CN104836652A - Space time block coding (STBC) multiple input multiple output-orthogonal frequency division multiplexing (MIMO-OFDM) system time frequency synchronization new method under low signal to noise ratio - Google Patents

Abstract

The invention discloses a space time block coding (STBC) multiple input multiple output-orthogonal frequency division multiplexing (MIMO-OFDM) system time frequency synchronization new method under low signal to noise ratio. The method is characterized by being provided for solving the problem that the STBC MIMO-OFDM system is hard to get accurate time frequency synchronization under the condition of low signal to noise ratio. According to the method, firstly, a group of complex complementary sequence sets and an orthogonal matrix are used for constructing a complex orthogonal complementary sequence, and the sequence is extended through conversion of the orthogonal matrix, so as to obtain the complex orthogonal complementary sequence of appropriate length. To improve the autocorrelation and mutual correlation of the sequence, a training sequence is constructed through the conjugation negation characteristic of the sequence and is superposed on a data signal. At a receiving end, a good synchronization signal is obtained through association of the training sequence and the data signal, and the integer frequency offset estimation range can reach Epsilon between -N/4 and N/4. Under low signal to noise ratio, compared with the conventional methods, the synchronization method can obtain more accurate time synchronization and larger frequency deviation range at -15dB, and has lower computation complexity.

Description

The synchronous new method of STBC MIMO-OFDM system time frequency under a kind of low signal-to-noise ratio

Technical field

The present invention relates to MIMO-OFDM technical field, the synchronous new method of STBC MIMO-OFDM system time frequency particularly under a kind of low signal-to-noise ratio.

Background technology

Along with the fast development of wireless communication technology, people are to the transmission rate of wireless communication technology and stability and the concrete more and more higher requirement of reliability.Existing OFDM (OFDM) technology is a kind of Multicarrier Transmission Technology, the features such as the availability of frequency spectrum is higher, anti-multipath fading that it has.Multiple-input and multiple-output (MIMO) technology is then make full use of space resources, realizes the function of the many receptions of multi-emitting, when not increasing frequency spectrum resource and transmitting antenna power, improves the capacity of channel.The system that MIMO and OFDM combines is had the features such as stable signal transmission, the availability of frequency spectrum are high, high-speed transfer rate, the requirement of wireless-transmission network of future generation can be met well, and be also the inexorable trend of later stage wireless communication technology development.Its field of mainly applying comprises: WLAN (wireless local area network) (WLAN), wireless broadband Internet access (Wimax), future generation mobile radio system (B3G/4G) etc.

In order to improve the transmission rate of wireless communication technology and stability and reliability, in STBC MIMO-OFDM system under Low SNR, MIMO-OFDM technology has the features such as stable signal transmission, the availability of frequency spectrum are high, high-speed transfer rate, and adopt STBC to encode and can realize multiple transmit antenna combined coding technology, antenna diversity can resist the decline of channel, improves the reliability of wireless communication link; Redundant information when increasing transmission empty by STBC coding, improves the Stability and dependability of wireless communication transmissions.

Under Low SNR, existing MIMO-OFDM synchronized algorithm time-frequency synchronization performance is all affected.Therefore, under Low SNR, want to obtain Time and Frequency Synchronization performance of good performance, this time-frequency synchronization is worth further investigation.The people such as Ali Rachini propose STBC MIMO ofdm system Time synchronization algorithm, adopt the good CAZAC sequence of orthogonality and Walsh-Hadamard sequence as training sequence, utilize that local training sequence is relevant to data-signal obtains good time synchronized performance.Simulation result shows that, under Low SNR, time synchronized performance is more excellent.But do not consider that frequency deviation estimates the impact on systematic function.The people such as FAN Hui-li propose MIMO-OFDM system time frequency synchronized algorithm, adopt the CAZAC sequence of conjugated structure as training sequence, utilize receiving terminal data message auto-correlation to obtain synchronizing information.Simulation result shows that the method Time and Frequency Synchronization under Low SNR is better, but the method deficiency to be frequency offset estimation range less.Documents: [1] R.N.Ali, B.D.Ali, N.Fabienne and B.Bilal.Timing synchronization method for MIMO-OFDM system using orthogonal preamble [C] .19th International Conference on Telecommunications (ICT 2012), 2012:1-6. [2] H.L.Fan, J.F.Sun, P.Yang, D.S.Li.A Robust Timing and Frequency Synchronization Algorithm for HF MIMO OFDM Systems [C] .IEEE Conference on Global Mobile Congress (GMC), these synchronous method of 2010:1-4. are all the synchronous method be inserted into by training sequence on OFDM data symbol, it is not enough, and place is the availability of frequency spectrum of reduction system and the transmitting power of reduction transmitter.

Therefore, in order to overcome the deficiency that existing MIMO-OFDM system synchronization method exists, the present invention proposes the synchronous new algorithm of STBC MIMO-OFDM system time frequency under a kind of low signal-to-noise ratio, the training sequence of this synchronous method is added on OFDM data symbol, and it effectively can improve the transmitting power of band efficiency and transmitter.

Summary of the invention

The object of the invention is to be to propose the synchronous new method of STBC MIMO-OFDM system time frequency under a kind of low noise, thus overcome the deficiency of existing MIMO-OFDM system synchronization method.The present invention proposes synchronous new method, first be design the good multiple orthogonal complement sequence of a kind of orthogonality, next utilizes the conjugation negate characteristic of this sequence to construct the training sequence made new advances, and this training sequence has good autocorrelation and more weak cross correlation, and has excellent orthogonality.Training sequence is added to a complete OFDM data symbol becomes and transmits, at receiving terminal, utilize that receiving terminal data-signal is relevant with local training sequence can obtain accurate synchronizing information, utilize receiving terminal data-signal and local training sequence related operation to ask phase place to obtain integer frequency bias estimated value.Synchronous and the Frequency Synchronization of precise time can ensure the correct demodulation of data message of transmission.

To achieve these goals, the innovative point of the synchronous new method of STBC MIMO-OFDM system time frequency under a kind of low signal-to-noise ratio of proposing of the present invention:

(1) the present invention designs a kind of training sequence based on multiple orthogonal complement sequence, and this training sequence adopts conjugation negate characteristic to improve the autocorrelation of training sequence.

(2) the new time-frequency synchronization of the present invention's proposition, Time synchronization algorithm adopts receiving terminal data message and local training sequence to obtain synchronizing information, accurate timing synchronization position is determined by maximum, realizing on timing synchronous foundation, carry out Frequency Synchronization, and larger frequency deviation region can be estimated in time domain, thus ensure the correct demodulation of the data message of transmission.

The present invention proposes the synchronous new method of STBC MIMO-OFDM system time frequency under a kind of low signal-to-noise ratio, and the antenna of this system arranges N _tindividual transmitting antenna and N _rindividual reception antenna, its concrete steps are as follows:

Step 1: multiple orthogonal complement sequences Design:

(1) a multiple complementary sequence set { A is designed ₁, B ₁, C ₁, D ₁, wherein A ₁=[1,1, j ,-j], B ₁=[1,1 ,-j, j], with c ₂=B ₁, D ₂=-A ₁, wherein a ₁conjugation negate computing ,-A ₁represent A ₁negative value, j represents imaginary symbols, j ²=-1;

(2) utilize above-mentioned multiple complementary sequence set structure row vector to be 4, column vector is the matrix E of 8 _{4 × 8}, wherein $E_{4\×8}={\left(\begin{array}{cccc}{A}_1& B_1& C_1& D_1\\ A_2& B_2& C_2& D_2\end{array}\right)}^T,$ In formula () ^trepresenting matrix transposition;

(3) orthogonal matrix is utilized $H={\left(\begin{array}{cccc}1& 1& -1& -1\\ -j& j& -j& j\end{array}\right)}^T$ Re-constructing row vector is 16, and column vector is 16 new matrix E _{16 × 16}, its matrix $E_{16\×16}={\left(\begin{array}{cccc}{E}_{4\×8}& E_{4\×8}& -E_{4\×8}& -E_{4\×8}\\ -j*E_{4\×8}& j*E_{4\×8}& -j*E_{4\×8}& j*E_{4\×8}\end{array}\right)}^T,$ In formula () ^trepresenting matrix transposition;

(4) orthogonal matrix is utilized $F=\left(\begin{array}{cc}1& -j\\ 1& j\end{array}\right)$ Re-constructing row vector is 32, and column vector is 32 new matrix E _{32 × 32}, obtain matrix $E_{32\×32}=\left(\begin{array}{cc}{E}_{16\×16}& -j*E_{16\×16}\\ E_{16\×16}& j*E_{16\×16}\end{array}\right),$

(5) above-mentioned (4) step is repeated, can extended matrix thus obtain longer orthogonal sequence;

Step 2: utilize above-mentioned multiple orthogonal complement sequences Design step, then can obtain orthogonal matrix $E_{\frac{N}{4}\×\frac{N}{4}}=\left(\begin{array}{cc}{E}_{\frac{N}{4}\×\frac{N}{4}}& -j*E_{\frac{N}{4}\×\frac{N}{4}}\\ E_{\frac{N}{4}\×\frac{N}{4}}& j*E_{\frac{N}{4}\×\frac{N}{4}}\end{array}\right),$ Pass through matrix construct one group of sequence a _i(n), wherein the span i ∈ [1, N/4] of the span n ∈ [0, N/4-1] of n, i;

As i=1, citing structure length is the sequence a of N/4 ₁n (), by sequence a ₁n () is got conjugation negate computing and is obtained new sequence b ₁n (), has:

b ₁(n)＝-a ₁*(n),n∈[0,N/4-1] (1)

In formula: N represents sub-carrier number, represent sequence a ₁() sequence gets conjugate operation;

Step 3: by sequence a ₁(n) and b ₁n () forms length is the sequence t of N/2 ₁n (), has:

$t_1\left(n\right)=\left\{\begin{array}{cc}{a}_1\left(n\right)& ,n\∈[0,N/4-1]\\ b_1\left(n\right)& ,n\∈[N/4,N/2-1]\end{array}\right.---\left(2\right)$

Step 4: by sequence t ₁n () repeats once to construct the sequence c that length is N ₁n (), has:

$c_1\left(n\right)=\left\{\begin{array}{cc}{t}_1\left(n\right),& n\∈[0,N/2-1]\\ t_1(n-N/2),& n\∈[N/2,N-1]\end{array}\right.---\left(3\right)$

Step 5: in the complete OFDM symbol that is added to by training sequence, signal indication is:

$r_j\left(n\right)=\sqrt{P(1-\mathrm{\β})}{x}_i\left(n\right)+\sqrt{\mathrm{P\β}}{c}_i\left(n\right),n=\mathrm{0,1},\·\·\·N+\mathrm{Ng}-1---\left(4\right)$

In formula: P represents the gross power of transmitter, β is power allocation factor, and its span 0 < β < 1, is defined as $\mathrm{\&beta;}={\mathrm{\&sigma;}}_c^2/({\mathrm{\&sigma;}}_c^2+{\mathrm{\&sigma;}}_x^2),$ Wherein ${\mathrm{\&sigma;}}_c^2=E\left({\left|c\left(n\right)\right|}^2\right),{\mathrm{\&sigma;}}_x^2=E\left({\left|x\left(n\right)\right|}^2\right).$

Step 6: the local training sequence utilizing receiving terminal to receive acquisition time relevant with data message is synchronous, and timing metric function expression is:

$M\left(d\right)=\frac{|P\left(d\right){|}^2}{{\left|R\left(d\right)\right|}^2}---\left(5\right)$

Wherein:

$P\left(d\right)=\underset{i=1}{\overset{N_T}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{N_R}{\mathrm{\&Sigma;}}}\underset{m=0}{\overset{1}{\mathrm{\&Sigma;}}}\underset{n=0}{\overset{N/4-1}{\mathrm{\&Sigma;}}}[r_j^{*}(d+\frac{N}{4}+\frac{N}{2}m+n)c_i(\frac{N}{4}+\frac{N}{2}m+n)+r_j^{*}(d+m+n)c_i(m+n)]---\left(6\right)$

$R\left(d\right)=\underset{i=1}{\overset{N_T}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{N_R}{\mathrm{\&Sigma;}}}\underset{n=0}{\overset{N/2-1}{\mathrm{\&Sigma;}}}{\left|r_j(d+n)\right|}^2---\left(7\right)$

In formula: Nt is the number of transmitting antenna, Nr is the number of reception antenna, and Ng is the length of Cyclic Prefix, represent that receiving terminal data message gets conjugate operation, n is the number of sampled point, c _in () represents the training sequence on transmitting antenna, d is integer value, and it represents the relative sliding position of local training sequence and receiving terminal data message, and m represents cycle-index;

Conveniently mathematical analysis, supposes that the gross power P of transmitter is 1, and training sequence has nothing to do with data message and interchannel noise.

$\begin{array}{c}B\left(P\left(d\right)\right)=B\{r_j^{*}(d+\frac{N}{4}+\frac{N}{2}m+n)c_i(\frac{N}{4}+\frac{N}{2}m+n)+r_j^{*}(d+m+n)c_i(m+n)\}\\ =B\{(\sqrt{1-\mathrm{\&beta;}}{x}_j^{*}(d+\frac{N}{4}+\frac{N}{2}m+n)+\sqrt{\mathrm{\&beta;}}{c}_i^{*}(d+\frac{N}{4}+\frac{N}{2}m+n))c_i(\frac{N}{4}+\frac{N}{2}m+n)\\ +(\sqrt{1-\mathrm{\&beta;}}{x}_j^{*}(d+m+n)+\sqrt{\mathrm{\&beta;}}{c}_i^{*}(d+m+n))c_i(m+n)\}\\ =B\{\sqrt{\mathrm{\&beta;}}{c}_i^{*}(d+\frac{N}{4}+\frac{N}{2}m+n)c_i(\frac{N}{4}+\frac{N}{2}m+n)\&CenterDot;+\sqrt{\mathrm{\&beta;}}{c}_i^{*}(d+m+n))c_i(m+n)\}\end{array}---\left(8\right)$

(8) formula is brought into (5) formula obtain:

$\begin{array}{c}B\left\{P\left(d\right)\right\}=\underset{i=1}{\overset{N_T}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{N_R}{\mathrm{\&Sigma;}}}\underset{m=0}{\overset{1}{\mathrm{\&Sigma;}}}\underset{n=0}{\overset{N/4-1}{\mathrm{\&Sigma;}}}B\{\sqrt{\mathrm{\&beta;}}{c}_i^{*}(d+\frac{N}{4}+\frac{N}{2}m+n)\&CenterDot;c_i(\frac{N}{4}+\frac{N}{2}m+n)\\ +\sqrt{\mathrm{\&beta;}}{c}_i^{*}(d+m+n)\&CenterDot;c_i(m+n)\}\\ =\sqrt{\mathrm{\&beta;}}{\mathrm{\&sigma;}}_c^2\end{array}---\left(9\right)$

In formula: represent the transmitting power of training sequence.When the overlying training sequence of receiving terminal data-signal aligns with local training sequence data block, B{P (d) } reach maximum.Because timing metric function R (d) in formula (5) is normalization effect, then, when M (d) reaches maximum, can judge that d is this moment OFDM symbol starting position, synchronized algorithm is expressed as:

$\widehat{\mathrm{\&tau;}}=\underset{d}{\arg \max }\left\{\right|M\left(d\right)\left|\right\}---\left(10\right)$

In formula: represent Timing Synchronization estimated value;

Step 6: can find out that synchronous exact position is relevant with training sequence power allocation factor by (9) formula, when power allocation factor is larger, timing net synchronization capability is better;

Step 7: adopt receiving terminal data message autocorrelation, obtain decimal frequency bias and estimate, its mathematical formulae is expressed as:

$F\left(\widehat{\mathrm{\&tau;}}\right)=\underset{i=1}{\overset{N_T}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{N_R}{\mathrm{\&Sigma;}}}\underset{n=0}{\overset{N/2-1}{\mathrm{\&Sigma;}}}{r}_j^{*}(\widehat{\mathrm{\&tau;}}+n+N/2)r_j\left(n\right)---\left(11\right)$

${\widehat{\mathrm{\&epsiv;}}}_f=\frac{1}{2\mathrm{\&pi;}}\mathrm{angle}\left\{\frac{\mathrm{rea}\left[F\left(\widehat{\mathrm{\&tau;}}\right)\right]}{\mathrm{imag}\left[F\left(\widehat{\mathrm{\&tau;}}\right)\right]}\right\}---\left(12\right)$

In formula: real () represents get real part computing, imag () represents get imaginary-part operation, represent decimal frequency bias estimated value, its estimation range:

Step 8: be that the integer frequency bias of the laggard line frequency of fractional frequency migration estimated is estimated, ask phase place to obtain integer frequency bias estimated value, expression formula by local training sequence and receiving terminal data-signal related operation:

${\widehat{\mathrm{\&epsiv;}}}_i=-\frac{N}{4\mathrm{\&pi;}}\mathrm{angle}\left[Q{\left(d\right)}_{d=\widehat{\mathrm{\&tau;}}}\right]---\left(13\right)$

$Q\left(d\right)=\underset{i=1}{\overset{N_T}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{N_R}{\mathrm{\&Sigma;}}}\underset{m=0}{\overset{1}{\mathrm{\&Sigma;}}}\underset{n=0}{\overset{N/2-1}{\mathrm{\&Sigma;}}}{r}_j^{*}(d+\frac{N}{2}m+n)\&CenterDot;c_i(n+\frac{N}{2}m+\frac{N}{2})+r_j^{*}(d+n+\frac{N}{2}m+\frac{N}{2})\&CenterDot;c_i(n+\frac{N}{2}m)---\left(14\right)$

In formula: represent that integer frequency bias is estimated, its estimation range is:

Accompanying drawing explanation

In order to more describe the synchronous new method of STBC MIMO-OFDM system time frequency under a kind of low signal-to-noise ratio in detail, for involved in the present invention to chart be described in detail.

Fig. 1 is STBC MIMO-OFDM system block diagram of the present invention.In figure, STBC MIMO-OFDM system block diagram forms primarily of two parts, mainly contains at transmitting terminal: the data message of input is through modules such as MPSK modulation 102, Space Time Coding 103, IFFT modulation 104; Mainly contain at receiving terminal: synchronous and channel estimating 112, FFT demodulation 116, empty time decode 117, the module such as MPSK demodulation 118.

Fig. 2 is the structured flowchart that training sequence of the present invention is added in OFDM symbol.In figure, training sequence is added on each transmitting antenna OFDM data symbol.

Fig. 3 is the organigram of the multiple orthogonal complement sequence of the present invention.In figure, multiple orthogonal complement sequence utilizes multiple complementary sequence set structural matrix, and expand the orthogonal sequence of appropriate length by orthogonal matrix and matrixing.

Fig. 4 is the training sequence structure schematic diagram that the present invention is based on multiple orthogonal complement sequence.In figure, adopt the multiple orthogonal complement sequence of appropriate length, and re-construct the training sequence based on multiple orthogonal complement sequence by conjugation negate and repeat property.

Fig. 5 is STBC MIMO-OFDM system time frequency synchronized algorithm flow chart of the present invention.In figure, utilize local training sequence acquisition relevant with the data message timing synchronisation information of receiving terminal, timing synchronization position is judged by maximum, carry out decimal frequency bias estimation and compensation subsequently, finally adopt local training sequence and receiving terminal data-signal related operation to ask phase place to obtain integer frequency bias estimated value.Fig. 6 is the BER Simulation figure of different capacity distribution factor of the present invention.In figure, abscissa represents signal to noise ratio (SNR), and ordinate represents bit error rate performance (BER), and β represents power allocation factor, finds out by figure, and along with the increase of power allocation factor β, error rate of system performance reduces.

Fig. 7 is the synchronous correct probability figure of different capacity distribution factor of the present invention.In figure, abscissa represents signal to noise ratio (SNR), and ordinate represents time synchronized correct probability, and β represents power allocation factor, finds out by figure, and along with the increase of power allocation factor β, system time synchronization correct probability performance improves.

Fig. 8 is time synchronized correct probability Performance comparision figure of the present invention.In figure, abscissa represents signal to noise ratio (SNR), ordinate represents time synchronized correct probability, the people such as the Ali Rachini Walsh-Hadamard sequence that adopts orthogonality good and CAZAC sequence two kinds of Time synchronization algorithms respectively as training sequence in document [1], the CAZAC sequence time synchronized algorithm that the people such as Fan Hui-li adopt shift-orthogonal in document [2], found out by figure, the performance of the time synchronized correct probability of innovatory algorithm is obviously better than document [1] and [2].Documents: [1] R.N.Ali, B.D.Ali, N.Fabienne and B.Bilal.Timing synchronization method for MIMO-OFDM system using orthogonal preamble [C] .19th International Conference on Telecommunications (ICT 2012), 2012:1-6. [2] H.L.Fan, J.F.Sun, P.Yang, D.S.Li.A Robust Timing and Frequency Synchronization Algorithm for HF MIMO OFDM Systems [C] .IEEE Conference on Global Mobile Congress (GMC), 2010:1-4.

Fig. 9 is that frequency deviation of the present invention estimates mean square error Performance comparision figure.In figure, abscissa represents signal to noise ratio (SNR), ordinate represents that frequency deviation estimates mean square error (MSE), that the people such as Fan Hui-li propose to adopt Received signal strength auto-correlation to obtain decimal frequency bias algorithm for estimating in document [2], found out by figure, the frequency deviation of innovatory algorithm estimates that mean square error performance is more better than document [2].Documents: [2] H.L.Fan, J.F.Sun, P.Yang, D.S.Li.A Robust Timing and Frequency Synchronization Algorithm for HF MIMO OFDM Systems [C] .IEEE Conference on Global Mobile Congress (GMC), 2010:1-4.

Embodiment

Below in conjunction with accompanying drawing, patent of the present invention is described in further details.

Fig. 1 is that the present invention has N _ttransmit antennas and N _rthe STBC MIMO-OFDM system fundamental block diagram of root reception antenna.This system block diagram is mainly made up of transmitter and receiver; at transmitter terminal; the module mainly comprised has data bit flow module 101, modulation module (MPSK) 102, Space Time Coding module 103, inverse Fourier transform (IFFT) module 104, parallel serial conversion module 105, parallel training block 106, power division module 107, laminating module 108, insertion protection interval module 109 and digital-to-analogue conversion (D/A) module.At receiver end; the module mainly comprised has analog-to-digital conversion (A/D) module 111, synchronous and channel estimation module 112, training sequence and data separating module 113, goes to protect interval module 114, serial to parallel conversion module 115, Fourier transform (FFT) module 116; decoder module 117 time empty; demodulation (MPSK) module 118, data bit flow module 119.The present invention mainly adopts at transmitter terminal and is added on OFDM data symbol by parallel training block 106, can improve the availability of frequency spectrum and the efficiency of transmission of system like this, save bandwidth resources.Adopting Space Time Coding module 103 can realize multiple transmit antenna coding techniques before, antenna diversity can also resist the decline of channel, thus improves the reliability of communication link, in addition, redundant information when also increasing transmission empty, improves the Stability and dependability of wireless communication transmissions.At receiver end, synchronous and channel estimation module mainly determines the original position of the window of FFT and the orthogonality of each subcarrier of guarantee, and when carrying out sky subsequently, decoder module 117 and demodulation (MPSK) module 118 realize the correct demodulation to OFDM data module.

Fig. 2 is the structured flowchart that training sequence of the present invention is added in OFDM symbol, and on each transmitting antenna that is added to by the training sequence based on multiple orthogonal complement on a complete OFDM data symbol, 201 is training sequence c ₁n (), 202 is training sequence c ₂(n).Fig. 3 is the organigram of the multiple orthogonal complement sequence of the present invention, the multiple orthogonal complement sequence step of design: first 301 be the multiple complementary sequence set of design one, and 302 utilize above-mentioned multiple complementary sequence set structural matrix E _{4 × 8}, 303 utilize orthogonal matrix H _{2 × 4}structural matrix E _{16 × 16}, 304 utilize orthogonal matrix F _{2 × 2}re-construct matrix E _{32 × 32}, 305 repeat above-mentioned steps the 4th step can obtain longer orthogonal sequence.

Fig. 4 is the training sequence structure schematic diagram that the present invention is based on multiple orthogonal complement sequence, the structure of the training sequence of multiple orthogonal complement sequence sequence b ₁n (), 403 by sequence a ₁(n) and b ₁n () composite construction becomes length to be the sequence t of N/2 ₁n (), 404 by sequence t ₁n () repeats once to construct the training sequence c that length is N ₁(n).

Fig. 5 is STBC MIMO-OFDM system time frequency synchronized algorithm flow chart of the present invention, 501 is Received signal strength, 502 local training sequence acquisition relevant with the data message Timing Synchronizations utilizing receiving terminal, 503 judge timing synchronization position by maximum, 504 decimal frequency bias are estimated, 505 compensate of frequency deviation, 506 utilize local training sequence and receiving terminal data-signal related operation to ask phase place to obtain integer frequency bias estimated value.

The synchronous correct probability analogous diagram of Fig. 6 to be the BER Simulation figure of different capacity distribution factor of the present invention, Fig. 7 be different capacity distribution factor, main simulation parameter of the present invention is arranged: number of transmit antennas N _t=2, reception antenna number N _r=2, covering Caro simulation times is 10000 times, and modulation system adopts MPSK, sub-carrier number N=1024, and the length of Cyclic Prefix is Ng=128, and choosing channel circumstance is multidiameter fading channel, and frequency deviation is set to ε=10.35.By the optimal power allocation factor of error rate of system performance and synchronous correct probability performance tradeoff overlying training sequence power allocation factor β.When the error rate is less and synchronous correct probability is higher, determine that power allocation factor is now optimal value.As can be seen from Fig. 6 and Fig. 7, under identical signal to noise ratio, along with power allocation factor increases β, bit error rate performance reduces, but time synchronized correct probability strengthens.In order to algorithm performance compromise, choose power allocation factor value β=0.3 close with theoretical value herein and will carry out on this basis as follow-up simulation analysis.

Fig. 8 is time synchronized correct probability performance simulation figure of the present invention, and main simulation parameter of the present invention is arranged: number of transmit antennas N _t=2, reception antenna number N _r=2, covering Caro simulation times is 10000 times, and modulation system adopts MPSK, sub-carrier number N=1024, and the length of Cyclic Prefix is Ng=128, and choosing channel circumstance is multidiameter fading channel, and frequency deviation is set to ε=10.35, and power allocation factor is set to β=0.3.As seen from the figure, when SNR=-19dB, correct probability can reach more than 90% herein.Compare with documents [1] and [2], method for synchronizing time better performances in this paper.

Fig. 9 is that frequency deviation of the present invention estimates mean square error performance simulation figure, and main simulation parameter of the present invention is arranged: number of transmit antennas N _t=2, reception antenna number N _r=2, covering Caro simulation times is 10000 times, and modulation system adopts MPSK, sub-carrier number N=1024, and the length of Cyclic Prefix is Ng=128, and choosing channel circumstance is multidiameter fading channel, and frequency deviation is set to ε=10.35, and power allocation factor is set to β=0.3.As can be seen from Figure, the frequency deviation region of decimal frequency bias algorithm for estimating is herein | ε _f|≤1 has lower mean square error, overcomes conventional algorithm frequency deviation region to be | ε _f| the defect of≤0.5.Under identical signal to noise ratio condition, frequency excursion algorithm in this paper compares with the algorithm performance of documents [2].There is lower mean square error.

Claims (1)

1. the synchronous new method of STBC MIMO-OFDM system time frequency under low signal-to-noise ratio, is characterized in that:

Step 1: multiple orthogonal complement sequences Design:

(1) a multiple complementary sequence set { A is designed ₁, B ₁, C ₁, D ₁, wherein A ₁=[1,1, j ,-j], B ₁=[1,1 ,-j, j], with c ₂=B ₁, D ₂=-A ₁, wherein a ₁conjugation negate computing ,-A ₁represent A ₁negative value, j represents imaginary symbols, j ²=-1;

(2) utilize above-mentioned multiple complementary sequence set structure row vector to be 4, column vector is the matrix E of 8 _{4 × 8}, wherein $E_{4\&times;8}{\left(\begin{array}{cccc}{A}_1& B_1& C_1& D_1\\ A_2& B_2& C_2& D_2\end{array}\right)}^T,$ In formula () ^trepresenting matrix transposition;

(3) orthogonal matrix is utilized $H={\left(\begin{array}{cccc}1& 1& -1& -1\\ -j& j& -j& j\end{array}\right)}^T$ Re-constructing row vector is 16, and column vector is 16 new matrix E _{16 × 16}, its matrix $E_{16\&times;16}={\left(\begin{array}{cccc}{E}_{4\&times;8}& E_{4\&times;8}& -E_{4\&times;8}& -E_{4\&times;8}\\ -j*E_{4\&times;8}& j*E_{4\&times;8}& -j*E_{4\&times;8}& j*E_{4\&times;8}\end{array}\right)}^T;$

(4) orthogonal matrix is utilized $F=\left(\begin{array}{cc}1& -j\\ 1& j\end{array}\right)$ Re-constructing row vector is 32, and column vector is 32 new matrix E _{32 × 32}, obtain matrix $E_{32\&times;32}=\left(\begin{array}{cc}{E}_{16\&times;16}& -j*E_{16\&times;16}\\ E_{16\&times;16}& j*E_{16\&times;16}\end{array}\right);$

(5) repeat above-mentioned (4) one step process, change row vector size, longer orthogonal sequence can be obtained;

Step 2: utilize above-mentioned multiple orthogonal complement sequences Design step, then can obtain orthogonal matrix $E_{\frac{N}{4}\&times;\frac{N}{4}}=\left(\begin{array}{cc}{E}_{\frac{N}{4}\&times;\frac{N}{4}}& -j*E_{\frac{N}{4}\&times;\frac{N}{4}}\\ E_{\frac{N}{4}\&times;\frac{N}{4}}& j*E_{\frac{N}{4}\&times;\frac{N}{4}}\end{array}\right),$ Pass through matrix construct one group of sequence a _i(n), wherein the span i ∈ [1, N/4] of the span n ∈ [0, N/4-1] of n, i;

As i=1, citing structure length is the sequence a of N/4 ₁n (), by sequence a ₁n () is got conjugation negate computing and is obtained new sequence b ₁n (), has:

$b_1\left(n\right)=-a_1^{*}\left(n\right),n\&Element;[0,N/4-1]---\left(1\right)$

In formula: N represents sub-carrier number, represent sequence a ₁() sequence gets conjugate operation;

Step 3: by sequence a ₁(n) and b ₁n () forms length is the sequence t of N/2 ₁n (), has:

$t_1\left(n\right)=\left\{\begin{array}{cc}{a}_1\left(n\right),& n\&Element;[0,N/4-1]\\ b_1\left(n\right),& n\&Element;[N/4,N/2-1]\end{array}\right.---\left(2\right)$

Step 4: by sequence t ₁n () repeats once to construct the sequence c that length is N ₁n (), has:

$c_1\left(n\right)=\left\{\begin{array}{cc}{t}_1\left(n\right),& n\&Element;[0,N/2-1]\\ t_1(n-N/2),& n\&Element;[N/2,N-1]\end{array}\right.---\left(3\right)$

Step 5: in the complete OFDM symbol that is added to by training sequence, signal indication is:

$r_j\left(n\right)=\sqrt{P(1-\mathrm{\&beta;})}{x}_i\left(n\right)+\sqrt{\mathrm{P\&beta;}}{c}_i\left(n\right),n=\mathrm{0,1},...N+\mathrm{Ng}-1---\left(4\right)$

In formula: P represents the gross power of transmitter, β is power allocation factor, and its span 0 < β < 1, is defined as $\mathrm{\&beta;}={\mathrm{\&sigma;}}_c^2/({\mathrm{\&sigma;}}_c^2+{\mathrm{\&sigma;}}_x^2),$ Wherein ${\mathrm{\&sigma;}}_c^2=E\left({\left|c\left(n\right)\right|}^2\right),{\mathrm{\&sigma;}}_x^2=E\left({\left|x\left(n\right)\right|}^2\right);$

Step 6: the local training sequence utilizing receiving terminal to receive acquisition time relevant with data message is synchronous, and timing metric function expression is:

$M\left(d\right)=\frac{{\left|P\left(d\right)\right|}^2}{{\left|R\left(d\right)\right|}^2}---\left(5\right)$

Wherein:

$P\left(d\right)=\underset{i=1}{\overset{N_T}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{N_R}{\mathrm{\&Sigma;}}}\underset{m=0}{\overset{1}{\mathrm{\&Sigma;}}}\underset{n=0}{\overset{N/4-1}{\mathrm{\&Sigma;}}}[r_j^{*}(d+\frac{N}{4}+\frac{N}{2}m+n)c_i(\frac{N}{4}+\frac{N}{2}m+n)+r_j^{*}(d+m+n)c_i(m+n)]---\left(6\right)$

$R\left(d\right)=\underset{i=1}{\overset{N_T}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{N_R}{\mathrm{\&Sigma;}}}\underset{n=0}{\overset{N/2-1}{\mathrm{\&Sigma;}}}{\left|r_j(d+n)\right|}^2---\left(7\right)$

In formula: N _tfor the number of transmitting antenna, N _rfor the number of reception antenna, Ng is the length of Cyclic Prefix, represent receiving terminal data message r _j() gets conjugate operation, and n is the number of sampled point, c _in () represents the training sequence on transmitting antenna, d is integer value, and it represents the relative sliding position of local training sequence and receiving terminal data message, and m represents cycle-index;

Step 6: because timing metric function R (d) in formula (5) is normalization effect, then, when M (d) reaches maximum, can judge that d is this moment OFDM symbol starting position, synchronized algorithm is expressed as:

$\widehat{\mathrm{\&tau;}}=\underset{d}{\arg \max }\left\{\right|M\left(d\right)\left|\right\}---\left(8\right)$

In formula: represent Timing Synchronization estimated value;

Step 7: adopt receiving terminal data message autocorrelation, obtain decimal frequency bias and estimate, its mathematical formulae is expressed as:

$F\left(\widehat{\mathrm{\&tau;}}\right)=\underset{i=1}{\overset{N_T}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{N_R}{\mathrm{\&Sigma;}}}\underset{n=0}{\overset{N/2-1}{\mathrm{\&Sigma;}}}{r}_j^{*}(\widehat{\mathrm{\&tau;}}+n+N/2)r_j\left(n\right)---\left(9\right)$

${\widehat{\mathrm{\&epsiv;}}}_f=\frac{1}{2\mathrm{\&pi;}}\mathrm{angle}\left\{\frac{\mathrm{real}\left[F\left(\widehat{\mathrm{\&tau;}}\right)\right]}{\mathrm{imag}\left[F\left(\widehat{\mathrm{\&tau;}}\right)\right]}\right\}---\left(10\right)$

In formula: real () represents get real part computing, imag () represents get imaginary-part operation, represent decimal frequency bias estimated value, its estimation range: ${\widehat{\mathrm{\&epsiv;}}}_f\&Element;(\mathrm{0,1}];$

Step 8: the integer frequency bias of the laggard line frequency of fractional frequency migration estimated is estimated, asks phase place to obtain integer frequency bias estimated value, expression formula by local training sequence and receiving terminal data-signal related operation:

${\widehat{\mathrm{\&epsiv;}}}_i=-\frac{N}{4\mathrm{\&pi;}}\mathrm{angle}\left[Q{\left(d\right)}_{d=\widehat{\mathrm{\&tau;}}}\right]---\left(11\right)$

$Q\left(d\right)=\underset{i=1}{\overset{N_T}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{N_R}{\mathrm{\&Sigma;}}}\underset{m=0}{\overset{1}{\mathrm{\&Sigma;}}}\underset{n=0}{\overset{N/2-1}{\mathrm{\&Sigma;}}}{r}_j^{*}(d+\frac{N}{2}m+n)\&CenterDot;c_i(n+\frac{N}{2}m+\frac{N}{2})+r_j^{*}(d+n+\frac{N}{2}m+\frac{N}{2})\&CenterDot;c_i(n+\frac{N}{2}m)---\left(12\right)$

In formula: represent that integer frequency bias is estimated, its estimation range is: