CN1652492A - A Method for Realizing Time-Frequency Synchronization of OFDM Communication System Based on Frequency Domain Correlation Detection - Google Patents

Abstract

The present invention realizes the time-frequency synchronization method of OFDM communication system based on frequency domain correlation detection. The principle of random sequence correlation in the spectrum domain, time synchronization is achieved by detecting the correlation peak of the pseudo-random sequence in the spectrum domain, and the fractional frequency offset of the subcarrier interval of the initial frequency offset is estimated by the complex value corresponding to the correlation peak; then Using the feedback mechanism, the original received time-domain prefix is compensated for the subcarrier spacing fractional frequency offset, and then the fast Fourier transform is performed to obtain the frequency-domain sequence, and the frequency-domain sequence is matched with the local pseudo-random sequence to obtain the sub-carrier spacing The integer multiple frequency offset is estimated, and then the initial phase is estimated by matching the complex value corresponding to the correlation peak; the method of the present invention can accurately realize time-frequency synchronization in a channel environment with a low signal-to-noise ratio.

Description

Realize the method for ofdm communication system Time and Frequency Synchronization based on the frequency domain coherent detection

Technical field:

The invention belongs to OFDM (OFDM) mobile communication technology field, relate generally to ofdm communication system Time and Frequency Synchronization technology.

Background technology:

For the higher speed of transmission in limited spectral bandwidth, adopted OFDM (OFDM) technology in the third generation mobile communication system mostly.Though the successful application in WLAN (wireless local area network) of OFDM technology in quick and various mobile wireless environment, be used the data service at a high speed of OFDM technical transmission, also has many problems to need to solve.Wherein, the Time and Frequency Synchronization problem is substantially the most also to be one of the most key technology in the ofdm system.

U.S.'s " international electronics communicate by letter journal " (IEEE Trans.Commun. with The Institution of Electrical Engineers, vol.45, pp.1613-1621, Dec.1997) a kind of method of carrying out time synchronized and frequency offset estimating in ofdm system has simultaneously been proposed, this method construct one long be the prefix of two OFDM cells, use is directly related at the repetitive sequence of time-domain, utilizes the method for peak value detection and maximal possibility estimation, realizes time synchronized and Frequency Synchronization simultaneously.This simultaneous techniques is effective under the good situation of channel condition, but in abominable mobile communication environment, it is not obvious that the decision value of this method becomes, and can not judge the original position of OFDM cell exactly, thereby influence correct time synchronized and frequency offset estimating.

In real system, because the prefix that is used for Time and Frequency Synchronization need take certain overhead, in order to reduce the system resource that Time and Frequency Synchronization consumes, that Chinese scholars constantly propose is new, more save the time-frequency synchronization method of system resource.U.S.'s " international electronics communicate by letter magazine " (IEEE Journal.Commun. with The Institution of Electrical Engineers, vol.19, pp.2495-2503, Dec.2001) introduced a kind of method of using an OFDM cell to realize Time and Frequency Synchronization, the frame prefix of ofdm system of at first having used MLS (Maximum Length Sequence) sequence structure, then detect the existence and the initial frequency deviation of packet with the time-domain correlation, the initial frequency deviation that obtains is fed back to the time domain correlator, carrying out compensate of frequency deviation, to carry out time domain later on more relevant, obtain the original position of Frame, realize time synchronized, use at last that the phase place of correlation obtains frequency departure on the time domain.This method is when realizing that packet detects, owing to do not know the original position of Frame, can not estimate initial frequency deviation accurately, thereby in feedback control loop, the frequency deviation that receives data can not be offseted fully, this will influence the correct estimation of Frame original position, and therefore this method can not adapt to the bigger system of frequency shift (FS) in the frequency shift (FS) of carrying out not estimating when initial frequency deviation is estimated OFDM subcarrier spacing integral multiple.

Technology contents:

The present invention is directed to the shortcoming of the existing various Time and Frequency Synchronization technology of ofdm communication system, a kind of method that realizes the ofdm communication system Time and Frequency Synchronization based on the frequency domain coherent detection is proposed, utilize frequency domain coherent detection principle, by using the sharp-pointed correlation peak of pseudo random sequence (PN sequence) frequency domain accurately to realize time synchronized, use the phase place of correlation peak complex values to come frequency departure and initial phase are accurately estimated simultaneously.

Thisly realize the method for ofdm communication system Time and Frequency Synchronization, comprise that structure frame prefix, initial time are synchronously and initial frequency synchronization based on the frequency domain coherent detection; Initial frequency synchronization comprises subcarrier spacing decimal times original frequency estimation of deviation and subcarrier spacing integral multiple original frequency estimation of deviation;

It is characterized in that:

Described structure frame prefix be used in interpolation pseudo random sequence on the even subcarrier (PN sequence) with on the strange subcarrier for empty frequency domain sequence, obtain time domain sequences after through Fast Fourier Transform Inverse (IFFT), this time domain sequences the first half is constant, and half carries out the time domain frame prefix sequence that time domain sequences conduct that differential coding obtains sends with the first half the back; The recipient, with receive length be an OFDM cell time domain base-band data signal the first half with the back half carry out differential decoding, the time domain sequences that differential decoding is later is carried out fast Fourier transform (FFT) and is obtained frequency domain sequence, relevant with local PN sequence with this frequency domain sequence, it is synchronous to obtain initial time by the sharp-pointed correlation peak of PN sequence; Phase place by the corresponding complex values of the sharp-pointed correlation peak of PN sequence comes a subcarrier spacing decimal times original frequency deviation is estimated; An initial time synchronizing information and a subcarrier spacing decimal times original frequency deviation are fed back to the primitive frame prefix sequence, offset subcarrier spacing decimal overtones band deviation, carry out FFT with having offseted the later time domain sequences of subcarrier spacing decimal overtones band deviation, obtain frequency domain sequence, this frequency domain sequence and local PN sequence are mated relevant, the frequency departure and the initial phase of subcarrier spacing integral multiple are estimated by the complex values that correlation peak is corresponding with correlation peak; When realizing that initial time is synchronous, to finish the later time-domain data point of phase compensation to each and carry out FFT one time, a kind of fast algorithm of pointwise FFT and the structure of this algorithm correspondence of realizing of recursive nature design when realizing according to this demand utilization pointwise FFT.

The inventive method is based on following principle:

Described frame prefix is made of an OFDM cell, and the length of establishing the OFDM cell is N (N is 2 integral number power), will grow to be N ₁(N ₁=N/2) pseudo random sequence PN (K), K=0 ..., N ₁-1 inserts in the even subcarrier of OFDM cell, data on the strange subcarrier of OFDM are empty (representing with 0) here, after this OFDM cell carried out IFFT, obtain the long time domain sequences g of N (n) that is, again g (n) is obtained later time domain sequences p (n) through coding, p (n) and g (n) concern shown in following formula:

$p\left(n\right)=\left\{\begin{array}{cc}g\left(n\right)& 0\&le;n<N/2\\ a\&times;\frac{g(n-N/2)}{g\left(n\right)}\&times;(1+j)& N/2\&le;n<N\end{array}\right.---\left(1\right)$

Wherein a is the amplitude of each subcarrier average energy correspondence of OFDM frequency domain cell.

With reference to the accompanying drawings 2, the antenna data of establishing from radio-frequency module (1) is τ wherein _sBe channel transmission time delay, h is the fading factor in cell time-slot of OFDM, f is for by the skew of the crystal oscillator frequency of receiving-transmitting sides and multispectrally rein in the frequency shift (FS) that frequency deviation causes, is the phase factor that channel delay and crystal oscillator phase deviation cause, T is the time-domain sampling cycle, w (n) is the white Gaussian noise of σ for variance, and the data that outputed to data channel (3) by cache module (2) are y (n)=r (n) * (1+j) * r ^*(n+N ₁), in pointwise FFT module (4) with the y (n-1) of each y (n) and front, y (n-2) ..., y (n-N ₁+ 1) makes N ₁Point FFT obtains sequence Y later on ⁿ(K) (K=0 ..., N ₁-1), if n＜N ₁-1 in y (n) front benefit 0, with Y ⁿ(K) (K=0 ..., N ₁-1) output to after the correlator (6) by data channel (5) relevant with local sequence, by correlation $P\left(n\right)=\frac{1}{{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">N</figure-callout>}}_1}\underset{k=0}{\overset{N_1-1}{\mathrm{\&Sigma;}}}\mathrm{PN}\left(K\right)\&times;Y^n\left(K\right),$ Normalized energy $R\left(n\right)=\frac{1}{{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">N</figure-callout>}}_1}\underset{K=0}{\overset{N_1-1}{\mathrm{\&Sigma;}}}{\left|Y^n\left(K\right)\right|}^2,$ Decision threshold $M\left(n\right)=\frac{{\left|P\left(n\right)\right|}^2}{R\left(n\right)}$ Detect the original position of Frame jointly, when energy R (n) greater than certain thresholding, the correlation peak that then detects M (n) is the starting point of OFDM cell, establishing the point that detects correlation peak is n _Opt, the system bandwidth of OFDM is B, then the subcarrier spacing fractional part of frequency offset $f_F=-\frac{\&angle;P\left(n_{\mathrm{opt}}\right)}{2\mathrm{\&pi;}}\&times;\frac{\mathrm{<figure-callout\; id="1"\; label="B\; N"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">B</figure-callout>}}{N_{}},$ ∠ P (n wherein _Opt) expression correlation P (n _Opt) angle (unit is a radian).

Estimated value in the correlator (6) is fed back in the cache module (2) by data channel (7), in cache module (2), carry out compensate of frequency deviation and obtain sequence U (n) (n=0 on the time-domain later on frame prefix sequence and subcarrier spacing fractional part of frequency offset estimated value, ..., N ₁-1), the expression formula of sequence U (n) is $U\left(n\right)=r(n_{\mathrm{opt}}-{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">N</figure-callout>}}_1+1+n)\exp \{-2\mathrm{\&pi;}\&times;f_F\&times;\frac{1}{B}\&times;n\},$ After sequence U (n) outputed to fast Fourier transform module (10) by data channel (9), in fast Fourier transform module (10), carry out N ₁The point FFT obtain later on sequence Q (K) (K=0 ..., N ₁-1), Q (K) is outputed to coupling correlation module (12), the relevant correlation that obtains with local PN sequences match in coupling correlation module (12) by data channel (11) $M\left(g\right)=\frac{1}{{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">N</figure-callout>}}_1}\underset{K=0}{\overset{N_1-1}{\mathrm{\&Sigma;}}}Q\left[(K+g)\mathrm{mod}{N}_1\right]\mathrm{PN}\left(K\right),$ By $g_{\max }=\underset{g\&Element;G}{\max }\left\{{\left|M\left(g\right)\right|}^2\right\},G=[-\frac{{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">N</figure-callout>}}_1}{2},\frac{{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">N</figure-callout>}}_1}{2})$ Try to achieve the point of correlation peak correspondence, then the subcarrier spacing integer frequency offset $f_I=g_{\max }\&times;\frac{\mathrm{<figure-callout\; id="1"\; label="B\; N"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">B</figure-callout>}}{N_{}},$ The frequency offset estimating f=f of whole channel _I+ f _F, initial phase =∠ M (g _Max), ∠ M (g wherein _Max) expression M (g _Max) angle (unit is a radian).

If it is { y that pointwise FFT module (4) receives from the data in the data channel (3) _n, n=0,1,2 ..., wherein ${\mathrm{<figure-callout\; id="0"\; label="y"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">y</figure-callout>}}_0={\mathrm{<figure-callout\; id="1"\; label="y"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">y</figure-callout>}}_1=...=y_{N_1-1}=0,$ N to each antenna data that receives and front ₁-1 point data is all done FFT one time, and the data acquisition system that obtains is ${\{Y}_k^t:t=\mathrm{0,1,2},...;K=\mathrm{0,1},...,N_1-1\},$ By formula

$Y_k^t=\underset{m=0}{\overset{N_1-1}{\mathrm{\&Sigma;}}}{y}_{t+m}{e}^{-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20041001399700102.png"\; state="\{\{state\}\}">j</figure-callout>}\frac{2\mathrm{\&pi;}}{{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">N</figure-callout>}}_1}\mathrm{Km}}(K=0,...,N_1-1)---\left(2\right)$

$Y_K^{t+1}=\underset{m=0}{\overset{N_1-1}{\mathrm{\&Sigma;}}}{y}_{t+m+1}{e}^{-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20041001399700102.png"\; state="\{\{state\}\}">j</figure-callout>}\frac{2\mathrm{\&pi;}}{{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">N</figure-callout>}}_1}\mathrm{Km}}={\mathrm{<figure-callout\; id="2"\; label="e\; j"\; filenames="A20041001399700102.png"\; state="\{\{state\}\}">e</figure-callout>}}^j\left(\underset{j=1}{\overset{N_1}{\mathrm{\&Sigma;}}}{y}_{t+j}{e}^{-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20041001399700102.png"\; state="\{\{state\}\}">j</figure-callout>}\frac{2\mathrm{\&pi;}}{{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">N</figure-callout>}}_1}\mathrm{Kj}}\right)(K=0,...,N_1-1)---\left(3\right)$

Then can obtain

$Y_K^{t+1}-Y_K^t\&times;{\mathrm{<figure-callout\; id="2"\; label="e\; j"\; filenames="A20041001399700102.png"\; state="\{\{state\}\}">e</figure-callout>}}^j={\mathrm{<figure-callout\; id="2"\; label="e\; j"\; filenames="A20041001399700102.png"\; state="\{\{state\}\}">e</figure-callout>}}^j(y_{t+N_1}-y_t)(K=0,...,N_1-1)---\left(4\right)$

$Y_K^{t+1}=e^{\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20041001399700102.png"\; state="\{\{state\}\}">j</figure-callout>}\frac{2\mathrm{\&pi;}}{N_1}K}(y_{t+N_1}-y_t+Y_K^t)(K=0,...,N_1-1)---\left(5\right)$

The present invention is a kind of method based on frequency domain coherent detection realization ofdm communication system Time and Frequency Synchronization, can be used in burst transfer mode and time synchronized, initial frequency deviation estimation and the initial phase estimation of transmission means continuously in the ofdm communication system; Because having finished time synchronized and subcarrier spacing fractional part of frequency offset in correlator (6) simultaneously estimates, overcome in the conventional method time synchronized and frequency offset estimating is asynchronous finishes, the precision of time synchronized has influenced the accuracy of Frequency Estimation, and frequency departure has been realized the estimation of time synchronized and subcarrier spacing fractional part of frequency offset accurately conversely with the winding problem of influence time synchronous accuracy; Because in the subcarrier spacing fractional part of frequency offset is estimated ± the phase overturn problem of π will cause the frequency difference of an integral multiple subcarrier spacing, therefore carrying out the subcarrier spacing fractional part of frequency offset among the present invention earlier estimates, carry out the subcarrier spacing integer frequency offset then and estimate, can in the subcarrier spacing integer frequency offset is estimated, utilize the relevant principle syndrome carrier spacing fractional part of frequency offset of coupling estimate in ± the phase overturn problem of π; By formula (5) as can be known, we do not need point of every input to carry out FFT one time for pointwise FFT, can utilize the recursive nature of pointwise FFT calculating, simplify the amount of calculation of pointwise FFT greatly.

Description of drawings:

Fig. 1 is the frame prefix structure chart based on the relevant Time and Frequency Synchronization of frequency domain.

Fig. 2 is the realization flow figure of time synchronized and Frequency Synchronization.

Fig. 3 is for realizing pointwise FFT fast algorithm schematic diagram.

Fig. 4 is a time-frequency synchronization method analogous diagram of the present invention.

Embodiment:

Below in conjunction with accompanying drawing embodiments of the invention are described.

Embodiment 1:

This example is an example with the ofdm system down link, illustrate initial time that the present invention is used for the ofdm communication system link synchronously, frequency offset estimating and initial phase estimate.

If the number of sub carrier wave of ofdm communication system is N=2048, ${\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">N</figure-callout>}}_1=\frac{\mathrm{<figure-callout\; id="2"\; label="N"\; filenames="A20041001399700102.png"\; state="\{\{state\}\}">N</figure-callout>}}{2}=1024,$ Sample rate is f _s=25.6MHz, then system bandwidth B=f _s=25.6MHz, Frame subcarrier spacing B _DiThe subcarrier spacing that=12.5KHz, recipient receive prefix signal is B _Pi=25KHz, the pseudo random sequence that transmit leg sends is PN (K) at frequency domain, K=0 ..., N ₁-1.

Based on the frame prefix structure of the relevant Time and Frequency Synchronization of frequency domain as shown in Figure 1:

Frame prefix is made of an OFDM cell, and the length of establishing the OFDM cell is N (N is 2 integral number power), among Fig. 1 first rectangular in hypographous little the expression length be

The PN sequence, with little of blank expression empty data (representing with 0 here), so just obtained growing and be the frequency domain sequence of N; After this frequency domain sequence carried out IFFT, obtain longly being the time domain sequences g of N (n) that g (n) is N by length ₁Time domain sequences S1 be N with long ₁Time domain sequences S2 constitute, in Fig. 1, be expressed as second rectangular; Again g (n) is obtained time domain sequences p (n) later on through differential coding, p (n) is N by length ₁Time domain sequences S1 be N with long ₁Time domain sequences S3 constitute, wherein S3 is obtained with sequence S2 coding by sequence S1, in Fig. 1, be expressed as the 3rd rectangular, p (n) and g (n) concern shown in following formula:

$p\left(n\right)=\left\{\begin{array}{cc}g\left(n\right)& 0\&le;n<N/2\\ a\&times;\frac{g(n-N/2)}{g\left(n\right)}\&times;(1+j)& N/2\&le;n<N\end{array}\right.$

Wherein a is the amplitude of each subcarrier average energy correspondence of OFDM frequency domain cell.

Fig. 2 has provided the realization flow figure of time synchronized and Frequency Synchronization:

If the data flow from radio-frequency module (1) is r (n), be y (n)=r (n) * (1+j) * r by the data that output to data channel (3) in the cache module (2) ^*(n+N ₁), in pointwise FFT module (4) with the y (n-1) of each y (n) and front, y (n-2) ..., y (n-N ₁+ 1) makes N ₁Point FFT obtains sequence Y later on ⁿ(K) (K=0 ..., N ₁-1), if n＜N ₁-1 in y (n) front benefit 0, with Y ⁿ(K) (K=0 ..., N ₁-1) output to after the correlator (6) by data channel (5) relevant with local sequence, by correlation $P\left(n\right)=\frac{1}{{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">N</figure-callout>}}_1}\underset{K=0}{\overset{N_1-1}{\mathrm{\&Sigma;}}}\mathrm{PN}\left(K\right)\&times;Y^n\left(K\right),$ Normalized energy $R\left(n\right)=\frac{1}{{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">N</figure-callout>}}_1}\underset{K=0}{\overset{N_1-1}{\mathrm{\&Sigma;}}}{\left|Y^n\left(K\right)\right|}^2,$ Decision threshold $M\left(n\right)=\frac{{\left|P\left(n\right)\right|}^2}{R\left(n\right)}$ Detect the original position of Frame jointly, when energy R (n) greater than certain thresholding, the correlation peak that then detects M (n) is the starting point of OFDM cell, establishing the point that detects correlation peak is n _Opt, the subcarrier spacing fractional part of frequency offset is $f_F=-\frac{\&angle;p\left(n_{\mathrm{opt}}\right)}{2\mathrm{\&pi;}}\&times;\frac{\mathrm{<figure-callout\; id="1"\; label="B\; N"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">B</figure-callout>}}{N_{}},$ ∠ p (n wherein _Opt) expression correlation p (n _Opt) angle (unit is a radian), so far realized that at correlator (6) time synchronized and subcarrier spacing fractional part of frequency offset estimate.

Estimated value in the correlator (6) is fed back in the cache module (2) by data channel (7), in cache module (2), carry out compensate of frequency deviation and obtain sequence U (n) (n=0 on the time-domain later on frame prefix sequence and subcarrier spacing fractional part of frequency offset estimated value, ..., N ₁-1), the expression formula of sequence U (n) is $U\left(n\right)=r(n_{\mathrm{opt}}-{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">N</figure-callout>}}_1+1+n)\exp \{-2\mathrm{\&pi;}\&times;f_F\&times;\frac{1}{B}\&times;n\},$ After sequence U (n) outputed to fast Fourier transform module (10) by data channel (9), carry out N in fast Fourier transform module (10) ₁The point FFT obtain later on sequence Q (K) (K=0 ..., N ₁-1), Q (K) is outputed to coupling correlation module (12), the relevant correlation that obtains with local PN sequences match in coupling correlation module (12) by data channel (11) $M\left(g\right)=\frac{1}{{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">N</figure-callout>}}_1}\underset{K=0}{\overset{N_1-1}{\mathrm{\&Sigma;}}}Q\left[(K+g)\mathrm{mod}{N}_1\right]\mathrm{PN}\left(K\right),$ By $g_{\max }=\underset{g\&Element;G}{\max }\left\{{\left|M\left(g\right)\right|}^2\right\},G=[-\frac{{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">N</figure-callout>}}_1}{2},\frac{{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">N</figure-callout>}}_1}{2})$ Try to achieve the point of correlation peak correspondence, then the subcarrier spacing integer frequency offset $f_I=g_{\max }\&times;\frac{\mathrm{<figure-callout\; id="1"\; label="B\; N"\; filenames="A20041001399700112.png"\; state="\{\{state\}\}">B</figure-callout>}}{N_{}},$ Initial phase =∠ M (g _Max), ∠ M (g wherein _Max) expression M (g _Max) angle (unit is a radian), so far realized that at coupling correlator (12) the subcarrier spacing integer frequency offset is estimated and initial phase is estimated the frequency offset estimating f=f of whole channel _I+ f _F

Accompanying drawing 3 has provided realization pointwise FFT fast algorithm schematic diagram, through the data flow y of differential decoding _nAt cache module A ₁, A ₂..., A _N1-1, A _N1Middle buffer memory N ₁Inferior, according to formula (5), the value Y that obtains after the output process time delay module (16) with FFT _K ^tWith time-domain signal y _T+N1,-y _tIn adder (14), carry out the later output valve of addition process in multiplier (15) with Multiply each other, just can access current FFT output Y as a result _K ^T+1Symbol D in the accompanying drawing in the time delay module (16) ^-1The expression time delay factor, the N when symbol t represents moment t ₁Individual point carries out pointwise FFT, t=0, and 1 ... the span of expression t is a nature manifold and zero; Symbol K is illustrated in K the FFT output point of t constantly, and the span of K is K=0, and 1 ..., N ₁-1.

Accompanying drawing 4 has provided the simulation result when signal to noise ratio (snr) is 0dB, this figure has comprised that frequency domain is relevant, frequency offset estimating and three subgraphs of phase estimation: by the relevant subgraph of frequency domain as can be seen, in signal to noise ratio is that the method that provides among the present invention under the channel condition of 0dB can access sharp-pointed correlation peak, realizes that precise time is synchronous; By the frequency offset estimating subgraph as can be seen, be that the frequency departure that estimates under the channel condition of 0dB can be accurate in the scope of [M-0.1, M+0.1] in signal to noise ratio, wherein M is the original frequency deviation that normalizes to subcarrier spacing; By the phase estimation subgraph as can be seen, be that the phase deviation that estimates under the channel condition of 0dB can be accurate in the scope of [-0.1 * π, +0.1 * π] in signal to noise ratio, wherein is original phase factor.

Claims (1)

1. A method for realizing time-frequency synchronization of an OFDM communication system based on frequency-domain correlation detection, comprising constructing a frame prefix, initial time synchronization, and initial frequency synchronization; initial frequency synchronization includes subcarrier spacing fractional times the initial frequency The bias estimate is an integer multiple of the subcarrier spacing and the initial frequency bias estimate: It is characterized by: The frame prefix construction is used to interpolate a pseudo-random sequence on an even subcarrier and an empty frequency domain sequence on an odd subcarrier, and obtain a time domain sequence after inverse fast Fourier transform. Half remains unchanged, and the time-domain sequence obtained by differential encoding of the second half and the first half is used as the time-domain frame prefix sequence sent out; at the receiver, the length of an OFDM cell time-domain baseband data signal will be received The first half and the second half of the differential decoding are performed, and the time domain sequence after differential decoding is fast Fourier transformed to obtain the frequency domain sequence. The frequency domain sequence is correlated with the local pseudo random sequence, and the sharp correlation peak of the pseudo random sequence is used to determine the Obtain initial time synchronization; estimate the initial frequency deviation of fractional times of subcarrier spacing by the phase of the sharp correlation peak of the pseudo-random sequence corresponding to the complex value; feed back the initial time synchronization information and the initial frequency deviation of fractional times of subcarrier spacing to the original The frame prefix sequence cancels the fractional multiple frequency deviation of the subcarrier spacing, performs fast Fourier transform on the time domain sequence after canceling the fractional multiple frequency deviation of the subcarrier spacing, obtains the frequency domain sequence, and combines the frequency domain sequence with The local pseudo-random sequence is matched and correlated, and the frequency deviation and the initial phase of the integer multiple of the subcarrier interval are estimated by the correlation peak and the complex value corresponding to the correlation peak; when realizing the initial time synchronization, it is necessary to perform phase compensation for each A fast Fourier transform is performed on the data points in the time domain. According to this requirement, a fast algorithm for implementing point-by-point fast Fourier transform and the structure corresponding to this algorithm are designed by using the recursive nature of the point-by-point fast Fourier transform. .

Images