) technology is widely used.OFDM technology mainly comprises time-domain synchronization OFDM technology (Time Domain Synchronous OFDM, TDS-OFDM), Cyclic Prefix orthogonal frequency (Cyclic-Prefix OFDM, and a few class such as zero padding OFDM (Zero-Padding OFDM, ZP-OFDM) CP-OFDM).Because OFDM has the good characteristic overcoming frequency selectivity very well, it has been applied in various digital signal transmission system, as electric line communication system standard (the International Telecommunications Union of International Telecommunication Association, ITU-T is G.9960), WLAN (wireless local area network) (WLAN), European Digital Video terrestrial broadcasting (Digital Video Broadcasting-Terrestrial, and Chinese terrestrial DTV transmission standard (Digital Television Multimedia Broadcast, DTMB) etc. DVB-T).The combination of multiple-input and multiple-output (Multiple-Input-Multiple-Output, MIMO) technology and OFDM technology, can further capacity, improve systematic function.

In convenient, the resourceful power line channel of transmission, carry out the high-efficiency digital communication based on OFDM technology, obtain investigation and application widely, but power line channel bad environments, especially there is serious impulse noise impact; Also there is the problems such as time domain impulse noise in ground system of digital television broadcast and wireless communication system.Impulse noise can reduce channel estimation accuracy, affects correct demapping and the decoding of data, has a strong impact on the correct transmission of data.But the method performance of traditional opposing impulse noise is not ideal enough, especially under the impulse noise of higher-strength, the transmission performance of data can severe exacerbation.The means such as traditional time domain intertexture, although can reduce the impact of impulse noise to a certain extent, fundamentally cannot eliminate impulse noise impact, poor effect under serious impulse noise condition; In addition, the method complexity of traditional anti-impulse noise is high, and needs the condition such as extra frequency spectrum resource and parametric statistics information, under complicated bad transmission condition, cause systematic function and decrease in efficiency.

Based on digital signal processing theory-compressed sensing (Compressive Sensing) emerging in recent years, the observation sequence far fewer than measured signal dimension can be utilized, by compressing perception algorithm based on convex optimization or greedy algorithm etc., Exact recovery has openness high dimensional signal.Compressed sensing algorithm obtains increasing concern in academia, is widely used in fields such as signal transacting, channel estimating, image compression.Because impulse noise has in time domain natural openness in essence, compressive sensing theory can be introduced and estimate.

External existing tradition is based on the impulse noise method of estimation of compressed sensing, utilize the special null subcarrier of frequency domain to form observation sequence to estimate, but there is certain defect in it: on the one hand, according to compressive sensing theory, the existence of ground noise on algorithm effect impact significantly, when stronger additive white Gaussian noise (AWGN), the method has to take the special null subcarrier of larger amt as observation sequence, otherwise cannot accurately estimate dryly to make an uproar than (Interference-to-Noise Ratio, INR) lower impulse noise signal, loss of spectral efficiency will be caused serious.And preserved sub-carrier quantity in existing power line communication standard and radio standard is very restricted, in a lot of application scenarios, impulse noise is made an uproar than not high relative to the dry of Gaussian noise, and the method not only performance will be had a strong impact on; On the other hand, compressed sensing algorithm iterative computation from initial condition of the method, the number of times of iteration is more, and computation complexity needs reduction badly.Further, at present still not for the spatial coherence that the impulse noise of mimo system has, the feature that the time-domain position of the impulse noise signal namely on multiple receive antenna is identical, carries out research that structuring multivariate joint probability impulse noise estimates or technology.

To sum up, traditional anti-impact swashs the problem that loss of spectral efficiency is serious, performance is low, computation complexity is high that Noise Method and the existing impulse noise method of estimation based on compressed sensing exist.

Summary of the invention

The present invention is intended at least one of solve the problems of the technologies described above.

For this reason, one object of the present invention is to propose a kind of impulse noise removing method.The method accurately can estimate the impulse noise of multi-I/O OFDM (MIMO-OFDM), the spectrum efficiency of effective elevator system and robustness.

Another object of the present invention is to propose a kind of impulse noise cancellation element.

To achieve these goals, the embodiment of a first aspect of the present invention discloses a kind of impulse noise removing method, comprise the following steps: S1: send signal frame frame head and signal frame frame respectively by T transmitting antenna of antenna system, wherein, described antenna system comprises T transmitting antenna and R reception antenna; S2: each reception antenna obtains the value on the frequency domain preserved sub-carrier of described signal frame frame, form the domain observations vector of impulse noise signal, and form multidimensional impulse noise combined sampling matrix according to R domain observations vector of a described R reception antenna, and based on the time-frequency domain Fourier transform relation of impulse noise signal, obtain structuring compressed sensing model; S3: according to described structuring compressed sensing model, adopt structure based compressed sensing algorithm to estimate time-domain position and the strength factor of the impulse noise signal of a described R reception antenna, obtain the estimation of the time domain impulse noise signal corresponding to a described R reception antenna; S4: the estimation time-domain signal of the signal frame frame of a described R reception antenna being deducted corresponding time domain impulse noise signal, R reception antenna signal frame body data after the impulse noise signal that is eliminated.

In addition, impulse noise removing method according to the above embodiment of the present invention can also have following additional technical characteristic:

In some instances, described signal frame frame is OFDM data block, and the frequency domain of described signal frame frame comprises multiple preserved sub-carrier, and described signal frame frame head is made up of the Cyclic Prefix of described signal frame frame or signal frame frame head training sequence.

In some instances, described step S2 also comprises: the time-domain signal according to the signal frame frame of current reception carries out rough estimate to the time-domain position of impulse noise signal, specifically comprise: obtain the time domain sequences absolute value of the signal frame frame of a described R reception antenna square, and square all to compare R described time domain sequences absolute value with pre-determined threshold, wherein, described pre-determined threshold is the multiple of the mean value of the quadratic sum of the time domain sequences absolute value of the signal frame frame of a described R reception antenna; Using the rough estimate of position as the time-domain position of described impulse noise signal square being all greater than described pre-determined threshold of absolute value in R time domain sequences of a described R reception antenna.

In some instances, described step S2, comprise further: S21: when described preserved sub-carrier is pilot sub-carrier, calculate the value on the pilot sub-carrier of the described signal frame frame frequency domain of current reception, and the known transmission pilot value value on described pilot sub-carrier deducted in respective frequencies is multiplied by the estimated value gained sequence of channel frequency response, obtain the domain observations sequence of described impulse noise signal, when described preserved sub-carrier is vacant subcarrier, the domain observations sequence of described impulse noise signal is that the value on the vacant subcarrier of frequency domain forms domain observations vector successively, S22: R the domain observations vector according to a described R reception antenna forms multidimensional impulse noise combined sampling matrix, S23: based on the time-frequency domain Fourier transform relation of impulse noise signal, obtains described structuring compressed sensing model.

In some instances, in described step S2, described structuring compressed sensing model is multidimensional impulse noise signal relationship between frequency and time equation.

In some instances, when described signal frame frame head is made up of described signal frame frame head training sequence, the acquisition pattern of the value on the pilot sub-carrier of described signal frame frame frequency domain or the value on vacant subcarrier is: carry out loop restructuring and training sequence interference elimination to the time domain sequences of the described signal frame frame of current reception; When described signal frame frame head is made up of the Cyclic Prefix of described signal frame frame, the acquisition pattern of the value on the pilot sub-carrier of described signal frame frame frequency domain or the value on vacant subcarrier is: directly carry out discrete Fourier transform to the time domain sequences of the described signal frame frame of current reception.

In some instances, in described step S3, described structuring compressed sensing algorithm is based on the convex optimized algorithm of multivariate joint probability sparse signal or the greedy algorithm based on multidimensional structure compressed sensing, and wherein, described convex optimized algorithm comprises interior point method, single order norm minimum algorithm; Described greedy algorithm comprises structuring synchronous orthogonal matching pursuit algorithm, structural sparse Adaptive matching back tracking method.

In some instances, in described step S3, adopt the rough estimate of the time-domain position of described impulse noise signal, as the prior information of structuring compressed sensing algorithm, carry out the auxiliary structuring compressed sensing iteration of prior information and impulse noise estimation.

In some instances, after described step S3, also comprise: the estimated accuracy optimizing the coefficient of described impulse noise signal, specifically comprise: to each reception antenna, on the time-domain position of the described impulse noise signal estimated, be multiplied by residuals squares principle between impulse noise time-domain signal to be estimated and described impulse noise domain observations sequence to minimize described Fourier transform matrix, carry out least-squares estimation.

The embodiment of second aspect present invention discloses a kind of impulse noise cancellation element, comprise: compressed sensing model generation module, the value on the frequency domain preserved sub-carrier of described signal frame frame is received for obtaining R reception antenna, form the domain observations vector of impulse noise signal, and form multidimensional impulse noise combined sampling matrix according to R domain observations vector of a described R reception antenna, and based on the time-frequency domain Fourier transform relation of impulse noise signal, obtain structuring compressed sensing model; Compressed sensing estimation module, for according to described structuring compressed sensing model, adopt structure based compressed sensing algorithm to estimate time-domain position and the strength factor of the impulse noise signal of a described R reception antenna, obtain the estimation of the time domain impulse noise signal corresponding to a described R reception antenna; Impulse noise cancellation module, the time-domain signal for the signal frame frame by a described R reception antenna deducts the estimation of corresponding time domain impulse noise signal, R reception antenna signal frame body data after the impulse noise signal that is eliminated.

According to impulse noise removing method and the device of the embodiment of the present invention, can in severe complicated transmission channel, with relatively high spectrum efficiency, estimate accurately and fast and eliminate impulse noise based on the spatial coherence between multiple antennas and structuring compressed sensing mode, the spectrum efficiency of elevator system and transmission robustness.

Additional aspect of the present invention and advantage will part provide in the following description, and part will become obvious from the following description, or be recognized by practice of the present invention.

Accompanying drawing explanation

Above-mentioned and/or additional aspect of the present invention and advantage will become obvious and easy understand from accompanying drawing below combining to the description of embodiment, wherein:

Fig. 1 is the flow chart of impulse noise removing method according to an embodiment of the invention;

Fig. 2 is that in embodiments of the invention 1, CP-OFDM system adopts the impulse noise of pilot sub-carrier observation sequence to estimate and eliminates time frequency processing schematic diagram;

Fig. 3 is that in embodiments of the invention 2, TDS-OFDM system adopts the impulse noise of null subcarrier observation sequence to estimate and eliminates time frequency processing schematic diagram; And

Fig. 4 is the structured flowchart of impulse noise cancellation element according to an embodiment of the invention.

Embodiment

Be described below in detail embodiments of the invention, the example of described embodiment is shown in the drawings, and wherein same or similar label represents same or similar element or has element that is identical or similar functions from start to finish.Being exemplary below by the embodiment be described with reference to the drawings, only for explaining the present invention, and can not limitation of the present invention being interpreted as.

In describing the invention, it will be appreciated that, term " " center ", " longitudinal direction ", " transverse direction ", " on ", D score, " front ", " afterwards ", " left side ", " right side ", " vertically ", " level ", " top ", " end ", " interior ", orientation or the position relationship of the instruction such as " outward " are based on orientation shown in the drawings or position relationship, only the present invention for convenience of description and simplified characterization, instead of indicate or imply that the device of indication or element must have specific orientation, with specific azimuth configuration and operation, therefore limitation of the present invention can not be interpreted as.In addition, term " first ", " second " only for describing object, and can not be interpreted as instruction or hint relative importance.

In describing the invention, it should be noted that, unless otherwise clearly defined and limited, term " installation ", " being connected ", " connection " should be interpreted broadly, and such as, can be fixedly connected with, also can be removably connect, or connect integratedly; Can be mechanical connection, also can be electrical connection; Can be directly be connected, also indirectly can be connected by intermediary, can be the connection of two element internals.For the ordinary skill in the art, concrete condition above-mentioned term concrete meaning in the present invention can be understood.

Below in conjunction with accompanying drawing description according to the impulse noise removing method of the embodiment of the present invention and device.

Fig. 1 is the flow chart of impulse noise removing method according to an embodiment of the invention.As shown in Figure 1, impulse noise removing method according to an embodiment of the invention, comprises the following steps:

S1: send signal frame frame head and signal frame frame respectively by T transmitting antenna of antenna system, wherein, antenna system comprises T transmitting antenna and R reception antenna, that is: for the multi-input multi-output antenna system containing T transmitting antenna, a R reception antenna, T transmitting antenna sends signal frame frame head and signal frame frame respectively.

S2: each reception antenna obtains the value on the frequency domain preserved sub-carrier of signal frame frame, form the domain observations vector of impulse noise signal, and form multidimensional impulse noise combined sampling matrix according to R domain observations vector of R reception antenna, and based on the time-frequency domain Fourier transform relation of impulse noise signal, obtain structuring compressed sensing model.

S3: according to structuring compressed sensing model, adopts structure based compressed sensing algorithm to estimate time-domain position and the strength factor of the impulse noise signal of R reception antenna, obtains the estimation of the time domain impulse noise signal corresponding to R reception antenna.

S4: the estimation time-domain signal of the signal frame frame of R reception antenna being deducted corresponding time domain impulse noise signal, R reception antenna signal frame body data after the impulse noise signal that is eliminated.

Wherein, signal frame frame is OFDM data block, and the frequency domain of described signal frame frame comprises multiple preserved sub-carrier, and described signal frame frame head is made up of the Cyclic Prefix of described signal frame frame or signal frame frame head training sequence.

In one embodiment of the invention, step S2 also comprises: the time-domain signal according to the signal frame frame of current reception carries out rough estimate to the time-domain position of impulse noise signal, specifically comprise: obtain the time domain sequences absolute value of the signal frame frame of a described R reception antenna square, and square all to compare R described time domain sequences absolute value with pre-determined threshold, wherein, described pre-determined threshold is the multiple of the mean value of the quadratic sum of the time domain sequences absolute value of the signal frame frame of a described R reception antenna; Using the rough estimate of position as the time-domain position of described impulse noise signal square being all greater than described pre-determined threshold of absolute value in R time domain sequences of a described R reception antenna.

In one embodiment of the invention, step S2 comprises further:

S21: when described preserved sub-carrier is pilot sub-carrier, calculate the value on the pilot sub-carrier of the described signal frame frame frequency domain of current reception, and the known transmission pilot value value on described pilot sub-carrier deducted in respective frequencies is multiplied by the estimated value gained sequence of channel frequency response, obtain the domain observations sequence of described impulse noise signal, when described preserved sub-carrier is vacant subcarrier, the domain observations sequence of described impulse noise signal is that the value on the vacant subcarrier of frequency domain forms domain observations vector successively.

S22: R the domain observations vector according to a described R reception antenna forms multidimensional impulse noise combined sampling matrix.

S23: based on the time-frequency domain Fourier transform relation of impulse noise signal, obtains described structuring compressed sensing model.

In step s 2, structuring compressed sensing model is multidimensional impulse noise signal relationship between frequency and time equation, that is: multidimensional impulse noise combined signal sampling matrix equals partial Fourier inverse-transform matrix and is multiplied by impulse noise combined signal matrix to be estimated and adds frequency domain ground noise signal matrix.

In one embodiment of the invention, when described signal frame frame head is made up of described signal frame frame head training sequence, the acquisition pattern of the value on the pilot sub-carrier of described signal frame frame frequency domain or the value on vacant subcarrier is: carry out loop restructuring and training sequence interference elimination to the time domain sequences of the described signal frame frame of current reception; When described signal frame frame head is made up of the Cyclic Prefix of described signal frame frame, the acquisition pattern of the value on the pilot sub-carrier of described signal frame frame frequency domain or the value on vacant subcarrier is: directly carry out discrete Fourier transform to the time domain sequences of the described signal frame frame of current reception.

In one embodiment of the invention, in step s3, described structuring compressed sensing algorithm is based on the convex optimized algorithm of multivariate joint probability sparse signal or the greedy algorithm based on multidimensional structure compressed sensing, and wherein, described convex optimized algorithm comprises interior point method, single order norm minimum algorithm; Described greedy algorithm comprises structuring synchronous orthogonal matching pursuit algorithm, structural sparse Adaptive matching back tracking method.

Further, in step s3, adopt the rough estimate of the time-domain position of described impulse noise signal, as the prior information of structuring compressed sensing algorithm, carry out the auxiliary structuring compressed sensing iteration of prior information and impulse noise estimation.

Further, after step s 3, also comprise: the estimated accuracy optimizing the coefficient of impulse noise signal, specifically comprise: to each reception antenna, on the time-domain position of the described impulse noise signal estimated, be multiplied by residuals squares principle between impulse noise time-domain signal to be estimated and described impulse noise domain observations sequence to minimize described Fourier transform matrix, carry out least-squares estimation.

[embodiment 1]

As shown in Figure 2, and composition graphs 1, for CP-OFDM in the present embodiment (pilot tone, priori is assisted) system adopts the impulse noise based on spatial coherence and structuring compressed sensing of pilot sub-carrier observation sequence to estimate and removing method time frequency processing schematic diagram.

S1: send signal frame frame head and signal frame frame respectively by T transmitting antenna of antenna system, wherein, described antenna system comprises T transmitting antenna and R reception antenna, that is, for the multi-input multi-output antenna system containing T transmitting antenna, a R reception antenna in the embodiment of the present invention, send signal frame frame head and signal frame frame respectively by a described T transmitting antenna.

In the present embodiment, the signal frame frame that each transmitting antenna sends is OFDM data block, the frequency domain of signal frame frame comprises multiple preserved sub-carrier, described signal frame frame head is made up of the Cyclic Prefix of described signal frame frame or signal frame frame head training sequence, OFDM sub-carrier number is N=4096, frame head is the Cyclic Prefix of frame OFDM data block, and the signal frame frame OFDM data block that reception antenna p receives is denoted as:

${\left\{y^{\left(p\right)}\left(n\right)\right\}}_{n=1}^{4096},$

Preserved sub-carrier is pilot sub-carrier, and the value on described pilot sub-carrier is known pilot value, and the frequency domain value of the orthogonal guide frequency subcarrier of p transmitting antenna transmission is denoted as:

${\left\{P^{\left(p\right)}\right[i\left]\right\}}_{i=1}^M,$

Wherein M=128 is pilot sub-carrier quantity, and its position is the optional position of OFDM subcarrier or is determined by corresponding communication system standard, and location sets is:

Ω ₀＝{k _i|1≤k _i≤4096,i＝1,2…,M}，

The frequency-region signal of p the corresponding pilot frequency locations of reception antenna is denoted as:

${\left\{Y^{\left(p\right)}\right[k_i\left]\right\}}_{i=1}^M.$

S2: each reception antenna obtains the value on the frequency domain preserved sub-carrier of described signal frame frame, form the domain observations vector of impulse noise signal, and form multidimensional impulse noise combined sampling matrix according to R domain observations vector of a described R reception antenna, and based on the time-frequency domain Fourier transform relation of impulse noise signal, obtain structuring compressed sensing model, namely each reception antenna value obtained on the frequency domain preserved sub-carrier of described signal frame frame forms the domain observations vector of described impulse noise signal, R domain observations vector of a described R reception antenna forms multidimensional impulse noise combined sampling matrix, based on the time-frequency domain Fourier transform relation of impulse noise signal, obtain structuring compressed sensing model.

With the time-domain signal of the signal frame frame of current reception, rough estimate is carried out to the time-domain position of impulse noise signal, obtains the prior information of impulse noise signal time-domain position, be specially: the signal frame frame time domain sequences received by R reception antenna, is designated as:

${\left\{y^{\left(p\right)}\left(n\right)\right\}}_{n=1}^{4096},p=1,...R,$

The signal frame frame time domain sequences that R reception antenna receives is taken absolute value square, the absolute value of R reception antenna square all exceeded pre-determined threshold ε _thposition as the prior information of described time domain impulse interference moment position, obtain the rough estimate location sets Γ of impulse noise time-domain position ₀, then have:

${\mathrm{\Γ}}_0=\left\{n\right|{\left|y^{\left(p\right)}\left(n\right)\right|}^2>{\mathrm{\ϵ}}_{\mathrm{th}},p=\mathrm{1,2},...R\},$

Pre-determined threshold is chosen for the constant times of the mean value of current Received Signal time domain sequences squared absolute value, and as preferably, constant is 3, that is:

${\mathrm{\ϵ}}_{\mathrm{th}}=3\·\frac{1}{4096R}\underset{p=1}{\overset{R}{\mathrm{\Σ}}}\underset{n=1}{\overset{4096}{\mathrm{\Σ}}}{\left|y^{\left(p\right)}\left(n\right)\right|}^2.$

In the present embodiment, preserved sub-carrier is pilot sub-carrier, the acquisition pattern of the value on p reception antenna pilot sub-carrier for directly to received signal frame frame time domain sequences do discrete Fourier transform and obtain, that is:

${\left\{Y^{\left(p\right)}\right[k_i\left]\right\}}_{i=1}^M,$

Wherein k _i∈ Ω ₀for preserved sub-carrier location sets; T antenna of transmitting terminal sends orthogonal pilot frequency sequence, and impulse noise domain observations sequence, is designated as:

${\left\{S^{\left(p\right)}\right[i\left]\right\}}_{i=1}^M,$

The acquisition pattern of impulse noise domain observations sequence be current Received Signal frame frame frequency domain pilot sub-carrier on the value known transmission pilot value deducted in respective frequencies be multiplied by the estimated value gained sequence of channel frequency response, that is:

S ^(p)[i]＝Y ^(p)[k _i]-P ^(p)[i]·H ^(p)[k _i]，i＝1,2,…M，

Wherein, for corresponding to p the estimated value sent out, receive the channel frequency response in antenna pilot frequency.

In the present embodiment, structuring compressed sensing algorithm model is multidimensional impulse noise signal relationship between frequency and time equation, namely described multidimensional impulse noise combined signal sampling matrix equals partial Fourier inverse-transform matrix and is multiplied by impulse noise combined signal matrix to be estimated and adds frequency domain ground noise signal matrix, that is:

$\tilde{S}=F_{M}E+\tilde{W},$

Wherein, for multidimensional impulse noise combined signal sampling matrix; Matrix F _mfor the Fourier transform matrix that M × N is corresponding, its kth _ithe element of row, the n-th row is:

${\left\{F_M\right\}}_{k_i,n}=\exp (-j\frac{2\mathrm{\π}}{N}(k_i-1)(n-1)),i=\mathrm{1,2},...,M,n=\mathrm{1,2},...N;$

$E=({\stackrel{\→}{e}}^{\left(1\right)},{\stackrel{\→}{e}}^{\left(2\right)},...,{\stackrel{\→}{e}}^{\left(R\right)})$ For impulse noise combined signal matrix to be estimated; $\tilde{W}=({\tilde{w}}^{\left(1\right)},{\tilde{w}}^{\left(2\right)},...{\tilde{w}}^{\left(R\right)})$ For frequency domain ground noise signal matrix, be additive white Gaussian noise (AWGN) in the present embodiment;

According to multidimensional impulse noise signal relationship between frequency and time equation adopt based on the auxiliary convex optimized algorithm of impulse noise signal time-domain position prior information, as the quick greedy algorithm (OMP algorithm, CoSaMP algorithm or SAMP algorithm etc.) utilizing impulse noise signal time-domain position prior information auxiliary, time domain impulse noise Signal estimation can be obtained that is:

${\hat{e}}^{\left(p\right)}={[{\hat{e}}^{\left(p\right)}\left(1\right),...{\hat{e}}^{\left(p\right)}\left(N\right)]}^T;$

Because time domain impulse noise signal has stronger openness, estimate gained by compressed sensing element only have a few locations non-zero, all the other positions are zero.? the set of nonzero element position be denoted as then have:

${\hat{e}}^{\left(p\right)}\left(n\right)=\left\{\begin{array}{c}0,n\∉\mathrm{\Γ}\left({\hat{e}}^{\left(p\right)}\right)\\ {\hat{e}}_n^{\left(p\right)},n\∈\mathrm{\Γ}\left({\hat{e}}^{\left(p\right)}\right)\end{array}\right..$

To promote estimated accuracy further, after obtaining described time domain impulse noise Signal estimation, can in the position of the time domain impulse noise Signal estimation of gained place, residuals squares principle between impulse noise time-domain signal to be estimated and described impulse noise domain observations sequence is multiplied by further to minimize described Fourier transform matrix, carry out least-squares estimation, to promote the estimated accuracy of the coefficient of time domain impulse noise signal, namely solve least square problem:

$\underset{{\hat{e}}^{\left(p\right)}}{\min }{\left|\right|{\tilde{s}}^{\left(p\right)}-F_M{\hat{e}}^{\left(p\right)}\left|\right|}_2,$ Wherein ${\hat{e}}^{\left(p\right)}\left(n\right)=\left\{\begin{array}{c}0,n\∉\mathrm{\Γ}\left({\hat{e}}^{\left(p\right)}\right)\\ {\hat{e}}_n^{\left(p\right)},n\∈\mathrm{\Γ}\left({\hat{e}}^{\left(p\right)}\right)\end{array}\right.;$

The time-domain signal time-domain signal of signal frame frame being deducted impulse noise signal is estimated, the signal frame body data after the impulse noise signal that is eliminated;

By the time-domain signal frame body data block that antenna p receives deduct described time domain impulse noise Signal estimation be eliminated the body data after impulse noise that is:

$x^{\left(p\right)}\left(n\right)=y^{\left(p\right)}\left(n\right)-{\hat{e}}^{\left(p\right)}\left(n\right),n=\mathrm{1,2},...N.$

[embodiment 2]

As shown in Figure 3, and composition graphs 1, for TDS-OFDM in the present embodiment (null subcarrier, without priori) system adopts the impulse noise based on spatial coherence and structuring compressed sensing of null subcarrier observation sequence estimate and eliminate time frequency processing schematic diagram.

S1: send signal frame frame head and signal frame frame respectively by T transmitting antenna of antenna system, wherein, described antenna system comprises T transmitting antenna and R reception antenna, that is, for the multi-input multi-output antenna system containing T transmitting antenna, a R reception antenna in the embodiment of the present invention, send signal frame frame head and signal frame frame respectively by a described T transmitting antenna.

In the present embodiment, it is OFDM data block that each transmitting antenna sends signal frame frame, and OFDM sub-carrier number is N=3780, and the Received signal strength frame frame OFDM data block that antenna p receives is denoted as:

${\left\{y^{\left(p\right)}\left(n\right)\right\}}_{n=1}^{3780},$

Frame head training sequence is the inverse discrete Fourier transform of one section of time domain binary pseudo-random or frequency domain binary pseudo-random; Preserved sub-carrier is null subcarrier, and the value on described null subcarrier is zero, null subcarrier quantity M=64, and its position is the optional position of OFDM subcarrier or is determined by corresponding communication system standard, and location sets is Ω ₀={ k _i| 1≤k _i≤ 3780, i=1,2 ..., M}; The frequency-domain received signal of the corresponding null subcarrier position of receiving terminal is denoted as:

${\left\{Y^{\left(p\right)}\right[k_i\left]\right\}}_{i=1}^M.$

S2: each reception antenna obtains the value on the frequency domain preserved sub-carrier of described signal frame frame, form the domain observations vector of impulse noise signal, and form multidimensional impulse noise combined sampling matrix according to R domain observations vector of a described R reception antenna, and based on the time-frequency domain Fourier transform relation of impulse noise signal, obtain structuring compressed sensing model.

In the present embodiment, preserved sub-carrier is null subcarrier, the acquisition pattern of the value on p reception antenna null subcarrier is carry out loop restructuring and training sequence interference elimination to the signal frame frame time domain sequences of current reception, namely the convolution of frame end and channel impulse response is added in current Received Signal frame frame front end, and deduct the convolution of next frame frame head training sequence and channel impulse response, obtain the circular convolution of signal frame frame and channel impulse response, and do Fourier transform and obtain, the value on the null subcarrier obtained is wherein k _i∈ Ω ₀for preserved sub-carrier location sets; Impulse noise domain observations sequence for the sequence that the value on the vacant subcarrier of frequency domain is formed successively, i.e. S ^(p)[i]=Y ^(p)[k _i], i=1,2 ... M;

In the present embodiment, structuring compressed sensing algorithm model is multidimensional impulse noise signal relationship between frequency and time equation, namely described multidimensional impulse noise combined signal sampling matrix equals partial Fourier inverse-transform matrix and is multiplied by impulse noise combined signal matrix to be estimated and adds frequency domain ground noise signal matrix, that is:

$\tilde{S}=F_{M}E+\tilde{W},$

Wherein, for multidimensional impulse noise combined signal sampling matrix; Matrix F _mfor the Fourier transform matrix that M × N is corresponding, its kth _ithe element of row, the n-th row is:

${\left\{F_M\right\}}_{k_i,n}=\exp (-j\frac{2\mathrm{\π}}{N}(k_i-1)(n-1)),i=\mathrm{1,2},...,M,n=\mathrm{1,2},...N;$

$E=({\stackrel{\→}{e}}^{\left(1\right)},{\stackrel{\→}{e}}^{\left(2\right)},...,{\stackrel{\→}{e}}^{\left(R\right)})$ For impulse noise combined signal matrix to be estimated; $\tilde{W}=({\tilde{w}}^{\left(1\right)},{\tilde{w}}^{\left(2\right)},...{\tilde{w}}^{\left(R\right)})$ For frequency domain ground noise signal matrix, be additive white Gaussian noise (AWGN) in the present embodiment;

According to multidimensional impulse noise signal relationship between frequency and time equation adopt based on the auxiliary convex optimized algorithm of the time-domain position prior information of impulse noise signal, as the quick greedy algorithm (OMP algorithm, CoSaMP algorithm or SAMP algorithm etc.) utilizing impulse noise signal time-domain position prior information auxiliary, time domain impulse noise Signal estimation can be obtained namely because time domain impulse noise signal has stronger openness, estimate gained by compressed sensing element only have a few locations non-zero, all the other positions are zero.? the set of nonzero element position be denoted as

then have: ${\hat{e}}^{\left(p\right)}\left(n\right)=\left\{\begin{array}{c}0,n\∉\mathrm{\Γ}\left({\hat{e}}^{\left(p\right)}\right)\\ {\hat{e}}_n^{\left(p\right)},n\∈\mathrm{\Γ}\left({\hat{e}}^{\left(p\right)}\right)\end{array}\right..$

To promote estimated accuracy further, after the time-domain signal obtaining described impulse noise signal is estimated, the position can estimated at the time-domain signal of the time domain impulse noise of gained place, residuals squares principle between impulse noise time-domain signal to be estimated and described impulse noise domain observations sequence is multiplied by further to minimize described Fourier transform matrix, carry out least-squares estimation, to promote the estimated accuracy of the coefficient of time domain impulse noise signal, namely solve least square problem:

$\underset{{\hat{e}}^{\left(p\right)}}{\min }{\left|\right|{\tilde{s}}^{\left(p\right)}-F_M{\hat{e}}^{\left(p\right)}\left|\right|}_2,$ Wherein ${\hat{e}}^{\left(p\right)}\left(n\right)=\left\{\begin{array}{c}0,n\∉\mathrm{\Γ}\left({\hat{e}}^{\left(p\right)}\right)\\ {\hat{e}}_n^{\left(p\right)},n\∈\mathrm{\Γ}\left({\hat{e}}^{\left(p\right)}\right)\end{array}\right.;$

The time-domain signal time-domain signal of signal frame frame being deducted impulse noise signal is estimated, the signal frame body data after the impulse noise signal that is eliminated;

By the time-domain signal frame body data block that antenna p receives deduct described time domain impulse noise Signal estimation be eliminated the body data after impulse noise namely

$x^{\left(p\right)}\left(n\right)=y^{\left(p\right)}\left(n\right)-{\hat{e}}^{\left(p\right)}\left(n\right),n=\mathrm{1,2},...N.$

According to the impulse noise removing method of the embodiment of the present invention, can in the complicated multipath channel that impulse noise is severe, with higher spectrum efficiency, accurately estimate and eliminate impulse noise fast, elevator system transmission robustness, is applicable to the dry application scenarios larger than dynamic range of making an uproar in various ofdm system.

Embodiments of the invention disclose a kind of impulse noise cancellation element, as shown in Figure 4, according to the impulse noise cancellation element 400 of the embodiment of the present invention, comprising: compressed sensing model generation module 410, compressed sensing estimation module 420 and impulse noise cancellation module 430.

Wherein, compressed sensing model generation module 410, the value on the frequency domain preserved sub-carrier of described signal frame frame is received for obtaining R reception antenna, form the domain observations vector of impulse noise signal, and form multidimensional impulse noise combined sampling matrix according to R domain observations vector of a described R reception antenna, and based on the time-frequency domain Fourier transform relation of impulse noise signal, obtain structuring compressed sensing model.

Compressed sensing estimation module 420, for according to described structuring compressed sensing model, adopt structure based compressed sensing algorithm to estimate time-domain position and the strength factor of the impulse noise signal of a described R reception antenna, obtain the estimation of the time domain impulse noise signal corresponding to a described R reception antenna.

Impulse noise cancellation module 430, the time-domain signal for the signal frame frame by a described R reception antenna deducts the estimation of corresponding time domain impulse noise signal, R reception antenna signal frame body data after the impulse noise signal that is eliminated.

According to the impulse noise cancellation element of the embodiment of the present invention, can in the complicated multipath channel that impulse noise is severe, with higher spectrum efficiency, accurately estimate and eliminate impulse noise fast, elevator system transmission robustness, is applicable to the dry application scenarios larger than dynamic range of making an uproar in various ofdm system.

It should be noted that, the specific implementation of the specific implementation of the impulse noise cancellation element of the embodiment of the present invention and the impulse noise removing method of the embodiment of the present invention is similar, specifically refers to the description of method part, in order to reduce redundancy, does not repeat.

In the description of this specification, specific features, structure, material or feature that the description of reference term " embodiment ", " some embodiments ", " example ", " concrete example " or " some examples " etc. means to describe in conjunction with this embodiment or example are contained at least one embodiment of the present invention or example.In this manual, identical embodiment or example are not necessarily referred to the schematic representation of above-mentioned term.And the specific features of description, structure, material or feature can combine in an appropriate manner in any one or more embodiment or example.

Although illustrate and describe embodiments of the invention, those having ordinary skill in the art will appreciate that: can carry out multiple change, amendment, replacement and modification to these embodiments when not departing from principle of the present invention and aim, scope of the present invention is by claim and equivalency thereof.

Claims (10)

1. an impulse noise removing method, is characterized in that, comprises the following steps:

S1: send signal frame frame head and signal frame frame respectively by T transmitting antenna of antenna system, wherein, described antenna system comprises T transmitting antenna and R reception antenna;

S2: each reception antenna obtains the value on the frequency domain preserved sub-carrier of described signal frame frame, form the domain observations vector of impulse noise signal, and form multidimensional impulse noise combined sampling matrix according to R domain observations vector of a described R reception antenna, and based on the time-frequency domain Fourier transform relation of impulse noise signal, obtain structuring compressed sensing model;

S3: according to described structuring compressed sensing model, adopt structure based compressed sensing algorithm to estimate time-domain position and the strength factor of the impulse noise signal of a described R reception antenna, obtain the estimation of the time domain impulse noise signal corresponding to a described R reception antenna;

S4: the estimation time-domain signal of the signal frame frame of a described R reception antenna being deducted corresponding time domain impulse noise signal, R reception antenna signal frame body data after the impulse noise signal that is eliminated.

2. method according to claim 1, it is characterized in that, wherein, described signal frame frame is OFDM data block, the frequency domain of described signal frame frame comprises multiple preserved sub-carrier, and described signal frame frame head is made up of the Cyclic Prefix of described signal frame frame or signal frame frame head training sequence.

3. method according to claim 1 and 2, is characterized in that, described step S2 also comprises:

Time-domain signal according to the signal frame frame of current reception carries out rough estimate to the time-domain position of impulse noise signal, specifically comprises:

Obtain the time domain sequences absolute value of the signal frame frame of a described R reception antenna square, and square all to compare R described time domain sequences absolute value with pre-determined threshold, wherein, described pre-determined threshold is the multiple of the mean value of the quadratic sum of the time domain sequences absolute value of the signal frame frame of a described R reception antenna;

Using the rough estimate of position as the time-domain position of described impulse noise signal square being all greater than described pre-determined threshold of absolute value in R time domain sequences of a described R reception antenna.

4. method according to claim 1 and 2, is characterized in that, described step S2, comprises further:

S21: when described preserved sub-carrier is pilot sub-carrier, calculate the value on the pilot sub-carrier of the described signal frame frame frequency domain of current reception, and the known transmission pilot value value on described pilot sub-carrier deducted in respective frequencies is multiplied by the estimated value gained sequence of channel frequency response, obtain the domain observations sequence of described impulse noise signal, when described preserved sub-carrier is vacant subcarrier, the domain observations sequence of described impulse noise signal is that the value on the vacant subcarrier of frequency domain forms domain observations vector successively;

S22: R the domain observations vector according to a described R reception antenna forms multidimensional impulse noise combined sampling matrix;

S23: based on the time-frequency domain Fourier transform relation of impulse noise signal, obtains described structuring compressed sensing model.

5. method according to claim 1, is characterized in that, in described step S2, described structuring compressed sensing model is multidimensional impulse noise signal relationship between frequency and time equation.

6. the method according to claim 1 or 4, it is characterized in that, when described signal frame frame head is made up of described signal frame frame head training sequence, the acquisition pattern of the value on the pilot sub-carrier of described signal frame frame frequency domain or the value on vacant subcarrier is: carry out loop restructuring and training sequence interference elimination to the time domain sequences of the described signal frame frame of current reception;

When described signal frame frame head is made up of the Cyclic Prefix of described signal frame frame, the acquisition pattern of the value on the pilot sub-carrier of described signal frame frame frequency domain or the value on vacant subcarrier is: directly carry out discrete Fourier transform to the time domain sequences of the described signal frame frame of current reception.

7. method according to claim 1, it is characterized in that, in described step S3, described structuring compressed sensing algorithm is based on the convex optimized algorithm of multivariate joint probability sparse signal or the greedy algorithm based on multidimensional structure compressed sensing, wherein, described convex optimized algorithm comprises interior point method, single order norm minimum algorithm; Described greedy algorithm comprises structuring synchronous orthogonal matching pursuit algorithm, structural sparse Adaptive matching back tracking method.

8. the method according to claim 1 or 7, it is characterized in that, in described step S3, adopt the rough estimate of the time-domain position of described impulse noise signal, as the prior information of structuring compressed sensing algorithm, carry out the auxiliary structuring compressed sensing iteration of prior information and impulse noise estimation.

9. method according to claim 1, is characterized in that, after described step S3, also comprises: the estimated accuracy optimizing the coefficient of described impulse noise signal, specifically comprises:

To each reception antenna, on the time-domain position of the described impulse noise signal estimated, be multiplied by residuals squares principle between impulse noise time-domain signal to be estimated and described impulse noise domain observations sequence to minimize described Fourier transform matrix, carry out least-squares estimation.

10. an impulse noise cancellation element, is characterized in that, comprising:

Compressed sensing model generation module, the value on the frequency domain preserved sub-carrier of described signal frame frame is received for obtaining R reception antenna, form the domain observations vector of impulse noise signal, and form multidimensional impulse noise combined sampling matrix according to R domain observations vector of a described R reception antenna, and based on the time-frequency domain Fourier transform relation of impulse noise signal, obtain structuring compressed sensing model;

Compressed sensing estimation module, for according to described structuring compressed sensing model, adopt structure based compressed sensing algorithm to estimate time-domain position and the strength factor of the impulse noise signal of a described R reception antenna, obtain the estimation of the time domain impulse noise signal corresponding to a described R reception antenna;

Impulse noise cancellation module, the time-domain signal for the signal frame frame by a described R reception antenna deducts the estimation of corresponding time domain impulse noise signal, R reception antenna signal frame body data after the impulse noise signal that is eliminated.