CN101414990B - Method for capturing carrier frequency bias and time delay of single carrier frequency domain equalizing system - Google Patents

Abstract

The invention discloses a method used for capturing carrier frequency offset and time delay in a single-carrier frequency domain balancing system, comprising the steps as follows: frame-arrival detection is carried out on the received data; when effective data frame is detected, the frame head position of the data frame is determined roughly; the determined frame head position is used as an initial position to carry out rough frequency offset estimation and compensation on the received data frame; the rough timing estimation is carried out so as to determine the initial position of the data block; the timing deviation is locked within one code element offset; the precise timing estimation is carried out by combining a maximum-likelihood searching method with an O&M timing estimation arithmetic, thus gaining precise code element timing information; precise frequency offset estimation and compensation are carried out; synchronous error lock judgment and correction are carried out. The method designs unique preamble code data form and structure, precisely carries out the estimation and compensation on carrier frequency offset and time delay during the synchronous capturing stage, gains precise carrier frequency offset and timing information, and determines the initial position of Fourier transform windows more exactly.

Description

Carrier wave frequency deviation in a kind of single-carrier frequency domain equalization system and the catching method of time delay

Technical field

The present invention relates to SC-FDE (single carrier system with frequency domainequalization, single carrier frequency domain equalization) technology, relate in particular to carrier wave frequency deviation in a kind of SC-FDE system and the catching method of time delay.

Background technology

At present, along with new communication service demand increases rapidly, transmission rate to wireless communication system and WLAN (wireless local area network) is had higher requirement, and the raising of transmission rate has brought the problems such as larger carrier wave frequency deviation, more serious ISI (intersymbol-interference, intersymbol interference) for conventional single-carrier system.OFDM (Orthogonal Frequency Division Multiplexing, OFDM) technology can effectively overcome the intersymbol interference that the weak channel of frequency selectivity brings, become gradually the study hotspot of radio communication and moving communicating field, in multiple standards, be called as support technology.But the OFDM technology is relatively more responsive to carrier synchronization, and PAPR (Peak-to-Average Power Ratio, peak-to-average power ratio) is larger, so the SC-FDE technology is suggested.The SC-FDE technology is the method that a kind of up-and-coming anti-multipath disturbs in the broadband wireless transmission, the same piecemeal of taking with OFDM transmits, and adopt CP (Cyclic Prefix, Cyclic Prefix) mode, so just can be converted into circular convolution to the linear convolution of signal and channel impulse response, and eliminate the interference of the data block that multipath causes.Adopt simple frequency-domain equalization technology just can eliminate intersymbol interference at receiving terminal.The SC-FDE system compares ofdm system and does not have the PAPR problem, thus do not need to use expensive linear power amplifier, simultaneously neither be responsive especially to carrier synchronization, so the SC-FDE technology is subject to increasing attention at present.

A most crucial problem in the digital communication technology field is exactly stationary problem, is divided into technically carrier synchronization and Timing Synchronization.In the baseband sampling signal that obtains through the receiving terminal behind the wireless mobile channel, exist carrier frequency offset, phase deviation and timing offset, these deviations generally derive from crystal oscillator between the transmitter and receiver do not mate with wireless mobile channel in Doppler effect and these several aspects of frequency selectivity, even if in the indoor environment of channel relatively flat, also can there be these deviations.Use the SC-FDE system of piece transmission technology that these synchronous deviations are had higher requirement.The simultaneous techniques of SC-FDE system is divided into carrier synchronization and Timing Synchronization, yet, just the synchronizing process of SC-FDE system is divided in the prior art and catches and follow the trail of two stages, do not have concrete algorithm.

Summary of the invention

The receiving terminal in the SC-FDE system the invention provides carrier wave frequency deviation in a kind of SC-FDE system and the catching method of time delay, so that can be determined the original position of Fourier transform window more accurately when receive data.

Technical scheme of the present invention is as follows:

Carrier wave frequency deviation in a kind of single-carrier frequency domain equalization system and the catching method of time delay, the sample rate of receiving terminal is 4 samplings, comprises step:

A, the data that receive are carried out frame arrive to detect, when detecting the valid data frame, determine the frame head position of Frame;

B, the described frame head position of determining carried out to the Frame that receives as original position thick frequency deviation is estimated and compensation;

C, carry out thick timing estimation, determine the original position of data block, timing offset is locked within the code element deviation;

D, in conjunction with maximum likelihood searching algorithm and O﹠amp; M timing estimation algorithm carries out smart timing estimation, obtains accurate symbol timing information;

E, carry out that smart frequency deviation is estimated and compensation;

F, carry out that synchronization error lock is judged and correct.

Preferably, describedly the data that receive carried out the implementation procedure that frame arrive to detect be:

A1, the original position point d of sliding window is set to d=1;

A2, obtain continuously two window length take d as original position and be the sampled value of 4G sampled point, R _i=(r (d+4 (i-1) G) ..., r (d+4iG-1)) ^T, i=1,2, G=32 wherein, the Baud Length that expression is obtained, r (n), { n=1,2,3......} represent to receive signal sampling value;

A3, determine R ₁, R ₂Correlation P (d), $P\left(d\right)={R_1}^H{\mathrm{<figure-callout\; id="2"\; label="R"\; filenames="H2008102278779E00051.png"\; state="\{\{state\}\}">R</figure-callout>}}_2=\underset{k=d}{\overset{d+4G-1}{\mathrm{\&Sigma;}}}{r}^{*}\left(k\right)r(k+4G);$

A4, determine the average energy of two windows getting access to $R\left(d\right)=\frac{1}{2}\underset{k=d}{\overset{d+8G-1}{\mathrm{\&Sigma;}}}{\left|r\left(k\right)\right|}^2;$

A5, calculating

Judge M _A(d) whether more than or equal to predetermined threshold value M _AIf,, then stopping mobile sliding window, the sliding window original position point of this moment is the frame head position of Frame, otherwise, the original position point of mobile sliding window is added 1, return steps A 2.

Preferably, described predetermined threshold value is M _A=0.7.

Preferably, among the described step B, the detailed process of the Frame that receives being carried out thick frequency deviation estimation and compensation is:

The frame head location positioning of B1, the Frame that obtains during frame arrive detected is the original position of sliding window, and the length of sliding window is N _B* L _BBaud Length, i.e. 4N _B* L _BIndividual sampled point, initial count value m are 0;

B2, m is added 1, judge that whether m is less than default mobile number of times threshold value M _BIf,, execution in step B3, otherwise, execution in step B5;

B3, begin to choose continuously 4N from the original position of sliding window _B* L _BThe observation data R of individual sampled point length _m, the sliding window original position is moved 4L to data transfer direction _BAfter the individual sampled point length, the original position from sliding window begins to choose continuously 4N again _B* L _BThe observation data R of individual sampled point length _M+1

Wherein, R _m=[r (1+4 (m-1) L _B) ..., r (4 (N _B+ m-1) L _B)], m={1,2.....M _B, the signal sampling value of r (n) expression receiving terminal;

B4, determine R _mAnd R _M+1Correlation P _m, $P_m=R_m{R}_{m+1}^H=\underset{n=1+4(m-1)L_B}{\overset{4(N_B+m-1)L_B}{\mathrm{\&Sigma;}}}r\left(n\right)r^{*}(n+4L_B),$ Then determine correlation P _mArgument X _mAnd give record, return step B2;

B5, the M to having recorded _BIndividual X _mValue is weighted summation, with the result of weighted sum divided by related interval L _B, 2 π and code element transmission time T product obtain frequency deviation estimated value Δ f;

B6, each sampled value of receive data is carried out compensate of frequency deviation: r (n)=r (n) e ^{-j2 π n Δ f/T}

Preferably, described related interval L _B=8 code elements, N _B=4, default mobile number of times threshold value M _B=6.

Preferably, among the described step C, the detailed process of thick timing estimation is:

C1, the original position point d of sliding window is set to 0, with mark position 0;

C2, the d value is added 1, d=1 the position for estimate through thick frequency deviation and compensation after this moment the sampled point place the position, the sampled value R that it is 4J sampled point as J Baud Length all that the original position take d as sliding window is obtained 4 window length continuously _i, R _i=(r (d+4 (i-1) J) ..., r (d+4iJ-1)) ^T, i=1,2,3,4;

Measure value M under C3, the calculating current location d _C(d), $M_C\left(d\right)=\frac{{|{R_1}^H{R}_2-{R_3}^H{R}_4-{R_2}^H{R}_3|}^2+{|{R_2}^H{R}_4-{R_1}^H{R}_3|}^2+{\left|{R_1}^H{R}_4\right|}^2}{14{\left|R_s\right|}^2},$ Wherein, R _i ^HR _jBe the conjugation of 4J sampled value of i the sequence sampled value corresponding with j sequence product cumulative and, R _sThe average energy value for sliding window;

C4, judgement M _C(d) whether estimate threshold value M more than or equal to default _CIf,, the M that then this is calculated _C(d) value is added in the Measure Sequences, with mark position 1, returns step C2; Otherwise, execution in step C5;

C5, judge whether current mark bit is 1, if estimate maximum, the location point that it is corresponding in the search Measure Sequences

Be the synchronous position of piece, otherwise, step C2 returned.

Preferably, the described default threshold value M that estimates _C=0.4.

Preferably, among the described step D, the detailed process of smart timing estimation is:

D1, utilize based on the maximum likelihood searching algorithm of BAKER code near the sampled point Q of timing position ₁And Q ₁Offset Δ T;

D2, with Q ₁Be original position, data intercept section [r (Q ₁) ..., r (Q ₁+ 4M-1)], with symbol S=[s (1) ..., s (4M)] and expression, i.e. S=[s (1) ..., s (4M)]=[r (Q ₁) ..., r (Q ₁+ 4M-1)], wherein M is the code element number in the data segment of intercepting, the length of the data block that sends with transmitting terminal is identical, and S is undertaken obtaining behind the bandpass filtering by the first band pass filter

D3, calculating Z=1.....4M, F () is nonlinear function, with what obtain

Undertaken obtaining behind the bandpass filtering by the second band pass filter ${\stackrel{\&RightArrow;}{S}}_2\left(z\right)=[s_2\left(1\right),...,s_2\left(4M\right)];$

D4, calculating residue timing offset

Make V=4 represent that the sample rate of receiving terminal is 4 samplings, namely

Total timing estimation value is

Wherein T is the code element transmission time.

Preferably, described step D1 specifically comprises step:

The sampling point position of determining synchronously according to piece obtains the estimated value P of the original position of BAKER code in the lead data;

The hunting zone of being determined the original position of BAKER code by estimated value P is: centered by the P point, front and back each I sampled point is determined with the BAKER code by the hunting zone and to be made the needed receiving sequence R of correlation computations;

Calculate the mould square C (k) of the correlation of 2I+1 BAKER code sequence B and receiving sequence R, $C\left(k\right)={\left|\underset{n=1}{\overset{n}{\mathrm{\&Sigma;}}}\stackrel{\&OverBar;}{r}\right[k+4(n-1)\left]B^{*}\left(n\right)\right|}^2,$ K=1....2I+1, wherein sequence $R=[\stackrel{\&OverBar;}{r}\left(1\right),...,\stackrel{\&OverBar;}{r}(52+2I)]=[r(P-I),...,r(P+52+I)],$ R () is for receiving the sampled value of signal;

Ask sequence number K corresponding to C (k) maximum, obtain the offset Δ T=K-I-1 of relative estimated position P, determine thus the new original position Q of BAKER code ₁, its position is Q ₁=P+ Δ T, this position is near timing position.

Preferably, in the described step e, the detailed process of carrying out smart frequency deviation estimation and compensation is:

E1, mobile sliding window position are to the original position of the leader block that is used for channel estimating, and the length of described sliding window is W Baud Length, i.e. 4W sampled point, W=128;

E2, to choose continuous length be that the data segment of 4W sampled point is observation data R ₁, the 4W sampled point is moved to data transfer direction in the position that sliding window is long, and the data segment of choosing continuous length and be the 4W sampled point is observation data R ₂

E3, obtain R ₁, R ₂Correlation P ₁, to the correlation P that obtains ₁Ask argument to obtain X;

E4, X is obtained smart frequency deviation estimated value Δ f and compensation divided by the product of correlation length W, 2 π and element duration T, compensation formula is r (n)=r (n) e ^{-j2 π n Δ f/T}

Preferably, in the described step F, carry out that synchronization error lock is judged and the detailed process of correcting as:

F1, after finishing preceding step, obtain for receive data corresponding to the unique word sequence of channel estimating by the smart timing position that obtains, it is carried out 4 times of down-sampled sequences that obtain The length of unique word sequence that is used for channel estimating is identical with the length of the data block of transmitting terminal transmission, is the M code element, therefore

It also is the M Baud Length;

F2, to described sequence

Carry out Fourier transform and obtain R _Fft, calculate channel estimation results H=R _Fft/ U _Fft, U _FftFourier transform value for described unique word sequence for channel estimating;

F3, described channel estimation results H is carried out inverse fourier transform obtain h, the hunting zone is defined as rear L the code element of h, wherein L is the length of Cyclic Prefix in the data block that sends of transmitting terminal;

F4, first channel gain of search is greater than the multipath component of default channel gain thresholding in described hunting zone, if search, then block synchronization error lock appears in expression, obtains the position Q of this multipath component in the hunting zone ₂, with the opposite direction skew M+1-Q of sync bit towards transfer of data ₂Individual code element.

Preferably, described Frame is made of leader block and several data blocks;

Described leader block is comprised of two lead data pieces, and first lead data piece comprises continuous 16 L _BThe sequence [UW2, UW2 ,-UW2, UW2] that the unique word sequence UW2 of the unique word sequence UW1 of Baud Length, 4 J Baud Lengths synthesizes, the data block of 16 Baud Lengths, the data block of this 16 Baud Length comprises the BARKER code of 3 invalid codes and 13 Baud Lengths; Second lead data piece is comprised of the unique word sequence UW3 of Cyclic Prefix and two W Baud Lengths, and this Cyclic Prefix is the rear L position of UW3 sequence;

Described data block is comprised of useful data and Cyclic Prefix, and this Cyclic Prefix is the unique word sequence C P of L Baud Length, and useful data is the N Baud Length, and each data block is the M=L+N Baud Length.

Preferably, described L _B=8, J=20, W=128, L=32, M=256, described BAKER code are the sequence B of 13 Baud Lengths, B=[B (1) ... ..B (13)]=[1-1-1-1-1 1 1-1-1 1-1 1-1].

Beneficial effect of the present invention is as follows:

Technical scheme of the present invention for SC-FDE system in the prior art to the characteristics of timing offset sensitivity, distinctive preamble code data and structure have been designed, estimation and the compensation of carrier wave frequency deviation and time delay have been carried out accurately at synchronous acquisition phase, thereby comparatively accurate carrier wave frequency deviation and timing information have been obtained, determined more accurately the original position of Fourier transform window, having utilized channel estimation results to carry out the mistake lock at acquisition phase to track phase judges and correction, has eliminated the possibility of synchronization error lock in the real process.Adopt technical scheme of the present invention not only can alleviate the complexity of synchronous tracking, and can greatly improve synchronization accuracy and the stability of whole system, its clear in structure, well arranged, realize simply being easy to encapsulation.

Description of drawings

Fig. 1 is the structural representation of the data block that the SC-FDE system adopts in the frame transmission course among the present invention;

Fig. 2 is the structural representation of the leader block that the SC-FDE system adopts in the frame transmission course among the present invention;

Fig. 3 is the realization flow figure of the catching method of carrier wave frequency deviation in the SC-FDE of the present invention system and time delay;

Fig. 4 carries out frame to the data that receive in the method for the invention to arrive the flow chart that detects;

Fig. 5 carries out the flow chart that thick frequency deviation is estimated and compensated to the Frame that receives in the method for the invention;

Fig. 6 is the flow chart that carries out thick timing estimation in the method for the invention;

Fig. 7 is in conjunction with maximum likelihood searching algorithm and O﹠amp in the method for the invention; M timing estimation algorithm carries out the flow chart of smart timing estimation;

Fig. 8 carries out the flow chart that smart frequency deviation is estimated and compensated in the method for the invention;

Fig. 9 carries out the flow chart that synchronization error lock is judged and corrected in the method for the invention.

Embodiment

The SC-FDE system is the system of transmitting frame by frame, and every frame is made of leader block and several data blocks.See also Fig. 1, this figure is the structural representation of the data block that the SC-FDE system adopts in the frame transmission course among the present invention, and as seen from the figure, each data block is comprised of useful data and Cyclic Prefix, Cyclic Prefix is taken as the unique word sequence C P of L Baud Length, useful data is the N Baud Length, and each data block is the vectorial s of M=L+N Baud Length, and s is by after the transmission, when receiving terminal is the V sampling, its corresponding sampled signal stream is r=[r (1) ..., r (VM)].See also Fig. 2, this figure is the structural representation of the leader block that the SC-FDE system adopts in the frame transmission course among the present invention, and as seen from the figure, the leader block in the SC-FDE system-frame is comprised of two leader block data blocks, and first lead data piece comprises continuous 16 L _BThe sequence [UW2, UW2 ,-UW2, UW2] that the unique word sequence UW2 of the unique word sequence UW1 of Baud Length, 4 J Baud Lengths forms, the data block of 16 Baud Lengths, this data block comprise the BARKER code of 3 invalid codes and 13 Baud Lengths; Second lead data piece is comprised of the unique word sequence UW3 sequence of Cyclic Prefix and two W Baud Lengths, and the Cyclic Prefix in second lead data piece is the rear L position of UW3 sequence.

In the frame structure of above-mentioned SC-FDE system, LB=8, J=20, W=128, L=32, M=256.The sequence that described BAKER code is 13 Baud Lengths [1-1-1-1-1 1 1-1-1 1-1 1-1].The unique word sequence that relates in this frame structure is generally chu sequence, frank-zadaff sequence or the PN sequence of IEEE802.16a standard code.All elect all unique word sequences that relate in the frame structure as the Chu sequence in the embodiment of the invention, its real part and imaginary part are respectively:

I(k)＝cos(θ(k))，Q(k)＝sin(θ(k))，0≤k＜L

θ(k)＝πk ²/L，0≤k＜L

Be that UW1 is the Chu sequence of 8 Baud Lengths, UW2 is the Chu sequence of 20 Baud Lengths, and UW3 is the Chu sequence of 128 Baud Lengths, and CP is the Chu sequence of 32 Baud Lengths.

Adopt the reason of specific lead data structure as follows among the present invention:

The BAKER code has strong autocorrelation performance, length be 13 BAKER code the strongest at auto-correlation position energy be 13, and the energy of all the other positions all is 1, even just because of this point so that there is the sampled point that also can accurately search under certain frequency deviation and the weak condition near timing position.The invalid code that increases is in order to ensure the strong correlation in the hunting zone.

See also Fig. 3, this figure is the realization flow figure of the catching method of carrier wave frequency deviation in the SC-FDE of the present invention system and time delay, and the sample rate of receiving terminal is 4 samplings, and it mainly comprises step:

Step 1, the data that receive are carried out frame arrive to detect, when detecting the valid data frame, determine the frame head position of Frame;

Step 2, the described frame head position of determining carried out to the Frame that receives as original position thick frequency deviation is estimated and compensation;

Step 3, carry out thick timing estimation, determine the original position of data block, timing offset is locked within the code element deviation;

Step 4, in conjunction with maximum likelihood searching algorithm and O﹠amp; M timing estimation algorithm carries out smart timing estimation, obtains accurate symbol timing information;

Step 5, the smart frequency deviation of carrying out are estimated and compensation;

Step 6, carry out that synchronization error lock is judged and correct.

The below gives further detailed explanation to the specific implementation process of above-mentioned each step.

See also Fig. 4, this figure carries out frame to the data that receive in the method for the invention to arrive the flow chart that detects, and comprises the steps:

Step 10, the original position point d of sliding window is set to d=1;

Step 11, obtain continuously two window length take d as original position and be the sampled value of 4G sampled point, R _i=(r (d+4 (i-1) G) ..., r (d+4iG-1)) ^T, i=1,2, G=32 wherein, the Baud Length that expression is obtained, r (n), { n=1,2,3.......} represent to receive signal sampling value;

Step 12, determine R ₁, R ₂Correlation P (d), $P\left(d\right)={R_1}^{\mathrm{<figure-callout\; id="2"\; label="H\; R"\; filenames="H2008102278779E00051.png"\; state="\{\{state\}\}">H</figure-callout>}}{R}_{}{}_2=\underset{k=d}{\overset{d+4G-1}{\mathrm{\&Sigma;}}}{r}^{*}\left(k\right)r(k+4G);$

Step 13, determine the average energy of two windows getting access to $R\left(d\right)=\frac{1}{2}\underset{k=d}{\overset{d+8G-1}{\mathrm{\&Sigma;}}}{\left|r\left(k\right)\right|}^2;$

Step 14, calculating $M_A\left(d\right)=\frac{{\left|P\left(d\right)\right|}^2}{{\left(R\left(d\right)\right)}^2};$

Step 15, judgement M _A(d) whether more than or equal to predetermined threshold value M _A=0.7, if, execution in step 16, otherwise, execution in step 17;

Step 16, stop mobile sliding window, the sliding window original position point d of this moment is the frame head position of Frame;

Step 17, with the mobile step-length of the original position point of mobile sliding window, be about to d=d+1, return step 11.

See also Fig. 5, this figure carries out the flow chart that thick frequency deviation is estimated and compensated to the Frame that receives in the method for the invention, comprises the steps:

Step 20, the frame head location positioning that above-mentioned frame is arrived the Frame that obtains in detecting are the original position of sliding window, and the sampled signal of establishing after this is r (t), t=1,2.....The length of setting sliding window is N _B* L _BIndividual code element, i.e. 4N _B* L _BIndividual sampled point, wherein, related interval L _B=8 code elements, N _B=4, initial count value m is 0.

Step 21, m is added 1.

Step 22, judge that m is whether less than default mobile number of times threshold value M _B, M _B=6, if, execution in step 23, otherwise, execution in step 26;

Step 23, to begin to choose continuously length from the original position of sliding window be 4N _B* L _BThe data segment of sampled point is observation data R _m, with sliding window mobile 4L backward _BAfter the sampled point length, beginning to choose continuously length from the original position of sliding window is 4N _B* L _BThe data segment of sampled point is observation data R _M+1Wherein, R _m=[r (1+4 (m-1) L _B) ..., r (4 (N _B+ m-1) L _B)], m={1,2.....M _B, the signal sampling value of r (n) expression receiving terminal;

Described observation data R _mWith observation data R _M+1Related interval be the interval L of two segment data original positions _BCode element; Described observation data R _mWith observation data R _M+1Correlation length be its data segment, length separately, i.e. 4N _B* L _BBaud Length.

Step 24, calculate R _mAnd R _M+1Correlation P _m, $P_m=R_m{R}_{m+1}^H=\underset{n=1+4(m-1)L_B}{\overset{4(N_B+m-1)L_B}{\mathrm{\&Sigma;}}}r\left(n\right)r^{*}(n+4L_B);$

Step 25, determine correlation P _mArgument X _mAnd give record, return step 21.

Step 26, to all have recorded in the step 25 M _BIndividual X _mValue is weighted summation, and formula is: $X=\underset{m=1}{\overset{M_B}{\mathrm{\&Sigma;}}}{\mathrm{\&alpha;}}_m{X}_m,$ Wherein, α _mBe weight coefficient.

Step 27, with the result of the weighted sum that obtains in the step 26 divided by related interval L _B, 2 π and code element transmission time T product obtain frequency deviation estimated value Δ f, i.e. Δ f=X/ (2 π L _BT);

Step 28, each sampled value of receive data is carried out compensate of frequency deviation: r (n)=r (n) e ^{-j2 π n Δ f/T}

The above-mentioned specific lead data structure of flow process utilization of the Frame that receives being carried out thick frequency deviation estimation and compensation, selection long by sliding window and sliding window displacement is related interval and two factors of correlation length with the Factorization that affects simultaneously capturing frequency deviation scope and estimated accuracy in the prior art, because related interval has determined the capturing frequency deviation scope, correlation length has determined frequency offset estimation accuracy, therefore when frequency deviation region is larger, reduce related interval by the displacement that reduces sliding window, to reach the effect that increases the capturing frequency deviation scope, increase correlation length by increasing sliding window length again simultaneously, fall low noise impact with this, improved frequency offset estimation accuracy.Thereby solved the contradiction that mutually restricts between existing frequency bias capture method frequency deviation estimated accuracy and the capture range, estimate weighted sum by the frequency deviation that diverse location is obtained simultaneously, slackened the dark weak impact on frequency offset estimation accuracy under the wireless mobile channel.

See also Fig. 6, this figure is the flow chart that carries out thick timing estimation in the method for the invention, comprises the steps:

Step 30, the original position point d of sliding window is set to 0, with mark position 0.

Step 31, the d value is added 1, wherein the position of d=1 be through thick frequency deviation estimate and compensation after, this moment the sampled point place the position.

Step 32, the original position take d as sliding window are obtained continuously 4 window length and are the sampled value R of 4J (J=20) sampled point _i, R _i=(r (d+4 (i-1) J) ..., r (d+4iJ-1)) ^T, i=1,2,3,4, r (t) is for receiving signal sampling value.

Measure value M corresponding to [+UW2 ,+UW2 ,-UW2 ,+UW2] under step 33, the calculating current location d _C(d).

In this step, calculate M _C(d) concrete steps are as follows:

1) calculates R ₁ ^HR ₂, R ₁ ^HR ₃, R ₁ ^HR ₄, R ₂ ^HR ₃, R ₂ ^HR ₄, R ₃ ^HR ₄And R _sValue, R _i ^HR _jIt is the correlation of two sections sequences, its account form be the conjugation of 4J sampled value of i the sequence sampled value corresponding with j sequence product cumulative and, try to achieve first the product of the conjugation of 4J sampled value of i the sequence sampled value corresponding with j sequence, then to the cumulative summation of this 4J product value.R _sBe the average energy value of sliding window, its account form is again divided by 4 with cumulative summation behind all sampled value moulds in 4 windows square;

2) carry out computing module-square, namely calculate | R ₁ ^HR ₂-R ₃ ^HR ₄-R ₂ ^HR ₃| ², | R ₂ ^HR ₄-R ₁ ^HR ₃| ², | R ₁ ^HR ₄| ²And | R _s| ²Value;

3) computing formula $M_C\left(d\right)=\frac{{|{R_1}^H{R}_2-{R_3}^H{R}_4-{R_2}^H{R}_3|}^2+{|{R_2}^H{R}_4-{R_1}^H{R}_3|}^2+{\left|{R_1}^H{R}_4\right|}^2}{14{\left|R_s\right|}^2}$

Step 34, judgement M _C(d) whether estimate threshold value M more than or equal to default _C, M _CIf=0.4, execution in step 35, otherwise, execution in step 36.

Step 35, this M that calculates (d) value is added among the Measure Sequences Metric, Measure Sequences Metric=[M (n1) ... .., M (n2)] (n1 is M _C(d) by less than thresholding M _CBe converted into greater than thresholding M _CThe position, n2 is M _C(d) by greater than thresholding M _CBe converted into less than thresholding M _CThe position), with mark position 1, return step 31 simultaneously.

Step 36, with this M that calculates _C(d) value is rejected, execution in step 37.

Step 37, judge whether current mark bit is 1, if, execution in step 38, otherwise, step 31 returned.

Step 38, from Measure Sequences Metric, search out the maximum of estimating wherein, the location point that it is corresponding

Be the synchronous position of piece that obtains through thick timing estimation.

Above-mentioned flow process of carrying out thick timing estimation has been adjusted Measure Algorithm by nonlinear characteristic, removed and opened the radical sign computing, reduced the realization resource, and utilize nonlinear characteristic to widen gap between secondary lobe and the main lobe, more be conducive to maximizing in search procedure, in the peaked process of traversal search, utilize first a specific thresholding to dwindle the hunting zone, and then in this scope maximizing, greatly reduce the realization difficulty, very high practical value is arranged.By having adjusted the normalization formula, use the energy of the whole training symbol that receives to come the energy of substitute symbol various piece as the denominator of timing metric simultaneously, further improved net synchronization capability, so that system can work under non-constant enveloped modulation signal.Further, above-mentioned flow process has improved the synchronous precision of piece by the mode of multiple sampling.

See also Fig. 7, this figure is in conjunction with maximum likelihood searching algorithm and O﹠amp in the method for the invention; M timing estimation algorithm carries out the flow chart of smart timing estimation, mainly comprises the steps:

The sampling point position that step 41, the thick timing estimation of basis are determined obtains the estimated value P of the original position of BAKER code in the lead data.

Step 42, determined that by estimated value P the hunting zone of the original position of BAKER code is: centered by the P point, front and back each I sampled point.Determined with the BAKER code by the hunting zone and to make the needed receiving sequence R of correlation computations;

Described BAKER code is the sequence B of 13 Baud Lengths, B=[B (1) ... ..B (13)]=[1-1-1-1-11 1-1-1 1-1 1-1].

Described R is: $R=[\stackrel{\&OverBar;}{r}\left(1\right),...,\stackrel{\&OverBar;}{r}(52+2I)]=[r(P-I),...,r(P+52+I)],$ 2I+13*4=2I+52 sampled point altogether needs the precision of the sampled point P that determines synchronously according to piece to select appropriate I value in the reality, r () is the sampled value of reception signal.

The mould square C (k) of the correlation of step 43,2I+1 BAKER code sequence B of calculating and receiving sequence R, $C\left(k\right)={\left|\underset{n=1}{\overset{n}{\mathrm{\&Sigma;}}}\stackrel{\&OverBar;}{r}\right[k+4(n-1)\left]B^{*}\left(n\right)\right|}^2,$ k＝1....2I+1。

If I value selected in the step 42 is 4, then need to calculate

With the mould square C (1) of the correlation of sequence B, With the mould square C (2) of the correlation of sequence B ... .., calculate

Mould square C (9) with the correlation of sequence B.

Step 44, ask sequence number K corresponding to maximum of C (k), obtain the offset Δ T=K-I-1 of relative estimated position P, determine thus the new original position Q of BAKER code ₁, its position is Q ₁=P+ Δ T, this position is adjusted the original position of Fourier transform window near timing position with this.

In this step, the value of described position offset Δ T be on the occasion of the time, the expression timing position need to be offset to data transfer direction, when the value of described position offset Δ T was negative value, the expression timing position need to be to the skew of the opposite direction of transfer of data.

After step 41 to step 44 is finished, can access near the sampled point of timing position, timing offset can be locked onto the deviation of a sampled point, i.e. the deviation of 1/4 code element.The below utilizes the O﹠amp after improving; The M algorithm further estimates to remain timing offset, the purpose of estimating to reach accurate timing.Detailed process is as follows:

Step 45, with Q ₁Be original position, data intercept section [r (Q ₁) ..., r (Q ₁+ 4M-1)], with symbol S=[s (1) ..., s (4M)] and expression, i.e. S=[s (1) ..., s (4M)]=[r (Q ₁) ..., r (Q ₁+ 4M-1)], wherein M is the code element number in the data segment of intercepting, the length of the data block that sends with transmitting terminal is identical;

Step 46, S is undertaken obtaining behind the bandpass filtering by the first band pass filter

${\stackrel{\&RightArrow;}{S}}_1\left(z\right)=[s_1\left(1\right),...,s_1\left(4M\right)].$

Step 47, calculating

Z=1.....4M, F () are nonlinear function, original O﹠amp; What the M algorithm adopted is quadratic nonlinearity

Z=1.....4M among the present invention is adjusted into it $F\left(\right|{\stackrel{\&RightArrow;}{S}}_1\left(z\right)\left|\right)={\left|s_1\left(z\right)\right|}^2+\left|s_1\left(z\right)\right|,z=1.....4M.$

Step 48, with what obtain in the step 17

Undertaken obtaining behind the bandpass filtering by the second band pass filter

${\stackrel{\&RightArrow;}{S}}_2\left(z\right)=[s_2\left(1\right),...,s_2\left(4M\right)].$

Step 49, calculating residue timing offset

Make V=4 represent that the sample rate of receiving terminal is 4 samplings, namely

Wherein T is the code element transmission time.

Step 50, basis

Proofread and correct the signal after obtaining regularly by the interpolation mould.

Above-mentioned the first band pass filter is that centre frequency is the band pass filter of 1/2T, and above-mentioned the second band pass filter is that centre frequency is the band pass filter of 1/T.

Total timing estimation value of determining by the flow process of above-mentioned smart timing estimation is

O﹠amp; Have the cyclic stationary characteristic if the basic principle of M algorithm is the signal of matched filtering output through the signal after the Nonlinear Processing, just have so spectral line to produce in the symbol rate position.This just can calculate spectral line with the Fourier series expansion.Adopt the O﹠amp after improving in the flow process of above-mentioned smart timing estimation; The M algorithm further estimates to remain timing offset, the O﹠amp after the improvement; The non-linear form of M algorithm not only can not reduce the shake variance under the non-weak channel, and can reduce the shake variance under the wireless mobile channel, improves estimated accuracy.O﹠amp after improvement; The input of M algorithm carries out the bandpass filtering first time to signal, carries out the bandpass filtering second time after non-linear, can reduce the impact of noise on estimating, reduces shake, improves estimated accuracy.

See also Fig. 8, this figure carries out the flow chart that smart frequency deviation is estimated and compensated in the method for the invention, specifically comprise the steps:

The original position of the leader block that is used for channel estimating is arrived in step 60, mobile sliding window position, and the length of described sliding window is W Baud Length, is 4W sampled point, W=128;

Step 61, to choose continuous length be that the data segment of 4W sampled point is observation data R ₁, the 4W sampled point is moved to data transfer direction in the position that sliding window is long, and the data segment of choosing continuous length and be the 4W sampled point is observation data R ₂

Step 62, obtain R ₁, R ₂Correlation P ₁, ${\mathrm{<figure-callout\; id="1"\; label="P"\; filenames="H2008102278779E00021.png,H2008102278779E00031.png"\; state="\{\{state\}\}">P</figure-callout>}}_1=R_1{R}_2^H=\underset{t=1}{\overset{2W}{\mathrm{\&Sigma;}}}r\left(t\right)r^{*}(t+2W);$

Step 63, the correlation P to obtaining ₁Ask argument to obtain X, X=angle (P ₁);

Step 64, X is obtained smart frequency deviation estimated value Δ f and compensation divided by the product of correlation length W, 2 π and element duration T, Δ f=X/ (2 π WT), compensation formula are r (n)=r (n) e ^{-j2 π n Δ f/T}

See also Fig. 9, this figure carries out the flow chart that synchronization error lock is judged and corrected in the method for the invention, specifically comprise the steps:

Step 71, after finishing preceding step, obtain for receive data corresponding to the UW sequence of channel estimating by the smart timing position that obtains, and it carried out 4 times of down-sampled sequences that obtain

The length of UW sequence that is used for channel estimating is identical with the length of the data block of transmitting terminal transmission, is the M code element, therefore

It also is the M Baud Length.

Step 72, to described sequence

Carry out the conversion of M point Fourier and obtain R _Fft, R _Fft=[R (1) ..., R (M)].

Step 73, calculating channel estimation results H=[H (1) ..., H (M)]=R _Fft/ U _Fft, U wherein _FftIt is the Fourier transform value that is used for the UW sequence of channel estimating among the present invention.

Step 74, channel estimation results H is carried out inverse fourier transform obtain h, h=[h (1) ..., h (M)].

Step 75, the hunting zone is defined as rear L the code element of h, namely the hunting zone be [h (M-L+1) ..., h (M)], wherein L is the length of Cyclic Prefix (CP) in the data block that sends of transmitting terminal.

Step 76, judge at hunting zone [h (M-L+1) that step 75 is determined, ..., h (M)] in can search gain greater than the multipath component of default channel gain thresholding a=0.1, if can search, then block synchronization error lock appears in expression, execution in step 77, otherwise block synchronization error lock, execution in step 78 do not appear in expression.

Step 77, obtain first gain of searching greater than the position Q of multipath component in the hunting zone of default channel gain thresholding a=0.1 ₂, Q ₂∈ [M-L+1, M] is with the opposite direction skew M+1-Q of sync bit towards transfer of data ₂Individual code element.

Step 78, need not to carry out the sync bit adjustment.

The above-mentioned flow process of carrying out the synchronization error lock judgement and correcting has been utilized the characteristic of channel estimation results on time domain, side-play amount when the different characteristic of channel estimation results on time domain in the time of the channel estimation results when locking by mistake occurring and accurate synchronization determined whether to occur wrong lock and wrong lock occurred, solved fast and effectively in the prior art and to have occurred in the transmission process dark when weak when article one footpath that arrives, the problem that the symbol original position of determining can be offset to the direction of transfer of data, its implementation procedure is simple, convenient.

Obviously, those skilled in the art can carry out various changes and modification to the present invention and not break away from the spirit and scope of the present invention.Like this, if of the present invention these are revised and modification belongs within the scope of claim of the present invention and equivalent technologies thereof, then the present invention also is intended to comprise these changes and modification interior.

Claims (12)

1. the carrier wave frequency deviation in the single-carrier frequency domain equalization system and the catching method of time delay, the sample rate of receiving terminal is 4 samplings, it is characterized in that, comprises step:

A, the data that receive are carried out frame arrive to detect, when detecting the valid data frame, determine the frame head position of Frame;

B, the described frame head position of determining carried out to the Frame that receives as original position thick frequency deviation is estimated and compensation;

C, carry out thick timing estimation, determine the original position of data block, timing offset is locked within the code element deviation;

D, in conjunction with maximum likelihood searching algorithm and O﹠amp; M timing estimation algorithm carries out smart timing estimation, obtains accurate symbol timing information;

E, carry out that smart frequency deviation is estimated and compensation;

F, carry out that synchronization error lock is judged and correct.

2. the method for claim 1 is characterized in that, describedly the data that receive are carried out the implementation procedure that frame arrive to detect is:

A1, the original position point d of sliding window is set to d=1;

A2, obtain continuously two window length take d as original position and be the sampled value of 4G sampled point, R _i=(r (d+4 (i-1) G) ..., r (d+4iG-1)) ^T, i=1,2, G=32 wherein, the Baud Length that expression is obtained, r (n), { n=1,2,3.......} represent to receive signal sampling value;

A3, determine R ₁, R ₂Correlation P (d), $P\left(d\right)={R_1}^H{R}_2=\underset{k=d}{\overset{d+4G-1}{\mathrm{\&Sigma;}}}{r}^{*}\left(k\right)r(k+4G);$ A4, determine the average energy of two windows getting access to

A5, calculating

Judge M _A(d) whether more than or equal to predetermined threshold value M _AIf,, then stopping mobile sliding window, the sliding window original position point of this moment is the frame head position of Frame, otherwise, the original position point of mobile sliding window is added 1, return steps A 2.

3. method as claimed in claim 2 is characterized in that, described predetermined threshold value is M _A=0.7.

4. the method for claim 1 is characterized in that, among the described step B, the detailed process of the Frame that receives being carried out thick frequency deviation estimation and compensation is:

The frame head location positioning of B1, the Frame that obtains during frame arrive detected is the original position of sliding window, and the length of sliding window is N _B* L _BBaud Length, i.e. 4N _B* L _BIndividual sampled point, initial count value m are 0;

B2, m is added 1, judge that whether m is less than default mobile number of times threshold value M _BIf,, execution in step B3, otherwise, execution in step B5;

B3, begin to choose continuously 4N from the original position of sliding window _B* L _BThe observation data R of individual sampled point length _m, the sliding window original position is moved 4L to data transfer direction _BAfter the individual sampled point length, the original position from sliding window begins to choose continuously 4N again _B* L _BThe observation data R of individual sampled point length _M+1

Wherein, R _m=[r (1+4 (m-1) L _B) ..., r (4 (N _B+ m-1) L _B)], m={1,2.....M _B, the signal sampling value of r (n) expression receiving terminal;

B4, determine R _mAnd R _M+1Correlation P _m, $P_m=R_m{R}_{m+1}^H=\underset{n=1+4(m-1)L_B}{\overset{4(N_B+m-1)L_B}{\mathrm{\&Sigma;}}}r\left(n\right)r^{*}(n+4L_B)$ Then determine correlation P _mArgument X _mAnd give record, return step B2;

B5, the M to having recorded _BIndividual X _mValue is weighted summation, with the result of weighted sum divided by related interval L _B, 2 π and code element transmission time T product obtain frequency deviation estimated value Δ f;

B6, each sampled value of receive data is carried out compensate of frequency deviation: r (n)=r (n) e-j ^{2 π n Δ f/T}Wherein,

Described related interval L _B=8 code elements, N _B=4, default mobile number of times threshold value M _B=6.

5. the method for claim 1 is characterized in that, among the described step C, the detailed process of thick timing estimation is:

C1, the original position point d of sliding window is set to 0, with mark position 0;

C2, the d value is added 1, d=1 the position for estimate through thick frequency deviation and compensation after this moment the sampled point place the position, the sampled value R that it is 4J sampled point as J Baud Length all that the original position take d as sliding window is obtained 4 window length continuously _i, R _i=(r (d+4 (i-1) J)., r (d+4iJ-1)) ^T, i=1,2,3,4;

Measure value M under C3, the calculating current location d _C(d), $M_C\left(d\right)=\frac{|{R_1}^H{R}_2-{R_3}^H{R}_4-{R_2}^H{R}_3{|}^2+|{R_2}^H{R}_4-{R_1}^H{R}_3{|}^2+|{R_1}^H{R}_4{|}^2}{14|R_s{|}^2},$ Wherein, R _i ^HR _jBe the conjugation of 4J sampled value of i the sequence sampled value corresponding with j sequence product cumulative and, R _sThe average energy value for sliding window;

C4, judgement M _C(d) whether estimate threshold value M more than or equal to default _CIf,, the M that then this is calculated _C(d) value is added in the Measure Sequences, with mark position 1, returns step C2; Otherwise, execution in step C5;

C5, judge whether current mark bit is 1, if estimate maximum, the location point that it is corresponding in the search Measure Sequences

Be the synchronous position of piece, otherwise, step C2 returned.

6. method as claimed in claim 5 is characterized in that, the described default threshold value M that estimates _C=0.4.

7. the method for claim 1 is characterized in that, among the described step D, the detailed process of smart timing estimation is:

D1, utilize based on the maximum likelihood searching algorithm of BAKER code near the sampled point Q of timing position ₁And Q ₁Offset Δ T;

D2, with Q ₁Be original position, data intercept section [r (Q ₁) ..., r (Q ₁+ 4M-1)], with symbol S=[s (1) ..., s (4M)] and expression, i.e. S=[s (1) ..., s (4M)]=[r (Q ₁) ..., r (Q ₁+ 4M-1)], wherein M is the code element number in the data segment of intercepting, the length of the data block that sends with transmitting terminal is identical, and S is undertaken obtaining behind the bandpass filtering by the first band pass filter

D3, calculating

Z=1.....4M, F () is nonlinear function, with what obtain Undertaken obtaining behind the bandpass filtering by the second band pass filter

${\stackrel{\&RightArrow;}{S}}_2\left(z\right)=[s_2\left(1\right),...,s_2\left(4M\right)];$

D4, calculating residue timing offset

Make V=4 represent that the sample rate of receiving terminal is 4 samplings, namely

Total timing estimation value is

Wherein T is the code element transmission time.

8. method as claimed in claim 7 is characterized in that, described step D1 specifically comprises step:

The sampling point position of determining synchronously according to piece obtains the estimated value P of the original position of BAKER code in the lead data;

The hunting zone of being determined the original position of BAKER code by estimated value P is: centered by the P point, front and back each I sampled point is determined with the BAKER code by the hunting zone and to be made the needed receiving sequence R of correlation computations;

Calculate the mould square C (k) of the correlation of 2I+1 BAKER code sequence B and receiving sequence R, $C\left(k\right)={\left|\underset{n=1}{\overset{n}{\mathrm{\&Sigma;}}}\stackrel{\&OverBar;}{r}\right[k+4(n-1)\left]B^{*}\left(n\right)\right|}^2,$ K=1....2I+1, wherein sequence $R=[\stackrel{\&OverBar;}{r}\left(1\right),...,\stackrel{\&OverBar;}{r}(52+2I)]=[r(P-I),...,r(P+52+I)],$ R () is for receiving the sampled value of signal;

Ask sequence number K corresponding to C (k) maximum, obtain the offset Δ T=K-I-1 of relative estimated position P, determine thus the new original position Q of BAKER code ₁, its position is Q ₁=P+ Δ T, this position is near timing position.

9. the method for claim 1 is characterized in that, in the described step e, the detailed process of carrying out smart frequency deviation estimation and compensation is:

E1, mobile sliding window position are to the original position of the leader block that is used for channel estimating, and the length of described sliding window is W Baud Length, i.e. 4W sampled point, W=128;

E2, to choose continuous length be that the data segment of 4W sampled point is observation data R ₁, the 4W sampled point is moved to data transfer direction in the position that sliding window is long, and the data segment of choosing continuous length and be the 4W sampled point is observation data R ₂

E3, obtain R ₁, R ₂Correlation P ₁, to the correlation P that obtains ₁Ask argument to obtain X;

E4, X is obtained smart frequency deviation estimated value Δ f and compensation divided by the product of correlation length W, 2 π and element duration T, compensation formula is r (n)=r (n) e ^{-j2 π n Δ f/T}

10. the method for claim 1 is characterized in that, in the described step F, carry out that synchronization error lock is judged and the detailed process of correcting as:

F1, after finishing preceding step, obtain for receive data corresponding to the unique word sequence of channel estimating by the smart timing position that obtains, it is carried out 4 times of down-sampled sequences that obtain

The length of unique word sequence that is used for channel estimating is identical with the length of the data block of transmitting terminal transmission, is the M code element, therefore It also is the M Baud Length;

F2, to described sequence Carry out Fourier transform and obtain R _Fft, calculate channel estimation results H=R _Fft/ U _Fft, U _FftFourier transform value for described unique word sequence for channel estimating;

F3, described channel estimation results H is carried out inverse fourier transform obtain h, the hunting zone is defined as rear L the code element of h, wherein L is the length of Cyclic Prefix in the data block that sends of transmitting terminal;

F4, first channel gain of search is greater than the multipath component of default channel gain thresholding in described hunting zone, if search, then block synchronization error lock appears in expression, obtains the position Q of this multipath component in the hunting zone ₂, with the opposite direction skew M+1-Q of sync bit towards transfer of data ₂Individual code element.

11. the method for claim 1 is characterized in that, described Frame is made of leader block and several data blocks;

Described leader block is comprised of two lead data pieces, and first lead data piece comprises continuous 16 L _BThe sequence [UW2, UW2 ,-UW2, UW2] that the unique word sequence UW2 of the unique word sequence UW1 of Baud Length, 4 J Baud Lengths synthesizes, the data block of 16 Baud Lengths, the data block of this 16 Baud Length comprises the BARKER code of 3 invalid codes and 13 Baud Lengths; Second lead data piece is comprised of the unique word sequence UW3 of Cyclic Prefix and two W Baud Lengths, and this Cyclic Prefix is the rear L position of UW3 sequence;

Described data block is comprised of useful data and Cyclic Prefix, and this Cyclic Prefix is the unique word sequence C P of L Baud Length, and useful data is the N Baud Length, and each data block is the M=L+N Baud Length.

12. method as claimed in claim 12 is characterized in that, described L _B=8, J=20, W=128, L=32, M=256, described BAKER code are the sequence B of 13 Baud Lengths, B=[B (1) ... ..B (13)]=[1-1-1-1-1 1 1-1-1 1-1 1-1].

Images