CN104954305A - Improved estimation method of phase noise in wireless communication system - Google Patents

Abstract

The invention belongs to the technical field of wireless communication, particularly relates to estimation of phase noise in a wireless communication system using a relevant algorithm, and provides a phase noise estimation method in the wireless communication system using the algorithm. According to the method, equivalent discrete time-domain channel impulse response is utilized to estimate CPE (common phase error) of the phase noise via an interpolation mode, and the phase noise is finally estimated by an iterative approach, thus system reliability is improved and error rate is reduced.

Description

A kind of phase noise improved estimator method under wireless communication system

Technical field

Wireless communication technology field of the present invention, particularly relates to the estimation being realized phase noise in a wireless communication system by respective algorithms.

Background technology

Modern digital communication systems has and uses frequency range accordingly, and system needs just can deliver in transmission medium after information sequence being transferred to designated frequency band, and receiving terminal then needs Received signal strength to be down-converted to base band to facilitate subsequent treatment.So all need to produce corresponding carrier wave to complete corresponding up-conversion and down-conversion operation at transmitting terminal and receiving terminal.But carrier wave is not desirable, because the crystal oscillator and phase-locked loop that produce carrier wave exist certain stability, thus cause and carry > ripple frequency and target frequency and there is random difference in short-term, and then cause produced sine wave signal generation random phase saltus step, show as phase noise.

Phase noise, is actually and characterizes the one of frequency source frequency stability.Under normal circumstances, frequency stability is divided into long-term frequency stability and short-term frequency stability.So-called short-term frequency stability, refers to the phase fluctuation or frequency fluctuation that are caused by random noise.As for the frequency slow drift caused because of temperature, aging etc., be then referred to as long-term frequency stability.Usually main it is considered that short-term stability problem, can think that phase noise is exactly short-term frequency stability, only two kinds of different representations of a physical phenomenon.For oscillator, frequency stability is that the one that it produces same frequency within the scope of whole official hour is measured.If signal frequency exists instantaneous change, can not remain unchanged, so signal source just also exists unsteadiness, and cause is exactly phase noise.

Phase noise is normally defined the dBc/Hz value at a certain given deviation frequency place, and wherein, dBc is the ratio of this frequency place power in units of dB and gross power.Oscillator is defined as the gross power ratio of signal power in this frequency place 1Hz bandwidth and signal at the phase noise at a certain deviation frequency place.

The estimation of phase noise and tracking can be divided into feedforward and feedback two large modes.Feed-forward mode needs the signal of reception to be divided into many segments, then estimates the average phase error of every section respectively.Current research major part is concentrated in an ofdm system.On the whole, have based on feedback and the phase noise compensation algorithm without feedback.In an ofdm system, phase noise can produce common phase error (CPE) and inter-carrier interference (ICI); In SC-FDE system, produce common phase error (CPE) and intersymbol interference (ISI), cause the error rate to increase.

Summary of the invention

The invention provides one in a wireless communication system, utilize the estimation of algorithm realization phase noise.The method utilizes equivalent dispersion time domain channel impulse to respond the common phase error (CPE) being estimated phase noise by interpolation method, and the method finally by iteration realizes the estimation of phase noise.

In order to describe content of the present invention easily, first the term used in the present invention is described:

Special word (UW, Unique Word).In order to carry out synchronous or parameter Estimation etc. at receiving terminal, transmitting terminal send have some particular characteristics, to the known special sequence of receiving terminal.

A kind of phase noise improved estimator method under wireless communication system, comprises the steps:

S1, phase noise estimate initialization, are specially:

S11, the form y that time-domain signal is expressed as matrix will be obtained at receiving terminal _{n × 1}=A _{n × N}h _{n × N}x _{n × 1}+ w _{n × 1}, wherein, y _{n × 1}the form of N × 1 column vector, A _{n × N}a N × N diagonal matrix formed by making an uproar mutually, h _{n × N}for Teoplitz (toeplitz) matrix that equivalent time domain channel impulse response is formed, x _{n × 1}for N × 1 column vector that transmission data are formed, w _{n × 1}for the noise vector of N × 1, N is non-vanishing natural number;

S12, channel estimating is utilized to obtain the impulse response of corresponding equivalent time domain channel special word (UW) is joined in the data sequence that will transmit, realizes removing intersymbol interference;

S13, the UW passed through in transmission data sequence, adopt the mode of interpolation to estimate make an uproar mutually constant a (i) and the ratio of a of i-th data block transmitted between i-th UW with the i-th+1 UW, get the draw value of described ratio as phase noise according to a preliminary estimate: $\frac{a\left(i\right)}{a}=\frac{1}{2}\×(\frac{a\_{\mathrm{uw}}_{\left(i\right)}}{a}+\frac{a\_{\mathrm{uw}}_{(i+1)}}{a});$

S2, iteration, be specially: the auto-correlation function utilizing time domain phase noise, constructs autocorrelation matrix, Eigenvalues Decomposition is carried out to described autocorrelation matrix, obtain characteristic value and corresponding orthogonal eigenvectors, recycling Karhunen-Loeve transformation realizes the estimation to phase noise;

S3, compensation phase noises realize the estimation of transmission data block transmission data.

The invention has the beneficial effects as follows:

The present invention utilizes equivalent dispersion time domain channel impulse to respond the common phase error (CPE) being estimated phase noise by interpolation method, is realized the estimation of phase noise by the method for iteration, improves the reliability of system, reduces the error rate.

Accompanying drawing explanation

Fig. 1 is the single-carrier frequency domain equalization system illustraton of model by effect of phase noise that the present invention uses.

Fig. 2 is the statistical model figure that makes an uproar mutually that the present invention uses.

Fig. 3 is the flow chart that the inventive method realizes eliminating effect of phase noise in single carrier frequency domain equalization communication system.

Embodiment

Below in conjunction with embodiment and accompanying drawing, describe technical scheme of the present invention in detail.

S1, phase noise estimate initialization, are specially:

S11, the form y that time-domain signal is expressed as matrix will be obtained at receiving terminal _{n × 1}=A _{n × N}h _{n × N}x _{n × 1}+ w _{n × 1}, wherein, y _{n × 1}the form of N × 1 column vector, A _{n × N}a N × N diagonal matrix formed by making an uproar mutually, h _{n × N}for Teoplitz (toeplitz) matrix that equivalent time domain channel impulse response is formed, x _{n × 1}for N × 1 column vector that transmission data are formed, w _{n × 1}for the noise vector of N × 1, N is non-vanishing natural number;

S12, channel estimating is utilized to obtain the impulse response of corresponding equivalent time domain channel special word (UW) is joined in the data sequence that will transmit, realizes removing intersymbol interference;

S13, the UW passed through in transmission data sequence, adopt the mode of interpolation to estimate make an uproar mutually constant a (i) and the ratio of a of i-th data block transmitted between i-th UW with the i-th+1 UW, get the draw value of described ratio as phase noise according to a preliminary estimate: $\frac{a\left(i\right)}{a}=\frac{1}{2}\×(\frac{a\_{\mathrm{uw}}_{\left(i\right)}}{a}+\frac{a\_{\mathrm{uw}}_{(i+1)}}{a});$

S2, iteration, be specially: the auto-correlation function utilizing time domain phase noise, constructs autocorrelation matrix, Eigenvalues Decomposition is carried out to described autocorrelation matrix, obtain characteristic value and corresponding orthogonal eigenvectors, recycling Karhunen-Loeve transformation realizes the estimation to phase noise.

For IEEE802.11ad standard, the phase noise PSD model provided can calculate the auto-correlation function drawing time domain phase noise: $\mathrm{\ρ}\left(\mathrm{\τ}\right)=E\left[\mathrm{\φ}\left(t\right)\mathrm{\φ}(t+\mathrm{\τ})\right]=\frac{K_{\mathrm{\φ}}{f}_p^2}{{f_z}^2}\mathrm{\δ}\left(\mathrm{\τ}\right)+K_{\mathrm{\φ}}{\mathrm{\πf}}_p(1-\frac{f_p^2}{{f_z}^2})e^{-2{\mathrm{\πf}}_p\left|\mathrm{\τ}\right|},$ The orthogonal eigenvectors group duplicate removal structure phase noise utilizing autocorrelation matrix Eigenvalues Decomposition to obtain and φ=Ua.

W _{n × 1}received signal strength time-domain representation is: y _{n × 1}=diag (p _{n × 1}) h _{n × N}x _{n × 1}+ w _{n × 1}, wherein, y _{n × 1}the matrix of the time domain formation of Received signal strength, diag (p _{n × N}) diagonal matrix that forms for time domain of making an uproar mutually, h _{n × N}for the toeplitz matrix that the time domain of channel is formed, x _{n × 1}for the matrix that transmission data time domain is formed, w _{n × 1}for the matrix that noise time domain is configured to.? as initial condition structure A _{1, N × N}, utilize e ^{j φ}≈ 1+j φ and y _{n × 1}=diag (p _{n × 1}) h _{n × N}x _{n × 1}+ w _{n × 1}and phase noise autocorrelation matrix decomposes these three conditions of characteristic vector group φ=Ua obtained.Gone out by continuous iterative estimate φ=Ua is utilized to realize the estimation of making an uproar to phase place.

S3, compensation phase noises realize the estimation of transmission data block transmission data.

Fig. 1 is the single-carrier frequency domain equalization system illustraton of model by effect of phase noise that the present invention uses.

Transmitting terminal uses single carrier block transmission (Single-Carrier Block Transmission, SCBT), and bit stream adds protection interval block transmission after chnnel coding, digit mapping.In SC-FDE system; protection is spaced apart known UW composition, and the benefit done like this is the equivalent cycle prefix (Cyclic Prefix, CP) both having constituted next block; also as the known array of receiving terminal parameter Estimation, data transmission efficiency can be improve.

Signal by receiving terminal noise effect, enters numeric field process after channel after sampling.From digital receiver, the equivalent channel of digital signal process comprises the matched filter of the forming filter of transmitter, physical propagation channel and receiver.The sampled signal of digital receiver is affected by the multiplicative of additive white Gaussian noise (Addictive White Gaussian Noise, the AWGN) Additive effect in channel and phase noise.

Receiver, the signal serioparallel exchange received, carries out frequency domain equalization through FFT to frequency domain after removing equivalent CP, then turns back to that time domain carries out adjudicating through IFFT, digital demodulation and channel-decoding, finally sends bit stream to terminal.

If the user data after serioparallel exchange is S _i=[s _i(0), s _i(1) ..., s _i(N _s-1)] ^t, wherein, [] ^tthe transposition of representing matrix or vector, N _sthe length of user data in i-th data block.U=[u (0), u (1) ..., u (N _cp-1)] ^tbe the UW inserted between data, the length of the UW of described insertion is N _cp.Add UW after user data and form i-th complete data block X _i=[s _i(0), s _i(1) ..., s _i(N _s-1), u (0) ..., u (N _cp-1)] ^t, described i-th complete data block X _ilength be N _b=N _s+ N _cp.The UW inserted, as the equivalent CP of next data block, adds the growth data block of equivalent CP be expressed as in the matrix form wherein, T _c=[0 _{ncp × (Nb-Ncp)}, I _ncp; I _nb], described T _cby N _cp× (N _b-N _cp) null matrix, N _cp× N _cpunit matrix and N _b× N _bunit matrix composition (N _b+ N _cp) × N _bcyclic Prefix matrix.In the SC-FDE system of UW equivalence CP, described in be i-th data block X _iform with afterbody UW in the i-th-1 data block.

Transmitting terminal i-th growth data block the equivalent dispersion CIR of process is h _i=[h _i(0), h _i(1) ... h _i(L-1)] ^Τ, wherein, L is the maximum length of channel.Suppose ideal synchronisation, remove and receive data block r _ithe time-domain received signal that obtains of CP by the Additive effect of additive white Gaussian noise and phase noise phase place deflection impact, obtaining time-domain received signal matrix is y _i=A _ih _ix _i+ n _i, wherein, A _i=diag{ φ _ithe diagonal matrix formed is deflected, diag{ [] by the phase place of phase noise ^trepresent using vector as cornerwise diagonal matrix, H _in _b× N _bcirculation Toeplitz matrix, described H _ifirst row be h _iadding 0, to extend to length be N _bchannel vector, n _in _bthe AWGN vector of × 1, described n _ibe 0 by average, variance is separate AWGN composition.To y _icarry out N _bit is Y that some FFT obtains frequency-domain received signal _i=Fy _i=Φ _ihH _ix _i+ N _i, wherein, Y _i=[Y _i(0), Y _i(1) ..., Y _i(N _b-1)] ^tthe frequency-domain received signal of data block, Φ _in _b× N _bcirculation Toeplitz matrix, described Φ _ifirst row be phase noise φ _in _bpoint DFT, HH _i=diag{ [H _i(0), H _i(1) ..., H _i(N _b-1)] ^t, H _ik () is h _in _bpoint DFT, 0≤k≤N _b-1, X _i=[X _i(0), X _i(1) ..., X _i(N _b-1)] ^tx _in _bpoint DFT, N _ifor frequency domain additive white Gaussian noise, described N _ifor n _in _bpoint DFT.

Do not consider the impact of noise item, frequency-domain received signal Y _ik () is only by H _ithe impact of (k) and by sending other frequencies of signal X _il (), the crosstalk of l ≠ k, the complexity of frequency domain equalization reduces greatly relative to time domain equalization.

Frequency-domain received signal Y _in after frequency domain equalization _bpoint IFFT turns back to time domain and obtains:

wherein, C _i=diag{ [C _i(0), C _i(1) ..., C _i(N _b-1)] ^Τn _b× N _bfrequency domain equalization coefficient matrix, be the noise of the linear conversion of time domain AWGN, its result is still AWGN.

Fig. 2 is the statistical model figure that makes an uproar mutually that the present invention uses.Phase noise generally uses its power spectral density (PSD) to characterize, and communication standard IEEE802.15.3c and IEEE802.11ad gives " pole/zero " model about phase noise PSD:

wherein, f represents the frequency of cheap carrier center, and PSD (0) is a constant, at IEEE802.15.3c and IEEE802.11ad standard value-87dbc/Hz and-90dbc/Hz respectively.F _p=1MHz is pole frequency, f _z=100MHz is zero frequency.

Stable Gaussian look phase noise is regarded as the band limit gauss heat source model that the white Gaussian noise of zero-mean is obtained by low pass filter.

The mould square of this low pass filter transfer function can be decomposed into:

${\left|H\left(f\right)\right|}^2=\frac{1+{(f/f_Z)}^2}{1+{(f/f_p)}^2}=\frac{1+j(f/f_Z)}{1+j(f/f_p)}\frac{1-j(f/f_Z)}{1-j(f/f_p)},$ From above formula, choose the transfer function of a systems stabilisation, obtain transfer function H (s) of simulation low-pass filter: wherein, s=j Ω=j2 π f, a=1/2 π f _z, b=1/2 π f _p, f _pand f _zbe respectively pole frequency and zero frequency.Utilize Bilinear transformation method to realize analog filter and be converted into digital filter: wherein, constant C=π (f _p+ f _z)/tan (π (f _p+ f _z)/2f _s) make at (f _p+ f _zit is linear that the analog frequency at this Frequency point place)/2 is transformed into numerical frequency.Obtain the transfer function of digital filter thus: $H\left(z\right)=\frac{(1+\mathrm{aC})+{(1-\mathrm{aC})z}^{-1}}{(1+\mathrm{bC})+{(1-\mathrm{bC})z}^{-1}}.$

Fig. 3 is the flow chart utilizing the phase noise method of estimation of this iteration to realize eliminating effect of phase noise in single carrier frequency domain equalization communication system.

Channel estimation sequence is utilized to estimate equivalent time domain channel impulse response a is constant of making an uproar mutually, and h is actual equivalent time domain channel impulse response.Then Interpolate estimation goes out the ratio of make an uproar mutually constant and a of i-th data block transmitted between i-th UW with the i-th+1 UW go the matrix notation of the time-domain received signal after CP: y _i=A _ih _ix _i+ n _i.? construct as initial condition $A_{1,i}=\left[\begin{array}{ccc}\frac{a\left(i\right)}{a}& \·\·\·& \·\·\·\\ \·\·\·& \frac{a\left(i\right)}{a}& \·\·\·\\ \·\·\·& \·\·\·& \frac{a\left(i\right)}{a}\end{array}\right].$ For the system of single carrier frequency domain equalization, data block length is N _bx _imiddle length is N _cpuW be known.Thus, can be able to be transformed to by splitting the matrix realizing the time-domain received signal after removing CP $y_i=A_i{\stackrel{\‾}{H}}_{1,i}{\stackrel{\‾}{x}}_i+A_i{\stackrel{\‾}{H}}_{2,i}{\stackrel{\‾}{x}}_{\mathrm{cp}}+n_i.$ We can estimate like this

Utilize e ^{j φ}≈ 1+j φ, can obtain:

Equation is deformed into:

Real part and imaginary part are got to above formula:

Imaginary part and real part are re-constructed matrix, and matrix is:

y′＝A′φ′+n′

Phase noise can launch with orthogonal basis, i.e. φ ' ≈ Ua, and in formula, a is the coefficient that will estimate, U represents that getting phase place makes an uproar

The matrix that before sound, several characteristic value characteristic of correspondence vector is enough.Equation is deformed into:

y′＝A′φ′+n′＝A′Ua+n′

Utilize Least Square Method a:

$\hat{a}=\arg \min {\left|\right|y^{\′}-A^{\′}\mathrm{Ua}\left|\right|}^2$

Utilization estimates reconstructed phase noise, the phase noise of reconstruct is compensated to received signal, then carry out frequency domain equalization according to MMMSE criterion and obtain data estimator x, again the x of estimation is gone in noise reduction process iterative feedback substitution formula again to estimate to upgrade phase noise, iteration several times after, finally obtain the estimated value of phase noise.

Claims (1)

1. a kind of phase noise improved estimator method under wireless communication system, is characterized in that, comprises the steps: that S1, phase noise estimate initialization, is specially:

S11, the form y that time-domain signal is expressed as matrix will be obtained at receiving terminal _{n × 1}=A _{n × N}h _{n × N}x _{n × 1}+ w _{n × 1}, wherein, y _{n × 1}the form of N × 1 column vector, A _{n × N}a N × N diagonal matrix formed by making an uproar mutually, h _{n × N}for Teoplitz (toeplitz) matrix that equivalent time domain channel impulse response is formed, x _{n × 1}for N × 1 column vector that transmission data are formed, w _{n × 1}for the noise vector of N × 1, N is non-vanishing natural number;

S12, channel estimating is utilized to obtain the impulse response of corresponding equivalent time domain channel special word (UW) is joined in the data sequence that will transmit, realizes removing intersymbol interference;

S13, the UW passed through in transmission data sequence, adopt the mode of interpolation to estimate make an uproar mutually constant a (i) and the ratio of a of i-th data block transmitted between i-th UW with the i-th+1 UW, get the draw value of described ratio as phase noise according to a preliminary estimate: $\frac{a\left(i\right)}{a}=\frac{1}{2}\×(\frac{a\_{\mathrm{uw}}_{\left(i\right)}}{a}+\frac{a\_{\mathrm{uw}}_{(i+1)}}{a});$

S2, iteration, be specially: the auto-correlation function utilizing time domain phase noise, constructs autocorrelation matrix, Eigenvalues Decomposition is carried out to described autocorrelation matrix, obtain characteristic value and corresponding orthogonal eigenvectors, recycling Karhunen-Loeve transformation realizes the estimation to phase noise;

S3, compensation phase noises realize the estimation of transmission data block transmission data.