CN1893337A - Emission diversity method for time-domain orthogonal frequency dividing duplexing system - Google Patents

Abstract

The invention includes steps: (1) dividing input sequence in frequency domain into odd and even sub sequences in length N/2 according to sequence number of sub carrier wave; (2) carrying out inversed discrete Fourier transform for continuous two frames of the said sub sequence, storing result to buffer; (3) calculating buffered data so as to obtain four signals in time domain in transmitting chain; (4) inserting different PN sequences as frame head into TDS-OFDM guard gaps of four transmitting chain respectively, building up integrated signal frame of each transmitting chain from the frame head and frame body obtained from previous step; (5) after forming filtering, D/A conversion, and process at front end, signal frames are transmitted out through four antennae at prearranged channel bandwidth. Advantages are: simple, quick, accuracy, holding transmission efficiency, more suitable to channel of dual selections of time and frequency, supporting 'soft failure'etc.

Description

The emission diversity method of time-domain synchronous orthogonal frequency-division multiplexing system

Technical field

The invention belongs to digital information transmission technical field, relate more specifically to a kind of time-domain synchronization OFDM (TimeDomain Synchronous OFDM, TDS-OFDM) in the system based on the time domain and frequency domain combined emission diversity method of space-time block code (Space-Time Block Code).

Background technology

In the wireless environment of complexity, object on every side (as house, building or trees etc.) can play the effect of reflection to radio wave.These barriers can produce the amplitude fading reflected wave different with phase delay.If launch a modulation signal, these a plurality of reflected waves that transmit will arrive reception antenna through different propagation delays from different directions so.These reflected signals according to the difference of its random phase, can play the effect of strengthening or weakening through after being positioned at receiver antenna reception everywhere to received signal.Can cause the changes in amplitude of receiving end signal thus, form decline.Statistics shows that in uniform avenue of barrier or forest environment, signal envelope rises and falls to be similar to and satisfies the Rayleigh distribution, so the multipath rapid fading is called the Rayleigh decline again.Short term rapid fading is because receiving and transmitting signal both sides' caused by relative motion: the existence of multipath signal causes the time diffusion, thereby causes the intersymbol interference of transmission signals; And the Doppler effect that relative motion causes can cause that the phase place of transmission signals changes rapidly, and different rapid fading characteristics is arranged under different test environments.

Except quick Rayleigh decline can appear in the instantaneous value of received signal, median of field strength also slow variation can occur.The reason that changes mainly contains two aspects: the one, cause by the shadow effect of the fixed obstacle in the mobile environment (as building, massif, forest etc.); The 2nd, because the variation of meteorological condition causes the vertical gradient generation of atmosphere relative dielectric constant gradual, promptly the radio wave refration coefficient changes in time, and the signal time delay of acceptance point also changes thereupon thereby multipath transmisstion arrives fixedly.This signal that is caused by shadow effect and meteorological reason changes, and is called slow fading.Especially, less by the variation that meteorological reason causes, ignore usually.Received signal under the slow fading environment is similar to obeys logarithm normal distribution, and amplitude of variation depends on barrier situation, operating frequency, rate of change, barrier and receiver translational speed etc.

Move in the reception at multipath, the Doppler frequency stretch that time delay expansion that multipath effect causes and Doppler effect cause exists simultaneously, is called frequency and time selective fading channels.Under this channel, the signal to noise ratio of received signal is very unstable, and received signal to noise ratio is low when channel is among the deep fading, and the probability of mistake in judgment is just big, seriously reduces the reliability of signal transmission.In order to improve the anti fading performance of system, can adopt various channel equalization techniques, OFDM (OFDM) multi-carrier modulation technology etc.And diversity technique is the effective technology that overcomes frequency and time selective fading, and it through several incoherent fading channels, synthesizes identical information then to received signal.Because it is lower that several channels are in the probability of deep fade simultaneously, therefore can reach level and smooth channel fading, increase signal to noise ratio, improve the purpose of receiver bit error performance.In the Digital Television Terrestrial Broadcasting network,, therefore under same transmitting power, can also expand the coverage of TV signal because diversity technique has reduced the signal to noise ratio (snr) threshold requirement of receiver.

Traditional signal diversifying technology such as time diversity, frequency diversity etc., the repeated encoding that in fact their effect is equivalent in the chnnel coding adds interleaving technology, though can improve the error performance of system and since identical information repeat transmit, often to sacrifice bigger efficiency of transmission.Another diversity mode commonly used is to adopt multiple antenna to carry out space diversity, and this technology can realize at transmitting terminal or receiving terminal, be called transmit diversity and receive diversity.

Wherein, the receive diversity that uses a plurality of reception antennas is a kind of tradition and effective diversity technique, it does not need to sacrifice efficiency of transmission, can adopt plain modes such as maximum switching, maximum ratio merging to finish the selection or the merging of a plurality of received signals at receiving terminal, and then decipher and adjudicate according to conventional method.Calendar year 2001, the researcher of French Harris company carried out European digital video broadcast-terrestrial DVB-T the experiment of reception antenna diversity (Faria G.Mobile DVB-T using antennadiversity receivers.2001.Available: Http: ∥ www.broadcastpapers.com), under various complicated multi-path environments, effect is fine, the about 6dB of average SNR thresholding decline of test prototype, and the anti-Doppler ability increases by 100%.This receive diversity scheme shown in figure l, receiving terminal adopt two groups independently radio-frequency front-end and OFDM separate the mediation channel estimation module, after the OFDM demodulation, the sample value of two-way received signal on k subcarrier is respectively:

$\left\{\begin{array}{c}{R}_{\mathrm{Rx}1}\left(k\right)=H_{\mathrm{Rx}1}\left(k\right)X\left(k\right)+N_{\mathrm{Rx}1}\left(k\right)\\ R_{\mathrm{Rx}2}\left(k\right)=H_{\mathrm{Rx}2}\left(k\right)X\left(k\right)+N_{\mathrm{Rx}2}\left(k\right)\end{array}\right.(0\≤k\≤N-1)$

Wherein, H _Rx1(k) and H _Rx2(k) be respectively frequency response values on k subcarrier of two RX path, N _Rx1(k) and N _Rx2(k) then represent noise on the respective paths respectively.

Suppose that the two-way received signal has experienced mutual incoherent channel fading and channel estimation results is correct, the maximum rate merging is that best signal merges mode so.Conjugation addition again with two group of received signals multiply by its sub-carrier frequencies response respectively obtains:

$R\left(k\right)=H_{\mathrm{Rx}1}^{*}\left(k\right)R_{\mathrm{Rx}1}\left(k\right)+H_{\mathrm{Rx}2}^{*}\left(k\right)R_{\mathrm{Rx}2}\left(k\right)$

$=({\left|H_{\mathrm{Rx}1}\left(k\right)\right|}^2+{\left|H_{\mathrm{Rx}2}\left(k\right)\right|}^2)X\left(k\right)+H_{\mathrm{Rx}1}^{*}\left(k\right)N_{\mathrm{Rx}1}\left(k\right)+H_{\mathrm{Rx}2}^{*}\left(k\right)N_{\mathrm{Rx}2}\left(k\right)$

Wherein, the composite signal of following formula is identical with the signal form that the non diversity system obtains, and can be directly used in decoding and judgement.Obviously, the signal to noise ratio of the signal after the merging is higher than the signal to noise ratio maximum of two tributary signals, has therefore obtained diversity gain.Multiple receive antenna also carries out the maximum rate merging in a manner described if adopt more, can also obtain bigger gain.

Receive diversity can obtain good effect, and deversity scheme is also very simple, but can be subjected to some restrictions in being applied to Digital Television Terrestrial Broadcasting (DTTB) field the time.The one, receive diversity needs receiver to have the parallel radio-frequency front-end of many covers to handle, and has increased the cost and the complexity of receiver, and this is uneconomical for broadcast system; The 2nd, make each road received signal uncorrelated, the distance of per two reception antennas will be 10 times of magnitudes of carrier wavelength, and in the residing VHF/UHF frequency range of DTTB, this distance is about 4～7m, and this is difficult to realize for a lot of moving with portable receiving terminal.On the contrary, for transmitter, above-mentioned restriction all is out of question, and therefore, transmit diversity techniques becomes the focus of research day by day.Because when adopting transmit diversity, between a plurality of transmitting antennas, often need to carry out the reasonable disposition of signal phasor to increase diversity order (Diversity Order) as far as possible, so this technology also is regarded as the another layer " ISN " that is placed on after conventional channel is encoded, and is called the diversity coding.The signal sampling of its receiving terminal is a plurality of stacks that transmit, and need separate and deciphers by suitable processing mode.

Have the achievement in research of a lot of transmit diversities in recent years.The diversity mode that adopts in document " Wittneben A.A new bandwidth efficienttransmit antenna modulation diversity scheme for linear digital modulation.in Proc.of IEEE ICC ' 93.Geneva; Switzerland:IEEE; 1993.1630-1634 " and " Winters J.The diversity gain of transmit diversityin wireless system with Rayleigh fading.in Proc.of IEEE ICC ' 94.New Orleans; LA:IEEE; 1994.1121-1125 " is with same signal time-delay emission, forms one " artificial multipath " and the mode that receives with similar Rake and is merged.Layering sign indicating number when document " Foschini G and Gans M.On limits of wireless communications in afading environment when using multiple antenna.Wireless Personal Communications; 1998; 6 (3): 311-335 " has been introduced empty in the Blast system, it is to transmit demultiplexing and carry out traditional chnnel coding respectively and interweave.1998, Tarokh is in the notion of having introduced trellis coding (STTC) when empty in " Tarokh V; Seshadri N; and Calderbank A.Space-time codes forhigh data rate wireless communications:performance criterion and code construction.IEEE Trans.onInformation Theory; 1998; 44 (2): 744-765 ", it is united chnnel coding and antenna diversity and designs, this method can obtain to gain to greatest extent in theory, but need to change the design of whole emission system, even under the situation of less number of transmit antennas and low order planisphere, decoding complexity is still very big.In order to address this problem, Alamouti has proposed a kind of space-time block code (STBC) scheme in 1998 in its classical paper " Alamouti S.A simple transmitdiversity technique for wireless communications.IEEE Trans.on Select Areas in Communications; 1998; 16 (8): 1451-1458 ", in two system of transmit antennas, use, its code construction and decoding algorithm are very simple, can obtain diversity gain equally.The situation that people such as Tarokh extended to any number of transmit antennas in 1999 with STBC has provided theory analysis and structure criterion to this scheme.The structure that above-mentioned these STBC codings are based on quadrature (orthogonal) proposes.On complex field, suppose to have k symbol (x ₁, x ₂..., x _k) p * n _TThe block encoding matrix is G (x ₁, x ₂..., x _k), the element among the G satisfies: each all is x for (1) _i, x _i ^*One of or its linear combination; (2) G ^HG=(| x ₁| ²+ | x ₂| ²+ ...+| x _k| ²) I _n, I _nBe unit matrix.When this code word is used for the transmission antenna diversity system, a continuous k incoming symbol is encoded according to shown in the G, each row of matrix are all represented for the symbol sebolic addressing that transmitting antenna sends.Because encoder matrix G has the property of orthogonality shown in (2), each symbol can be separated when making receiving terminal decoding, x _iDecipher respectively, so just greatly reduce the decoding complexity of receiving terminal.

The code efficiency of definition STBC code word is R=k/p, and wherein, k is the incoming symbol number, and p is an encoding time delay.For the 2 antenna transmit diversity systems that Alamouti proposes, R=1.But as number of transmit antennas (n _T) more than 2 o'clock, verified in people's such as Tarokh document, based on the code efficiency of the STBC sign indicating number of property of orthogonality design less than 1, i.e. p＞k, thereby lost effective transmission code rate.This just means the transmission code rate that will make the system that adopts behind the transmit diversity still keep original single-shot transmitter system, must increase the shared bandwidth of original system, this system to bandwidth fixed (as DTTB) is a very big shortcoming, therefore introduced STBC sign indicating number (" Jafarkhani H; A quasi-orthogonalspace-time block code.IEEE Trans.on Communications; 2001,49 (1): 1-4 ") based on the design of accurate quadrature (quasi-orthogonal) character.Accurate quadrature STBC has piled up the constraint of orthogonality condition loose, though reduced some diversity gains, can be so that code efficiency reaches 1, even in the system of bandwidth fixed, also can effectively use.On complex field, suppose that still the block encoding matrix is G, at this moment p=k.4 transmit antenna case that propose in the literature with Jafarkhani are example, and for the element among the accurate orthogonal design G, above-mentioned character (1) keeps, and (2) change into

$G^{H}G=\left[\begin{array}{cccc}a& 0& 0& b\\ 0& a& -b& 0\\ 0& -b& a& 0\\ b& 0& 0& a\end{array}\right]$

Wherein, $a=\underset{i=1}{\overset{4}{\mathrm{\Σ}}}{\left|x_i\right|}^2,$ b＝2Re(x ₁x ₄ ^*-x ₂x ₃ ^*)。Therefore, for each the column vector (V in the matrix _i, i=1,2,3,4), can be divided into 2 groups: (V ₁, V ₄) and (V ₂, V ₃).Non-orthogonal between the vector in every group, but be not quadrature between the vector on the same group.When receiving terminal carries out maximum likelihood (ML) decoding, can be two parts according to component with the judgement formula, operand is still very little like this.

Because STBC (orthogonal design or accurate orthogonal design) coded system has the advantage of fast decoding, so obtained broad research very soon, and expand in the OFDM channel of frequency selective fading by flat decline single carrier channel rapidly, formed Space-Time Block Coding based on ofdm system.Owing in the OFDM modulation technique, relate to time domain and two kinds of signals of frequency domain, so the STBC coding can carry out in time domain and frequency domain respectively.

If in the time domain incoming symbol, carry out STBC coding (STC-OFDM), be about to continuous k OFDM symbol encode according to shown in certain coded format G (" Lee K and Williams D.A space-time coded transmitter diversitytechnique for frequency selective fading channels.in Proc.IEEE Sensor Array and MultichannelSignal Processing Workshop.Cambridge; MA:IEEE, 2000.149-152 ").At this moment, in order to keep the character of transmission matrix G quadrature (or accurate quadrature), must suppose that channel is quasi-static, promptly channel remains unchanged in continuous k adjacent OFDM symbol time.This hypothesis can produce very mistake under fast fading channel, thereby picture is also inapplicable under the situations such as mobile reception of DTTB.

The STBC coding also can carry out (SFC-OFDM) in frequency domain, with the digital coding on the adjacent k subcarrier (" LeeK and Williams D.A space-frequency transmitter diversity technique for OFDM systems.in Proc.IEEE GLOBECOM ' 00.San Francisco; CA:IEEE; 2000,1473-1477 ").Like this, in order to keep the character of transmission matrix G quadrature (or accurate quadrature), need the frequency response on k the adjacent subcarrier of hypothesis identical.Though it is SFC-OFDM goes for fast fading channel, same because the error that the channel hypothesis is brought is also inapplicable in frequency selective fading channels.And in esse channel circumstance is double selectivity (time and frequency) mostly, so STC-OFDM and SFC-OFDM system all can bring than mistake in actual applications.

According to the engineering characteristic of Digital Television Terrestrial Broadcasting, the emission diversity method that the TDS-OFDM system adopts should be followed following design principle: (1) does not sacrifice efficiency of transmission.Diversity scheme should the assurance system keep original information handling capacity, can not introduce redundancy, that is to say that each radio-frequency channel can not reduce rate of information transmission because having increased diversity coding, thereby during more than the diversity system of 2 transmitting antennas, adopt above-mentioned quasi-orthogonal project organization in design; (2) do not change original emission system as far as possible.Make diversity scheme become one optional " accessory " of emission system, the network operator can determine whether needs adopt diversity, have also reduced the transmitter cost to greatest extent according to objective realities such as local channel characteristics and coverage conditions like this; (3) the receiver algorithm change is little, and complexity increases few.Increasing transmitting antenna can increase the functional module of receiver inevitably, as channel estimating and signal combiner etc.; (4) support " fail soft (Soft Failure) ".So-called support " fail soft " is meant that the signal phasor configuration between transmitting antenna should guarantee, when a RX path lost efficacy because of certain reason, another RX path still can make system normally receive, and only is to have sacrificed 1/2 mean receiving power.Therefore, this transmit diversity techniques has in fact also increased the reliability of system.

The Digital Television Terrestrial Broadcasting transmission standard mainly contains three kinds in the world at present: and the DVB-T (ground digital video terrestrial broadcasting Digital VideoTerrestrial Broadcasting-Terrestrial) in the ATSC of the U.S. (the Advanced Television Systems Committee of Advanced Television Systems Committee), Europe and the ISDB-T of Japan (floor synthetic service digital broadcasting Integrated ServiceDigital Broadcasting-Terrestrial, ISDB-T).China also began the research work of high definition TV from 1994.Under this background, Tsing-Hua University has proposed T-DMB (Digital Multimedia Broadcasting forTerrestrial, DMB-T) host-host protocol.

The time-domain synchronization OFDM that adopts among the DMB-T of Tsing-Hua University (TDS-OFDM) modulation belongs to multi-carrier modulation technology; but Coded Orthogonal Frequency Division Multiplexing (COFDM) (COFDM) technology that adopts with European DVB-T is different; in the TDS-OFDM system, do not insert the pilot tone signal; but in the protection of OFDM at interval, inserted pseudorandom (PN) sequence in the mode of time domain, be used for frame synchronization, Frequency Synchronization, regularly synchronously, channel transfer characteristic is estimated and follow the tracks of phase noise etc.

For realize quick and stable synchronously, the TDS-OFDM transmission system that Tsing-Hua University proposes has adopted hierarchical frame structure.The elementary cell of frame structure is called signal frame, as shown in Figure 2.200/225 signal frame is defined as a frame group, and 512 frame groups are defined as a superframe.The top layer of frame structure is called a day frame, is made up of superframe.Each signal frame among the frame group has unique frame number, and it is coded in the PN sequence of frame head.

The signal frame of TDS-OFDM transmission system uses the OFDM modulation of Domain Synchronous, and perhaps being called with the PN sequence is protection OFDM modulation at interval.A signal frame is made up of frame synchronization and frame two parts, and they have identical baseband signalling rate 7.56MS/s (1/T).A signal frame can be used as an OFDM (OFDM) piece.An OFDM piece further is divided into a protection interval and an inverse discrete Fourier transform (IDFT) piece.For TDS-OFDM, frame synchronization PN sequence is as the protection interval of OFDM, and frame is as the IDFT piece, as shown in Figure 3.

Seeing grant number for details about the correlation circumstance of DMB-T, TDS-OFDM is that 00123597.4 " ground digital multimedia TV broad cast system " by name, grant number are that 01115520.5 " time-domain synchronous orthogonal frequency division multiplex modulation method " by name, grant number are ZL01130659.9 " frame-synchronization generation method in the ground digital multimedia TV broad cast system " by name, and grant number is the Chinese invention patent that 01124144.6 " protection fill method at interval in the orthogonal FDM modulation system " by name waits Tsing-Hua University to apply for.

In order in the TDS-OFDM system, to realize transmission antenna diversity, must satisfy following condition:

(1) keeps enough distances between the transmitting antenna, so that it is independent to arrive each bar transmission channel statistics of receiver;

(2) receiving terminal can accurately estimate the channel information of each channel of current time when carrying out channel estimating;

(3) by suitable method the multiple signals that receive are separated, made it uncorrelated mutually, then isolated multiple signals are merged, obtain maximum signal to noise ratio.

At above-mentioned background, the present invention proposes a kind of time domain and frequency domain combined emission diversity method based on space-time block code at the TDS-OFDM system.

Summary of the invention

The object of the present invention is to provide a kind of time-domain synchronization OFDM (Time Domain Synchronous OFDM, TDS-OFDM) in the system based on space-time block code (Space Time Block Code, a kind of time domain and frequency domain combined emission diversity method STBC).

The present invention is directed to the transmit diversity problem in the digital tv ground broadcasting, proposed a kind of time domain and frequency domain combined emission diversity method based on accurate orthohormbic structure STBC coding.The emission diversity method block diagram that the present invention proposes as shown in Figure 4.

Time-domain synchronization OFDM of the present invention, be TDS-OFDM, time domain and frequency domain combined emission diversity method, it is characterized in that, described method is a kind of time domain and frequency domain combined emission diversity method based on space-time block code, and it is realized successively according to following steps in the special digital integrated circuit:

Step 1. note frequency domain list entries be X (k, l), wherein k represents the subcarrier sequence number, 0≤k≤N-1, N are the sub-carrier number in the ofdm system, l represents the signal frame sequence number, (k l) is divided into odd number subsequence X according to the subcarrier sequence number with X _o(k is l) with even number subsequence X _e(k, l), their length is N/2;

Step 2. is with X _o(k, l) and X _e(k l) makes N/2 point Inverse Discrete Fourier Transform respectively, and the time domain sequences that obtains is x _l ^To(n, l) and x _l ^Te(n, l);

Step 3. is imported x as a result after data are carried out Inverse Discrete Fourier Transform with two continuous frames ₁ ^To(n, l), x ₁ ^Te(n, l), x ₁ ^To(n, l+1) and x ₁ ^Te(n l+1) is deposited in the buffer memory;

Step 4. is carried out nonidentity operation with the data in the buffer memory according to following described four kinds of situations then, obtains being used for four required time-domain signals of antenna emission respectively:

(a) for first transmitting antenna Tx1, time-domain signal x _Tx1(n, l), x _Tx1(n l+1) is

$\left\{\begin{array}{c}{x}_{\mathrm{Tx}1}(n,l)=[x_1^{\mathrm{te}}(n,l)+x_1^{\mathrm{to}}(n,l)W_N^{-n}]/2,\\ x_{\mathrm{Tx}1}(n+N/2,l)=[x_1^{\mathrm{te}}(n,l)-x_1^{\mathrm{to}}(n,l)W_N^{-n}]/2,\\ x_{\mathrm{Tx}1}(n,l+1)=[x_1^{\mathrm{te}}(n,l+1)+x_1^{\mathrm{to}}(n,l+1)W_N^{-n}]/2,\\ x_{\mathrm{Tx}1}(n+N/2,l+1)=[x_1^{\mathrm{te}}(n,l+1)-x_1^{\mathrm{to}}(n,l+1)W_N^{-n}]/2,\end{array}\right.0\≤n\≤N/2-1;$

Wherein, $W_N^{-n}=e^{j\frac{2\mathrm{\π}}{N}n},$ N is the sub-carrier number in the ofdm system;

(b), earlier the data in the buffer memory are obtained through behind the space-frequency coding for second transmitting antenna Tx2

$\left\{\begin{array}{c}{x}_2^{\mathrm{te}}(n,l)={x_1^{\mathrm{to}}}^{*}({(-n)}_{N/2},l),\\ x_2^{\mathrm{to}}(n,l)=-{x_1^{\mathrm{te}}}^{*}({(-n)}_{N/2},l),\\ x_2^{\mathrm{te}}(n,l+1)={x_1^{\mathrm{to}}}^{*}({(-n)}_{N/2},l+1),\\ x_2^{\mathrm{to}}(n,l+1)=-{x_1^{\mathrm{te}}}^{*}({(-n)}_{N/2},l+1),\end{array}\right.0\≤n\≤N/2-1;$

Wherein, * represents the complex conjugate computing, (n) _N/2Expression obtains time-domain signal x then to n delivery N/2 computing _Tx2(n, l), x _Tx2(n l+1) is

$\left\{\begin{array}{c}{x}_{\mathrm{Tx}2}(n,l)=[x_2^{\mathrm{te}}(n,l)+x_2^{\mathrm{to}}(n,l)W_N^{-n}]/2,\\ x_{\mathrm{Tx}2}(n+N/2,l)=[x_2^{\mathrm{te}}(n,l)-x_2^{\mathrm{to}}(n,l)W_N^{-n}]/2,\\ x_{\mathrm{Tx}2}(n,l+1)=[x_2^{\mathrm{te}}(n,l+1)+x_2^{\mathrm{to}}(n,l+1)W_N^{-n}]/2,\\ x_{\mathrm{Tx}2}(n+N/2,l+1)=[x_2^{\mathrm{te}}(n,l+1)-x_2^{\mathrm{to}}(n,l+1)W_N^{-n}]/2,\end{array}\right.0\≤n\≤N/2-1;$

(c), earlier the data in the buffer memory are obtained through behind the Space Time Coding for the 3rd transmitting antenna Tx3

$\left\{\begin{array}{c}{x}_3^{\mathrm{te}}(n,l)={x_1^{\mathrm{te}}}^{*}({(-n)}_{N/2},l+1),\\ x_3^{\mathrm{to}}(n,l)={x_1^{\mathrm{to}}}^{*}({(-n)}_{N/2},l+1),\\ x_3^{\mathrm{te}}(n,l+1)={-x_1^{\mathrm{te}}}^{*}({(-n)}_{N/2},l),\\ x_3^{\mathrm{to}}(n,l+1)=-{x_1^{\mathrm{to}}}^{*}({(-n)}_{N/2},l),\end{array}\right.0\≤n\≤N/2-1;$

Obtain time-domain signal x then _Tx3(n, l), x _Tx3(n l+1) is

$\left\{\begin{array}{c}{x}_{\mathrm{Tx}3}(n,l)=[x_3^{\mathrm{te}}(n,l)+x_3^{\mathrm{to}}(n,l)W_N^{-n}]/2,\\ x_{\mathrm{Tx}3}(n+N/2,l)=[x_3^{\mathrm{te}}(n,l)-x_3^{\mathrm{to}}(n,l)W_N^{-n}]/2,\\ x_{\mathrm{Tx}3}(n,l+1)=[x_3^{\mathrm{te}}(n,l+1)+x_3^{\mathrm{to}}(n,l+1)W_N^{-n}]/2,\\ x_{\mathrm{Tx}3}(n+N/2,l+1)=[x_3^{\mathrm{te}}(n,l+1)-x_3^{\mathrm{to}}(n,l+1)W_N^{-n}]/2,\end{array}\right.0\≤n\≤N/2-1;$

(d) for the 4th transmitting antenna Tx4, the data in the buffer memory through Space Time Coding, obtain the x as a result as shown in (c) earlier ₃ ^To(n, l), x ₃ ^Te(n, l), x ₃ ^To(n, l+1) and x ₃ ^Te(n, l+1), and then the process space-frequency coding obtains

$\left\{\begin{array}{c}{x}_4^{\mathrm{te}}(n,l)={x_3^{\mathrm{to}}}^{*}({(-n)}_{N/2},l),\\ x_4^{\mathrm{to}}(n,l)=-{x_3^{\mathrm{te}}}^{*}({(-n)}_{N/2},l),\\ x_4^{\mathrm{te}}(n,l+1)={x_3^{\mathrm{to}}}^{*}({(-n)}_{N/2},l+1),\\ x_4^{\mathrm{to}}(n,l+1)=-{x_3^{\mathrm{te}}}^{*}({(-n)}_{N/2},l+1),\end{array}\right.0\≤n\≤N/2-1;$

Obtain time-domain signal x at last _Tx4(n, l), x _Tx4(n l+1) is

$\left\{\begin{array}{c}{x}_{\mathrm{Tx}4}(n,l)=[x_4^{\mathrm{te}}(n,l)+x_4^{\mathrm{to}}(n,l)W_N^{-n}]/2,\\ x_{\mathrm{Tx}4}(n+N/2,l)=[x_4^{\mathrm{te}}(n,l)-x_4^{\mathrm{to}}(n,l)W_N^{-n}]/2,\\ x_{\mathrm{Tx}4}(n,l+1)=[x_4^{\mathrm{te}}(n,l+1)+x_4^{\mathrm{to}}(n,l+1)W_N^{-n}]/2,\\ x_{\mathrm{Tx}4}(n+N/2,l+1)=[x_4^{\mathrm{te}}(n,l+1)-x_4^{\mathrm{to}}(n,l+1)W_N^{-n}]/2,\end{array}\right.0\≤n\≤N/2-1;$

Step 5. generates the PN sequence of four different respective length according to the length 420 or 945 of time-domain synchronous orthogonal frequency-division multiplexing system signal frame frame head;

Step 6. is according to the channel frame structure of time-domain synchronous orthogonal frequency-division multiplexing system; in the TDS-OFDM of transmitting antenna Tx1, Tx2, four links of Tx3 and Tx4 protection at interval, insert above-mentioned different PN sequence respectively, the frame x that frame head PN sequence and step (4) are obtained as frame head _Tx1(n, l), x _Tx2(n, l), x _Tx3(n, l), x _Tx4(n l) forms four signal frames that transmitting chain is complete separately respectively;

Above-mentioned complete TDS-OFDM signal is formed filtering to step 7. and digital to analog conversion is handled, then through comprising frequency up-converted and power amplifier in interior front-end processing, launch in predetermined channel bandwidth by antenna Tx1, Tx2, Tx3 and Tx4 respectively at last, finish transmission antenna diversity.

Time-domain synchronization OFDM of the present invention, i.e. TDS-OFDM, time domain and frequency domain combined emission diversity method is characterized in that, the equivalent space-time block code structure of described time domain and frequency domain combined coding is expressed as with matrix G:

$G=\left[\begin{array}{cccc}X(2k,l)& X^{*}(2k+1,l)& X^{*}(2k,l+1)& X(2k+1,l+1)\\ X(2k+1,l)& -X^{*}(2k,l)& X^{*}(2k+1,l+1)& -X(2k,l+1)\\ X(2k,l+1)& X^{*}(2k+1,l+1)& -X^{*}(2k,l)& -X(2k+1,l)\\ X(2k+1,l+1)& -X^{*}(2k,l+1)& -X^{*}(2k+1,l)& X(2k,l)\end{array}\right],(0\≤k\≤N/2-1);$

Wherein, (k l) is the frequency domain list entries to X, and k represents the subcarrier sequence number, and l represents the OFDM frame number, and this is space-time block code (STBC) structure of an accurate quadrature (quosi-orthogonal) design.

Time-domain synchronization OFDM of the present invention, i.e. TDS-OFDM, time domain and frequency domain combined emission diversity method, it is characterized in that, can keep original single transmit chain road (Tx1) constant substantially, and just add buffer therein, make net structure flexible.The time-domain signal of other three transmitting chains (Tx2, Tx3, Tx4) need only carry out simple process to the data in the buffer memory and can obtain, and therefore, need only do an IDFT computing in the time of transmission one frame ofdm signal, and computational complexity is very little.

Time-domain synchronization OFDM of the present invention, be TDS-OFDM, time domain and frequency domain combined emission diversity method, it is characterized in that, the strong robustness of diversity system work, promptly undesired as the work of fruit part transmission link, need not carry out any modification to original system, receiving terminal still can normally be deciphered, and error performance is not less than the situation under the single-shot transmitter system at least.In the present invention, if any transmitting chain breaks down, its excess-three transmitting chain still can be formed accurate quadrature STBC structure so, thereby obtains diversity gain.When having only two transmitting chains working properly, can be divided into three kinds of situations: 1) Tx1 and Tx3 (or Tx2 and Tx4), can form the quadrature STC-OFDM structure of one 2 antenna; 2) Tx1 and Tx2 (or Tx3 and Tx4) can form the quadrature SFC-OFDM structure of one 2 antenna; 3) Tx1 and Tx4 (or Tx2 and Tx3).1) and 2) under the situation, system still can obtain diversity gain.

Simultaneously, time domain and frequency domain combined emission diversity method proposed by the invention is without loss of generality, and can be transplanted to other multicarriers DTTB system easily.Emission diversity scheme of the present invention does not repel receive diversity, can introduce a plurality of reception antennas in the present invention and carry out receive diversity.

Below we analyze the principle and the performance of the time domain and frequency domain combined emission diversity method that proposes among the present invention.The application system structured flowchart of emission diversity method as shown in Figure 5.

Suppose the frequency domain input signal sequence be X (k, l), wherein k represents subcarrier sequence number (0≤k≤N-1, N are the sub-carrier number in the ofdm system), l represents the signal frame sequence number.In ofdm system, through anti-Discrete Fourier Transform (IDFT), the time-domain signal x that obtains _Tx1(n l) is

$x_{\mathrm{Tx}1}(n,l)=\frac{1}{N}\underset{k=0}{\overset{N-1}{\mathrm{\Σ}}}X(k,l)W_N^{-\mathrm{nk}},(0\≤n\≤N-1)$

Wherein, $W_N^k=e^{-j\frac{2\mathrm{\π}}{N}k}.$

(k l) is divided into odd number subsequence X according to the subcarrier sequence number with X _o(k is l) with even number subsequence X _e(k, l), their length is N/2, if note X _o(k, l) and X _e(k, l) result who does N/2 point IDFT conversion is x ₁ ^To(n, l) and x ₁ ^Te(n, l), time-domain signal x then _Tx1(n l) can change form and is expressed as:

$\left\{\begin{array}{c}{x}_{\mathrm{Tx}1}(n,l)=[x_1^{\mathrm{te}}(n,l)+x_1^{\mathrm{to}}(n,l)W_N^{-n}]/2\\ x_{\mathrm{Tx}1}(n+N/2,l)=[x_1^{\mathrm{te}}(n,l)-{\mathrm{<figure-callout\; id="1"\; label="x"\; filenames="A20051001212800252.png,A20051001212800262.png"\; state="\{\{state\}\}">x</figure-callout>}}_1^{\mathrm{to}}(n,l)W_N^{-n}]/2\end{array}\right.(0\&le;n\&le;N/2-1)$

This is x as a result _Tx1(n l) is exactly the time-domain signal that is used for first transmitting antenna Tx1.Simultaneously with x _Tx1(n, l), x _Tx2(n, l), x _Tx3(n, l), x _Tx4(n l) is saved in the buffer memory.

Similarly, the time-domain signal x that is used for second transmitting antenna Tx2 _Tx2(n, l) and x _Tx2(n l+1) can be expressed as:

$\left\{\begin{array}{c}{x}_{\mathrm{Tx}2}(n,l)=[x_2^{\mathrm{te}}(n,l)+x_2^{\mathrm{to}}(n,l)W_N^{-n}]/2\\ x_{\mathrm{Tx}2}(n+N/2,l)=[x_2^{\mathrm{te}}(n,l)-x_2^{\mathrm{to}}(n,l)W_N^{-n}]/2\\ x_{\mathrm{Tx}2}(n,l+1)=[x_2^{\mathrm{te}}(n,l+1)+x_2^{\mathrm{to}}(n,l+1)W_N^{-n}]/2\\ x_{\mathrm{Tx}2}(n+N/2,l+1)=[x_2^{\mathrm{te}}(n,l+1)-x_2^{\mathrm{to}}(n,l+1)W_N^{-n}]/2\end{array}\right.\left(0\&le;n\&le;N/2-1\right)$

X in the formula ₂ ^Te(n, l), x ₂ ^To(n, l), x ₂ ^Te(n, l+1) and x ₂ ^To(n is by obtaining behind the process space-frequency coding of the signal in the buffer memory (SFC) l+1):

$\left\{\begin{array}{c}{x}_2^{\mathrm{te}}(n,l)={x_1^{\mathrm{to}}}^{*}({(-n)}_{N/2},l)\\ x_2^{\mathrm{to}}(n,l)=-{x_1^{\mathrm{te}}}^{*}({(-n)}_{N/2},l)\\ x_2^{\mathrm{te}}(n,l+1)={x_1^{\mathrm{to}}}^{*}({(-n)}_{N/2},l+1)\\ x_2^{\mathrm{to}}(n,l+1)=-{x_1^{\mathrm{te}}}^{*}({(-n)}_{N/2},l+1)\end{array}\right.\left(0\&le;n\&le;N/2-1\right)$

Wherein, * represents the complex conjugate computing, (n) _N/2Expression is to n delivery N/2 computing.Computing character according to Discrete Fourier Transform (DFT) can get x ₂ ^Te(n, l), x ₂ ^To(n, l), x ₂ ^Te(n, l+1) and x ₂ ^To(n, N/2 point DFT transformation results l+1) can be written as:

$\left\{\begin{array}{c}{x}_2^{\mathrm{te}}(n,l)\stackrel{\mathrm{DFT}(N/2)}{\&RightArrow;}{X}_o^{*}(k,l)\\ x_2^{\mathrm{to}}(n,l)\stackrel{\mathrm{DFT}(N/2)}{\&RightArrow;}-X_e^{*}(k,l)\\ x_2^{\mathrm{te}}(n,l+1)\stackrel{\mathrm{DFT}(N/2)}{\&RightArrow;}{X}_o^{*}(k,l+1)\\ x_2^{\mathrm{to}}(n,l+1)\stackrel{\mathrm{DFT}(N/2)}{\&RightArrow;}-X_e^{*}(k,l+1)\end{array}\right.(0\&le;n,k\&le;N/2-1)$

Therefore, for second transmitting chain, its equivalent frequency domain input signal is

X _Tx2＝[X ^*(1)，-X ^*(0)…X ^*(2k+1)，-X ^*(2k)…X ^*(N-1)，-X ^*(N-2)](0≤k≤N/2-1)

Character shown in the following formula and OFDM frame number are irrelevant, so omitted second l in bracket.

For the 3rd transmitting chain Tx3, elder generation with the process of the signal in buffer memory Space Time Coding (STC), can get

$\left\{\begin{array}{c}{x}_3^{\mathrm{te}}(n,l)={x_1^{\mathrm{te}}}^{*}({(-n)}_{N/2},l+1)\\ x_3^{\mathrm{to}}(n,l)={x_1^{\mathrm{to}}}^{*}({(-n)}_{N/2},l+1)\\ x_3^{\mathrm{te}}(n,l+1)={-x_1^{\mathrm{te}}}^{*}({(-n)}_{N/2},l)\\ x_3^{\mathrm{to}}(n,l+1)=-{x_1^{\mathrm{to}}}^{*}({(-n)}_{N/2},l)\end{array}\right.\left(0\&le;n\&le;N/2-1\right)$

In like manner, use the computing character of DFT conversion, x ₃ ^Te(n, l), x ₃ ^To(n, l), x ₃ ^Te(n, l+1) and x ₃ ^To(n, the result of N/2 point DFT conversion l+1) can be expressed as

$\left\{\begin{array}{c}{x}_3^{\mathrm{te}}(n,l)\stackrel{\mathrm{DFT}(N/2)}{\&RightArrow;}{X}_e^{*}(k,l+1)\\ x_3^{\mathrm{to}}(n,l)\stackrel{\mathrm{DFT}(N/2)}{\&RightArrow;}{X}_o^{*}(k,l+1)\\ x_3^{\mathrm{te}}(n,l+1)\stackrel{\mathrm{DFT}(N/2)}{\&RightArrow;}{-X}_e^{*}(k,l)\\ x_3^{\mathrm{to}}(n,l+1)\stackrel{\mathrm{DFT}(N/2)}{\&RightArrow;}-X_o^{*}(k,l)\end{array}\right.(0\&le;n,k\&le;N/2-1)$

Therefore, the equivalent frequency domain input signal of the 3rd transmitting chain is

$\left\{\begin{array}{c}{X}_{\mathrm{Tx}3}(k,l)=X^{*}(k,l+1)\\ X_{\mathrm{Tx}3}(k,l+1)=-X^{*}(k,l)\end{array}\right.(0\&le;k\&le;N-1)$

Further, can get time-domain signal x _Tx3(n, l), x _Tx3(n l+1) is

$\left\{\begin{array}{c}{x}_{\mathrm{Tx}3}(n,l)=[x_3^{\mathrm{te}}(n,l)+x_3^{\mathrm{to}}(n,l)W_N^{-n}]/2\\ x_{\mathrm{Tx}3}(n+N/2,l)=[x_3^{\mathrm{te}}(n,l)-{\mathrm{<figure-callout\; id="3"\; label="x"\; filenames="A20051001212800251.png"\; state="\{\{state\}\}">x</figure-callout>}}_3^{\mathrm{to}}(n,l)W_N^{-n}]/2\\ x_{\mathrm{Tx}3}(n,l+1)=[x_3^{\mathrm{te}}(n,l+1)+x_3^{\mathrm{to}}(n,l+1)W_N^{-n}]/2\\ x_{\mathrm{Tx}3}(n+N/2,l+1)=[x_3^{\mathrm{te}}(n,l+1)-x_3^{\mathrm{to}}(n,l+1)W_N^{-n}]/2\end{array}\right.\left(0\&le;n\&le;N/2-1\right)$

Then, consider the 4th transmitting chain, will be through the data x that obtains behind the Space Time Coding ₃ ^Te(n, l), x ₃ ^To(n, l), x ₃ ^Te(n, l+1) and x ₃ ^To(n l+1) can get through space-frequency coding again

$\left\{\begin{array}{c}{x}_4^{\mathrm{te}}(n,l)={x_3^{\mathrm{to}}}^{*}({(-n)}_{N/2},l)\\ x_4^{\mathrm{to}}(n,l)=-{x_3^{\mathrm{te}}}^{*}({(-n)}_{N/2},l)\\ x_4^{\mathrm{te}}(n,l+1)={x_3^{\mathrm{to}}}^{*}({(-n)}_{N/2},l+1)\\ x_4^{\mathrm{to}}(n,l+1)=-{x_3^{\mathrm{te}}}^{*}({(-n)}_{N/2},l+1)\end{array}\right.\left(0\&le;n\&le;N/2-1\right)$

Same character according to the DFT transform operation, x ₄ ^Te(n, l), x ₄ ^To(n, l), x ₄ ^Te(n, l+1) and x ₄ ^To(n, the result of N/2 point DFT conversion l+1) can be expressed as

$\left\{\begin{array}{c}{x}_4^{\mathrm{te}}(n,l)\stackrel{\mathrm{DFT}(N/2)}{\&RightArrow;}{X_{\mathrm{Tx}3}^o}^{*}(k,l)=X_o(k,l+1)\\ x_2^{\mathrm{to}}(n,l)\stackrel{\mathrm{DFT}(N/2)}{\&RightArrow;}{{-X}_{\mathrm{Tx}3}^e}^{*}(k,l)=-X_e(k,l+1)\\ x_2^{\mathrm{te}}(n,l+1)\stackrel{\mathrm{DFT}(N/2)}{\&RightArrow;}{X_{\mathrm{Tx}3}^o}^{*}(k,l+1)=-X_o(k.l)\\ x_2^{\mathrm{to}}(n,l+1)\stackrel{\mathrm{DFT}(N/2)}{\&RightArrow;}{{-X}_{\mathrm{Tx}3}^e}^{*}(k,l+1)=X_e(k.l)\end{array}\right.(0\&le;n,k\&le;N/2-1)$

In the formula, X _Tx3 ^o(k, l), X _Tx3 ^e(k l) is respectively x ₃ ^To(n, l), x ₃ ^Te(n, the result of N/2 point DFT conversion l).Following formula is the equivalent frequency domain input signal of the 4th transmitting chain.

At last, obtain time-domain signal x _Tx4(n, l), x _Tx4(n l+1) is

$\left\{\begin{array}{c}{x}_{\mathrm{Tx}4}(n,l)=[x_4^{\mathrm{te}}(n,l)+x_4^{\mathrm{to}}(n,l)W_N^{-n}]/2\\ x_{\mathrm{Tx}4}(n+N/2,l)=[x_4^{\mathrm{te}}(n,l)-x_4^{\mathrm{to}}(n,l)W_N^{-n}]/2\\ x_{\mathrm{Tx}4}(n,l+1)=[x_4^{\mathrm{te}}(n,l+1)+x_4^{\mathrm{to}}(n,l+1)W_N^{-n}]/2\\ x_{\mathrm{Tx}4}(n+N/2,l+1)=[x_4^{\mathrm{te}}(n,l+1)-x_4^{\mathrm{to}}(n,l+1)W_N^{-n}]/2\end{array}\right.\left(0\&le;n\&le;N/2-1\right)$

The equivalent frequency domain input coding matrix that we can obtain four transmitting chains from top analysis is

$G=\left[\begin{array}{cccc}X(2k,l)& X^{*}(2k+1,l)& X^{*}(2k,l+1)& X(2k+1,l+1)\\ X(2k+1,l)& -X^{*}(2k,l)& X^{*}(2k+1,l+1)& -X(2k,l+1)\\ X(2k,l+1)& X^{*}(2k+1,l+1)& -X^{*}(2k,l)& -X(2k+1,l)\\ X(2k+1,l+1)& -X^{*}(2k,l+1)& -X^{*}(2k+1,l)& X(2k,l)\end{array}\right],(0\&le;k\&le;N/2-1);$

This is the STBC structure of an accurate quadrature (quosi-orthogonal) design.

Be without loss of generality, suppose to have only a receiver, so that analyze simple at receiving terminal.In order to obtain the channel estimation results of four transmitting chains respectively at receiving terminal, in the protection at interval of TDS-OFDM system, to insert different PN sequences respectively as frame head.In the analysis below, suppose that receiver can obtain the accurate channel information of every transmitting chain, the channel of the signal process of each transmitting antenna is incoherent, and is additive channel, and the signal of receiving with a reception antenna is each stack that transmits.After the OFDM demodulation, the sample value of received signal on 2k, a 2k+1 subcarrier can be expressed as:

$\left\{\begin{array}{c}{R}_l\left(2k\right)=H_{1,l}\left(2k\right)X(2k,l)+H_{2,l}\left(2k\right)X^{*}(2k+1,l)+\\ H_{3,l}\left(2k\right)X^{*}(2k,l+1)+H_{4,l}\left(2k\right)X(2k+1,l+1)+{\mathrm{\&eta;}}_l\left(2k\right)\\ R_l(2k+1)=H_{1,l}(2k+1)X(2k+1,l)-H_{2,l}(2k+1)X^{*}(2k,l)+\\ H_{3,l}(2k+1)X^{*}(2k+1,l+1)-H_{4,l}(2k+1)X(2k,l+1)+{\mathrm{\&eta;}}_l(2k+1)\\ R_{l+1}\left(2k\right)=H_{1,l+1}\left(2k\right)X(2k,l+1)+H_{2,l+1}\left(2k\right)X^{*}(2k+1,l+1)-\\ H_{3,l+1}\left(2k\right)X^{*}(2k,l)-H_{4,l+1}\left(2k\right)X(2k+1,l)+{\mathrm{\&eta;}}_{l+1}\left(2k\right)\\ R_{l+1}(2k+1)=H_{1,l+1}(2k+1)X(2k+1,l+1)-H_{2,l+1}(2k+1)X^{*}(2k,l+1)-\\ H_{3,l+1}(2k+1)X^{*}(2k+1,l)+H_{4,l+1}(2k+1)X(2k+l)+{\mathrm{\&eta;}}_{l+1}(2k+1)\end{array}\right.,(0\&le;k\&le;N/2-1)$

In the formula, H _{I, l}Be illustrated in the complex valued signals response vector of (time period represents to transmit a complete used time of OFDM frame, down together) i transmission link in l time period, η _lBe illustrated in complex value additive white Gaussian noise (AWGN) vector in l time period.

Suppose that the channel response between adjacent two time periods and adjacent two subcarriers is approximate identical, promptly

H _i(k)＝H _i，l(2k)≈H _i，l(2k+1)≈H _i，l+1(2k)≈H _i，l+1(2k+1)

(i＝1，2，3，4 0≤k≤N/2-1)

Then received signal can be simplified shown as

$\left[\begin{array}{c}{R}_l\left(2k\right)\\ {R_l}^{*}(2k+1)\\ {R_{l+1}}^{*}\left(2k\right)\\ R_{l+1}(2k+1)\end{array}\right]=\left[\begin{array}{cccc}{H}_1\left(k\right)& H_2\left(k\right)& H_3\left(k\right)& H_4\left(k\right)\\ -{H_2}^{*}\left(k\right)& {{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="A20051001212800252.png,A20051001212800262.png"\; state="\{\{state\}\}">H</figure-callout>}}_1}^{*}\left(k\right)& -{H_4}^{*}\left(k\right)& {{\mathrm{<figure-callout\; id="3"\; label="H"\; filenames="A20051001212800251.png"\; state="\{\{state\}\}">H</figure-callout>}}_3}^{*}\left(k\right)\\ -{{\mathrm{<figure-callout\; id="3"\; label="H"\; filenames="A20051001212800251.png"\; state="\{\{state\}\}">H</figure-callout>}}_3}^{*}\left(k\right)& -{H_4}^{*}\left(k\right)& {{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="A20051001212800252.png,A20051001212800262.png"\; state="\{\{state\}\}">H</figure-callout>}}_1}^{*}\left(k\right)& {H_2}^{*}\left(k\right)\\ H_4\left(k\right)& -H_3\left(k\right)& -H_2\left(k\right)& H_1\left(k\right)\end{array}\right]\left[\begin{array}{c}X(2k,l)\\ X^{*}(2k+1,l)\\ X^{*}(2k,l+1)\\ X(2k+1,l+1)\end{array}\right]+\left[\begin{array}{c}{\mathrm{\&eta;}}_l\left(2k\right)\\ {{\mathrm{\&eta;}}_l}^{*}(2k+1)\\ {{\mathrm{\&eta;}}_{l+1}}^{*}\left(2k\right)\\ {\mathrm{\&eta;}}_{l+1}(2k+1)\end{array}\right],(0\&le;k\&le;N/2-1)$

The note channel response matrix is H (k), promptly

$H\left(k\right)=\left[\begin{array}{cccc}{H}_1\left(k\right)& H_2\left(k\right)& H_3\left(k\right)& H_4\left(k\right)\\ -{H_2}^{*}\left(k\right)& {{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="A20051001212800252.png,A20051001212800262.png"\; state="\{\{state\}\}">H</figure-callout>}}_1}^{*}\left(k\right)& -{H_4}^{*}\left(k\right)& {{\mathrm{<figure-callout\; id="3"\; label="H"\; filenames="A20051001212800251.png"\; state="\{\{state\}\}">H</figure-callout>}}_3}^{*}\left(k\right)\\ -{{\mathrm{<figure-callout\; id="3"\; label="H"\; filenames="A20051001212800251.png"\; state="\{\{state\}\}">H</figure-callout>}}_3}^{*}\left(k\right)& -{H_4}^{*}\left(k\right)& {{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="A20051001212800252.png,A20051001212800262.png"\; state="\{\{state\}\}">H</figure-callout>}}_1}^{*}\left(k\right)& {H_2}^{*}\left(k\right)\\ H_4\left(k\right)& -H_3\left(k\right)& -H_2\left(k\right)& H_1\left(k\right)\end{array}\right](0\&le;k\&le;N/2-1)$

Hermit transformation matrix H with the equal premultiplication H in the two ends of received signal expression formula (k) ^H(k), can get final signal estimated value X ' (2k, l), X ' (2k+1, l), X ' (2k, l+1) and X ' (2k+1 l+1) is:

$\left[\begin{array}{c}{X}^{\&prime;}(2k,l)\\ X^{\&prime;*}(2k+1,l)\\ X^{\&prime;*}(2k,l+1)\\ X^{\&prime;}(2k+1,l+1)\end{array}\right]=\left[\begin{array}{cccc}a\left(k\right)& 0& 0& b\left(k\right)\\ 0& a\left(k\right)& -b\left(k\right)& 0\\ 0& -b\left(k\right)& a\left(k\right)& 0\\ b\left(k\right)& 0& 0& a\left(k\right)\end{array}\right]\left[\begin{array}{c}X(2k,l)\\ X^{*}(2k+1,l)\\ X^{*}(2k,l+1)\\ X(2k+1,l+1)\end{array}\right]+\left[\begin{array}{c}{\mathrm{\&eta;}}_l^{\&prime;}\left(2k\right)\\ {\mathrm{\&eta;}}_l^{\&prime;*}(2k+1)\\ {{\mathrm{\&eta;}}_{l+1}^{\&prime;}}^{*}\left(2k\right)\\ {\mathrm{\&eta;}}_{l+1}^{\&prime;}(2k+1)\end{array}\right](*),(0\&le;k\&le;N/2-1)$

Wherein, $a\left(k\right)=\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{\left|H_i\left(k\right)\right|}^2,$ B (k)=2Re (H ₁(k) H ₄ ^*(k)-H ₂(k) H ₃ ^*(k)), η _l' remain complex value additive white Gaussian noise (AWGN) vector.

As seen, carry out maximum likelihood (ML) when decoding, deterministic can be divided into two parts at receiving terminal: (X (2k, l), X (2k+1, l+1)) and ((2k+1, l), X (2k, l+1)) like this can be so that the operand of deciphering reduces X greatly.Can also see that from top analysis though adopt transmit diversity structure diversity gain proposed by the invention to reduce to some extent, keeping code efficiency is 1.

The time domain and frequency domain combined emission diversity method structure (Fig. 4) that proposes according to the present invention is utilized the computing character of DFT conversion, can see the time-domain signal x of four transmitting chains in the superincumbent analysis _Tx1(n, l), x _Tx2(n, l), x _Tx3(n, l), x _Tx4(n l) can be by the data x in the buffer memory ₁ ^To(n, l), x ₁ ^Te(n, l), x ₁ ^To(n, l+1) and x ₁ ^Te(n l+1) comes out by simple complex multiplication and additional calculation, also promptly on average does the DFT computing of a N point and 3N/2 complex multiplication and 3N complex addition in the time period of an OFDM frame.And in traditional STC-OFDM (" Lee K and Williams D.A space-time coded transmitterdiversity technique for frequency selective fading channels.in Proc.IEEE Sensor Array andMultichannel Signal Processing Workshop.Cambridge; MA:IEEE; 2000.149-152 ") and SFC-OFDM (" Lee K and Williams D.A space-frequency transmitter diversity technique for OFDM systems.inProc.IEEE GLOBECOM ' 00.SanFrancisco; CA:IEEE; 2000; 1473-1477 ") system, in the time period of an OFDM frame, on average do four N point DFT computing.Therefore, when the sub-carrier number N of ofdm system was enough big, the operand of the method that the present invention proposes was about 1/4 of original method.Especially, be example with the TDS-OFDM system, N=3780, the adopting said method operand can reduce about 70%.

Usually, in traditional STC-OFDM system, be about to continuous k OFDM symbol and shown in certain coded format G, encode.At this moment, in order to keep the character of transmission matrix G quadrature (or accurate quadrature), must suppose that channel is quasi-static, promptly channel remains unchanged in continuous k adjacent OFDM symbol time.This hypothesis can produce very mistake under fast fading channel, thereby picture is also inapplicable under the situations such as mobile reception of DTTB.And in traditional SFC-OFDM system, with the digital coding on the adjacent k subcarrier.Like this, in order to keep the character of transmission matrix G quadrature (or accurate quadrature), need the frequency response on k the adjacent subcarrier of hypothesis identical.It is same because the error that the channel hypothesis is brought is also inapplicable in frequency selective fading channels.And in esse channel circumstance is double selectivity (time and frequency) mostly, so STC-OFDM and SFC-OFDM system all can bring than mistake in actual applications.The time domain and frequency domain combined emission diversity method based on STBC that the present invention proposes need only suppose that the channel response between adjacent two time periods and adjacent two subcarriers is approximate identical, therefore is applicable to actual time, frequency dual-selection channel more.

Based on above-mentioned analysis, a kind of time domain and frequency domain combined emission diversity method based on space-time block code (STBC) in the TDS-OFDM system proposed by the invention has been carried out Computer Simulation, the system configuration of emulation is identical with Fig. 5.We adopt two kinds of channel model A shown in table 1 and the table 2 and B in emulation.Wherein the multipath delay of model A is shorter, and echo strength a little less than; Introduced strong multipath (multipath 6) in the Model B with long delay.

Table 1 diversity transmission channel simulation model A

Model A	Multipath	1	Multipath 2	Multipath 3	Multipath 4	Multipath 5	Multipath 6
								Type amplitude (dB) time-delay (us)	Rayleigh 0 0	Rayleigh -12 0.3	Rayleigh -4 3.5	Rayleigh -7 4.4	Rayleigh -15 9.5	Rayleigh -22 12.7

Table 2 diversity transmission channel simulation model B

Model B	Multipath	1	Multipath 2	Multipath 3	Multipath 4	Multipath 5	Multipath 6
								Type amplitude (dB) time-delay (us)	Rayleigh 0 0	Rayleigh -12 0.3	Rayleigh -4 3.5	Rayleigh -7 4.4	Rayleigh -15 9.5	Rayleigh 0 30

Adopt QPSK planisphere and protection to be spaced apart 3780 TDS-OFDM systems of data length 1/9 in the emulation, chnnel coding adopts the convolution code of 2/3 code check.And suppose that receiver can obtain the accurate channel information of every transmitting chain, the channel of the signal process of each transmitting antenna is uncorrelated.Fig. 6～9 have provided diversity not, traditional STC-OFDM, SFC-OFDM and bit error rate (BER) simulation result of deversity scheme under the different channels situation proposed by the invention.In order to make simulation result have comparativity, signal to noise ratio among the figure (Signal to Noise Ratio, SNR) be as the criterion with reception antenna, that is to say, the power of each transmitting antenna only is 1/4 of the middle transmitting power of single transmit antenna method (not diversity) in the emission diversity method.

Fig. 6 adopts channel model A for prolonging slow fading channel in short-term, does not add doppler shift effect.Three kinds of transmit diversity systems are compared diversity system not all clearly gain, and systematic function is approximate identical.

Fig. 7 still adopts channel model A for prolonging in short-term under the fast fading channel, maximum doppler frequency f _d=50Hz.Because the influence of channel hypothesis error, the STC-OFDM system can not provide diversity gain, and hierarchy system is also not poor for its performance even ratio.And the diversity scheme that SFC-OFDM system and the present invention are carried shows insensitive to Doppler frequency shift, has still kept higher gain.

Fig. 8 is the long delay slow fading channel, adopts channel model B, does not add doppler shift effect.As can be seen, because the influence of channel hypothesis error, the SFC-OFDM system can not provide diversity gain.And the diversity scheme that SFC-OFDM system and the present invention are carried still can provide higher gain.

Existing long delay of channel shown in Figure 9 and strong multipath have rapid fading again, adopt channel model B, maximum doppler frequency f _d=50Hz.Same because channel is supposed the influence of error, and STC-OFDM and SFC-OFDM system all can't provide diversity gain, and the diversity scheme that the present invention carried still can obtain higher gain.

In the TDS-OFDM system proposed by the invention based on the strong robustness of the time domain and frequency domain combined emission diversity method diversity system work of space-time block code (STBC), support " fail soft (Soft Failure) ", promptly undesired as the work of fruit part transmission link, need not carry out any modification to original system, receiving terminal still can normally be deciphered, and error performance is not less than the situation under the single-shot transmitter system at least.This be because, if certain transmitting chain can't operate as normal, in last judgement formula (*), the element zero setting in the corresponding channel response matrix need only get final product so, still can normally carry out ML and decipher.In the present invention, if any transmitting chain breaks down, its excess-three transmitting chain still can be formed accurate quadrature STBC structure so, thereby obtains diversity gain.When having only two transmitting chains working properly, can be divided into three kinds of situations: 1) Tx1 and Tx3 (or Tx2 and Tx4), can form the quadrature STC-OFDM structure of one 2 antenna; 2) Tx1 and Tx2 (or Tx3 and Tx4) can form the quadrature SFC-OFDM structure of one 2 antenna; 3) Tx1 and Tx4 (or Tx2 and Tx3).1) and 2) under the situation, system still can obtain diversity gain.Figure 10 has provided the computer artificial result under the situation of transmitter of the operate as normal that has different numbers in the system, adopts channel model A during emulation, does not add doppler shift effect.Simulation result has well been verified above-mentioned analysis result.

Description of drawings

Fig. 1 receive diversity scheme block diagram

Fig. 2 is a TDS-OFDM system level frame structure.

Fig. 3 is a TDS-OFDM system signal frame assumption diagram.

The time domain and frequency domain combined emission diversity method block diagram that Fig. 4 proposes for the present invention.

Fig. 5 is the application system structured flowchart of the emission diversity method of the present invention's proposition.

Fig. 6 is the contrast simulation result (f of the present invention to channel model A _d=0Hz).

Fig. 7 is the contrast simulation result (f of the present invention to channel model A _d=50Hz).

Fig. 8 is the contrast simulation result (f of the present invention to channel model B _d=0Hz).

Fig. 9 is the contrast simulation result (f of the present invention to channel model B _d=50Hz).

There is the contrast simulation result under the situation of transmitter of operate as normal of different numbers in the diversity system that Figure 10 proposes for the present invention.

Embodiment

See Fig. 4.Frequency domain list entries in the native system is the complex signal after ovennodulation mapping (planisphere mapping), and it can be non-code signal, also can be through the signal behind the coding.The frequency domain list entries at first is divided into odd number subsequence and even number subsequence according to the subcarrier sequence number, and their length is N/2.Then two continuous frames is imported the long subsequence of totally four N/2 of data respectively as the anti-Discrete Fourier Transform of N/2 point, it is x as a result ₁ ^To(n, l), x ₁ ^Te(n, l), x ₁ ^To(n, l+1) and x ₁ ^Te(n l+1) is deposited in the buffer memory and deposits in the buffer memory.Utilize the computing character of anti-Discrete Fourier Transform, divide following four kinds of situations that the data in the buffer memory are carried out simple complex multiplication and additional calculation, obtain the time-domain signal of four transmitting chains respectively:

1) for antenna Tx1:

$\left\{\begin{array}{c}{x}_{\mathrm{Tx}1}(n,l)=[x_1^{\mathrm{te}}(n,l)+x_1^{\mathrm{to}}(n,l)W_N^{-n}]/2\\ x_{\mathrm{Tx}1}(n+N/2,l)=[x_1^{\mathrm{te}}(n,l)-{\mathrm{<figure-callout\; id="1"\; label="x"\; filenames="A20051001212800252.png,A20051001212800262.png"\; state="\{\{state\}\}">x</figure-callout>}}_1^{\mathrm{to}}(n,l)W_N^{-n}]/2\\ x_{\mathrm{Tx}1}(n,l+1)=[x_1^{\mathrm{te}}(n,l+1)+x_1^{\mathrm{to}}(n,l+1)W_N^{-n}]/2\\ x_{\mathrm{Tx}1}(n+N/2,l+1)=[x_1^{\mathrm{te}}(n,l+1)-x_1^{\mathrm{to}}(n,l+1)W_N^{-n}]/2\end{array}\right.\left(0\&le;n\&le;N/2-1\right)$

2) for antenna Tx2, earlier through space-frequency coding:

$\left\{\begin{array}{c}{x}_2^{\mathrm{te}}(n,l)={x_1^{\mathrm{to}}}^{*}({(-n)}_{N/2},l)\\ x_2^{\mathrm{to}}(n,l)=-{x_1^{\mathrm{te}}}^{*}({(-n)}_{N/2},l)\\ x_2^{\mathrm{te}}(n,l+1)={x_1^{\mathrm{to}}}^{*}({(-n)}_{N/2},l+1)\\ x_2^{\mathrm{to}}(n,l+1)=-{x_1^{\mathrm{te}}}^{*}({(-n)}_{N/2},l+1)\end{array}\right.\left(0\&le;n\&le;N/2-1\right)$

Calculate time-domain signal then:

$\left\{\begin{array}{c}{x}_{\mathrm{Tx}2}(n,l)=[x_2^{\mathrm{te}}(n,l)+x_2^{\mathrm{to}}(n,l)W_N^{-n}]/2\\ x_{\mathrm{Tx}2}(n+N/2,l)=[x_2^{\mathrm{te}}(n,l)-x_2^{\mathrm{to}}(n,l)W_N^{-n}]/2\\ x_{\mathrm{Tx}2}(n,l+1)=[x_2^{\mathrm{te}}(n,l+1)+x_2^{\mathrm{to}}(n,l+1)W_N^{-n}]/2\\ x_{\mathrm{Tx}2}(n+N/2,l+1)=[x_2^{\mathrm{te}}(n,l+1)-x_2^{\mathrm{to}}(n,l+1)W_N^{-n}]/2\end{array}\right.\left(0\&le;n\&le;N/2-1\right)$

3) for antenna Tx3, earlier through Space Time Coding:

$\left\{\begin{array}{c}{x}_3^{\mathrm{te}}(n,l)={x_1^{\mathrm{te}}}^{*}({(-n)}_{N/2},l+1)\\ x_3^{\mathrm{to}}(n,l)={x_1^{\mathrm{to}}}^{*}({(-n)}_{N/2},l+1)\\ x_3^{\mathrm{te}}(n,l+1)={-x_1^{\mathrm{te}}}^{*}({(-n)}_{N/2},l)\\ x_3^{\mathrm{to}}(n,l+1)=-{x_1^{\mathrm{to}}}^{*}({(-n)}_{N/2},l)\end{array}\right.\left(0\&le;n\&le;N/2-1\right)$

Calculate time-domain signal then:

$\left\{\begin{array}{c}{x}_{\mathrm{Tx}3}(n,l)=[x_3^{\mathrm{te}}(n,l)+x_3^{\mathrm{to}}(n,l)W_N^{-n}]/2\\ x_{\mathrm{Tx}3}(n+N/2,l)=[x_3^{\mathrm{te}}(n,l)-{\mathrm{<figure-callout\; id="3"\; label="x"\; filenames="A20051001212800251.png"\; state="\{\{state\}\}">x</figure-callout>}}_3^{\mathrm{to}}(n,l)W_N^{-n}]/2\\ x_{\mathrm{Tx}3}(n,l+1)=[x_3^{\mathrm{te}}(n,l+1)+x_3^{\mathrm{to}}(n,l+1)W_N^{-n}]/2\\ x_{\mathrm{Tx}3}(n+N/2,l+1)=[x_3^{\mathrm{te}}(n,l+1)-x_3^{\mathrm{to}}(n,l+1)W_N^{-n}]/2\end{array}\right.\left(0\&le;n\&le;N/2-1\right)$

4), will pass through space-frequency coding again through the data behind the Space Time Coding for antenna Tx4:

$\left\{\begin{array}{c}{x}_4^{\mathrm{te}}(n,l)={x_3^{\mathrm{to}}}^{*}({(-n)}_{N/2},l)\\ x_4^{\mathrm{to}}(n,l)=-{x_3^{\mathrm{te}}}^{*}({(-n)}_{N/2},l)\\ x_4^{\mathrm{te}}(n,l+1)={x_3^{\mathrm{to}}}^{*}({(-n)}_{N/2},l+1)\\ x_4^{\mathrm{to}}(n,l+1)=-{x_3^{\mathrm{te}}}^{*}({(-n)}_{N/2},l+1)\end{array}\right.\left(0\&le;n\&le;N/2-1\right)$

Calculate time-domain signal then:

$\left\{\begin{array}{c}{x}_{\mathrm{Tx}4}(n,l)=[x_4^{\mathrm{te}}(n,l)+x_4^{\mathrm{to}}(n,l)W_N^{-n}]/2\\ x_{\mathrm{Tx}4}(n+N/2,l)=[x_4^{\mathrm{te}}(n,l)-x_4^{\mathrm{to}}(n,l)W_N^{-n}]/2\\ x_{\mathrm{Tx}4}(n,l+1)=[x_4^{\mathrm{te}}(n,l+1)+x_4^{\mathrm{to}}(n,l+1)W_N^{-n}]/2\\ x_{\mathrm{Tx}4}(n+N/2,l+1)=[x_4^{\mathrm{te}}(n,l+1)-x_4^{\mathrm{to}}(n,l+1)W_N^{-n}]/2\end{array}\right.\left(0\&le;n\&le;N/2-1\right)$

Channel frame structure according to the TDS-OFDM system; in the TDS-OFDM of four transmitting chains protection at interval, insert different PN sequences respectively, the frame that obtains in frame head PN sequence and the above-mentioned steps is formed four signal frames that transmitting chain is complete separately respectively as frame head.And then complete TDS-OFDM signal is formed filtering and digital to analog conversion handle, pass through front-end processings such as frequency up-converted and power amplifier then, launch in predetermined channel bandwidth by four antennas respectively at last, finish transmission antenna diversity.

At receiving terminal, suppose that the channel response between adjacent two time periods and adjacent two subcarriers is approximate identical, after the OFDM demodulation, the sample value of received signal on 2k, a 2k+1 subcarrier can be expressed as:

$\left[\begin{array}{c}{R}_l\left(2k\right)\\ {R_l}^{*}(2k+1)\\ {R_{l+1}}^{*}\left(2k\right)\\ R_{l+1}(2k+1)\end{array}\right]=\left[\begin{array}{cccc}{H}_1\left(k\right)& H_2\left(k\right)& H_3\left(k\right)& H_4\left(k\right)\\ -{H_2}^{*}\left(k\right)& {{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="A20051001212800252.png,A20051001212800262.png"\; state="\{\{state\}\}">H</figure-callout>}}_1}^{*}\left(k\right)& -{H_4}^{*}\left(k\right)& {{\mathrm{<figure-callout\; id="3"\; label="H"\; filenames="A20051001212800251.png"\; state="\{\{state\}\}">H</figure-callout>}}_3}^{*}\left(k\right)\\ -{{\mathrm{<figure-callout\; id="3"\; label="H"\; filenames="A20051001212800251.png"\; state="\{\{state\}\}">H</figure-callout>}}_3}^{*}\left(k\right)& -{H_4}^{*}\left(k\right)& {{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="A20051001212800252.png,A20051001212800262.png"\; state="\{\{state\}\}">H</figure-callout>}}_1}^{*}\left(k\right)& {H_2}^{*}\left(k\right)\\ H_4\left(k\right)& -H_3\left(k\right)& -H_2\left(k\right)& H_1\left(k\right)\end{array}\right]\left[\begin{array}{c}X(2k,l)\\ X^{*}(2k+1,l)\\ X^{*}(2k,l+1)\\ X(2k+1,l+1)\end{array}\right]+\left[\begin{array}{c}{\mathrm{\&eta;}}_l\left(2k\right)\\ {{\mathrm{\&eta;}}_l}^{*}(2k+1)\\ {{\mathrm{\&eta;}}_{l+1}}^{*}\left(2k\right)\\ {\mathrm{\&eta;}}_{l+1}(2k+1)\end{array}\right],(0\&le;k\&le;N/2-1)$

The note channel response matrix is H (k), promptly

$H\left(k\right)=\left[\begin{array}{cccc}{H}_1\left(k\right)& H_2\left(k\right)& H_3\left(k\right)& H_4\left(k\right)\\ -{H_2}^{*}\left(k\right)& {{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="A20051001212800252.png,A20051001212800262.png"\; state="\{\{state\}\}">H</figure-callout>}}_1}^{*}\left(k\right)& -{H_4}^{*}\left(k\right)& {{\mathrm{<figure-callout\; id="3"\; label="H"\; filenames="A20051001212800251.png"\; state="\{\{state\}\}">H</figure-callout>}}_3}^{*}\left(k\right)\\ -{{\mathrm{<figure-callout\; id="3"\; label="H"\; filenames="A20051001212800251.png"\; state="\{\{state\}\}">H</figure-callout>}}_3}^{*}\left(k\right)& -{H_4}^{*}\left(k\right)& {{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="A20051001212800252.png,A20051001212800262.png"\; state="\{\{state\}\}">H</figure-callout>}}_1}^{*}\left(k\right)& {H_2}^{*}\left(k\right)\\ H_4\left(k\right)& -H_3\left(k\right)& -H_2\left(k\right)& H_1\left(k\right)\end{array}\right](0\&le;k\&le;N/2-1)$

Hermit transformation matrix H with the equal premultiplication H in the two ends of received signal expression formula (k) ^H(k), can get final signal estimated value X ' (2k, l), X ' (2k+1, l), X ' (2k, l+1) and X ' (2k+1 l+1) is:

$\left[\begin{array}{c}{X}^{\&prime;}(2k,l)\\ X^{\&prime;*}(2k+1,l)\\ X^{\&prime;*}(2k,l+1)\\ X^{\&prime;}(2k+1,l+1)\end{array}\right]=\left[\begin{array}{cccc}a\left(k\right)& 0& 0& b\left(k\right)\\ 0& a\left(k\right)& -b\left(k\right)& 0\\ 0& -b\left(k\right)& a\left(k\right)& 0\\ b\left(k\right)& 0& 0& a\left(k\right)\end{array}\right]\left[\begin{array}{c}X(2k,l)\\ X^{*}(2k+1,l)\\ X^{*}(2k,l+1)\\ X(2k+1,l+1)\end{array}\right]+\left[\begin{array}{c}{\mathrm{\&eta;}}_l^{\&prime;}\left(2k\right)\\ {\mathrm{\&eta;}}_l^{\&prime;*}(2k+1)\\ {{\mathrm{\&eta;}}_{l+1}^{\&prime;}}^{*}\left(2k\right)\\ {\mathrm{\&eta;}}_{l+1}^{\&prime;}(2k+1)\end{array}\right],(0\&le;k\&le;N/2-1)$

Wherein, $a\left(k\right)=\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{\left|H_i\left(k\right)\right|}^2,$ B (k)=2Re (H ₁(k) H ₄ ^*(k)-H ₂(k) H ₃ ^*(k)), η _l' remain complex value additive white Gaussian noise (AWGN) vector.Carrying out maximum likelihood (ML) when decoding, deterministic is being divided into two parts: (X (2k, l), X (2k+1, l+1)) and ((2k+1, l), X (2k, l+1)) like this can be so that the operand of deciphering reduces X greatly.

Above general specific embodiment of the present invention be illustrated, but the present invention is not restricted to the foregoing description, under the spirit and scope situation of the claim that does not break away from the application, those skilled in the art can make various modifications or remodeling.

Claims (4)

1, the emission diversity method of time-domain synchronous orthogonal frequency-division multiplexing system is characterized in that, described method is a kind of time domain and frequency domain combined emission diversity method based on space-time block code, and it is realized successively according to following steps in the special digital integrated circuit:

Step 1. note frequency domain list entries be X (k, l), wherein k represents the subcarrier sequence number, 0≤k≤N-1, N are the sub-carrier number in the ofdm system, l represents the signal frame sequence number, (k l) is divided into odd number subsequence X according to the subcarrier sequence number with X _o(k is l) with even number subsequence X for x _e(k, l), their length is N/2;

Step 2. is with X _o(k, l) and X _e(k l) makes N/2 point Inverse Discrete Fourier Transform respectively, and the time domain sequences that obtains is x ₁ ^To(n, l) and x ₁ ^Te(n, l);

Step 3. is imported x as a result after data are carried out Inverse Discrete Fourier Transform with two continuous frames ₁ ^To(n, l), x ₁ ^Te(n, l), x ₁ ^To(n, l+1) and x ₁ ^Te(n l+1) is deposited in the buffer memory;

Step 4. is carried out nonidentity operation with the data in the buffer memory according to following described four kinds of situations then, obtains being used for four required time-domain signals of antenna emission respectively:

(a) for first transmitting antenna Tx1, time-domain signal x _Tx1(n, l), x _Tx1(n l+1) is

$\left\{\begin{array}{c}{x}_{\mathrm{Tx}1}(n,l)=[x_1^{\mathrm{te}}(n,l)+x_1^{\mathrm{to}}(n,l)W_N^{-n}]/2,\\ x_{\mathrm{Tx}1}(n+N/2,l)=[x_1^{\mathrm{te}}(n,l)-x_1^{\mathrm{to}}(n,l)W_N^{-n}]/2,\\ x_{\mathrm{Tx}1}(n,l+1)=[x_1^{\mathrm{te}}(n,l+1)+x_1^{\mathrm{to}}(n,l+1)W_N^{-n}]/2,\\ x_{\mathrm{Tx}1}(n+N/2,l+1)=[x_1^{\mathrm{te}}(n,l+1)-x_1^{\mathrm{to}}(n,l+1)W_N^{-n}]/2,\end{array}0\&le;n\&le;N/2-1;\right.$

Wherein, $W_N^{-n}=e^{j\frac{2\mathrm{\&pi;}}{N}n},$ N is the sub-carrier number in the ofdm system;

(b), earlier the data in the buffer memory are obtained through behind the space-frequency coding for second transmitting antenna Tx2

$\left\{\begin{array}{c}{x}_2^{\mathrm{te}}(n,l)=x_1^{{\mathrm{to}}^{*}}({(-n)}_{N/2},l),\\ x_2^{\mathrm{to}}(n,l)=-x_1^{t{e}^{*}}({(-n)}_{N/2},l),\\ x_2^{\mathrm{te}}(n,l+1)=x_1^{{\mathrm{to}}^{*}}({(-n)}_{N/2},l+1),\\ x_2^{\mathrm{to}}(n,l+1)=-x_1^{{\mathrm{te}}^{*}}({(-n)}_{N/2},l+1),\end{array}\right.0\&le;n\&le;N/2-1;$

Wherein, * represents the complex conjugate computing, (n) _N/2Expression obtains time-domain signal x then to n delivery N/2 computing _Tx2(n, l), x _Tx2(n l+1) is

$\left\{\begin{array}{c}{x}_{\mathrm{Tx}2}(n,l)=[x_2^{\mathrm{te}}(n,l)+x_2^{\mathrm{to}}(n,l)W_N^{-n}]/2,\\ x_{\mathrm{Tx}2}(n+N/2,l)=[x_2^{\mathrm{te}}(n,l)-x_2^{\mathrm{to}}(n,l)W_N^{-n}]/2,\\ x_{\mathrm{Tx}2}(n,l+1)=[x_2^{\mathrm{te}}(n,l+1)+x_2^{\mathrm{to}}(n,l+1)W_N^{-n}]/2,\\ x_{\mathrm{Tx}2}(n+N/2,l+1)=[x_2^{\mathrm{te}}(n,l+1)-x_2^{\mathrm{to}}(n,l+1)W_N^{-n}]/2,\end{array}0\&le;n\&le;N/2-1;\right.$

(c), earlier the data in the buffer memory are obtained through behind the Space Time Coding for the 3rd transmitting antenna Tx3

$\left\{\begin{array}{c}{x}_3^{\mathrm{te}}(n,l)=x_1^{{\mathrm{te}}^{*}}({(-n)}_{N/2},l+1),\\ x_3^{\mathrm{to}}(n,l)=x_1^{t{o}^{*}}({(-n)}_{N/2},l+1),\\ x_3^{\mathrm{te}}(n,l+1)=-x_1^{{\mathrm{te}}^{*}}({(-n)}_{N/2},l),\\ x_3^{\mathrm{to}}(n,l+1)=-x_1^{{\mathrm{to}}^{*}}({(-n)}_{N/2},l),\end{array}\right.0\&le;n\&le;N/2-1;$

Obtain time-domain signal x then _Tx3(n, l), x _Tx3(n l+1) is

$\left\{\begin{array}{c}{x}_{\mathrm{Tx}3}(n,l)=[x_3^{\mathrm{te}}(n,l)+x_3^{\mathrm{to}}(n,l)W_N^{-n}]/2,\\ x_{\mathrm{Tx}3}(n+N/2,l)=[x_3^{\mathrm{te}}(n,l)-x_3^{\mathrm{to}}(n,l)W_N^{-n}]/2,\\ x_{\mathrm{Tx}3}(n,l+1)=[x_3^{\mathrm{te}}(n,l+1)+x_3^{\mathrm{to}}(n,l+1)W_N^{-n}]/2,\\ x_{\mathrm{Tx}3}(n+N/2,l+1)=[x_3^{\mathrm{te}}(n,l+1)-x_3^{\mathrm{to}}(n,l+1)W_N^{-n}]/2,\end{array}0\&le;n\&le;N/2-1;\right.$

(d) for the 4th transmitting antenna Tx4, the data in the buffer memory through Space Time Coding, obtain the x as a result as shown in (c) earlier ₃ ^To(n, l), x ₃ ^Te(n, l), x ₃ ^To(n, l+1) and x ₃ ^Te(n, l+1), and then the process space-frequency coding obtains

$\left\{\begin{array}{c}{x}_4^{\mathrm{te}}(n,l)=x_3^{{\mathrm{to}}^{*}}({(-n)}_{N/2},l),\\ x_4^{\mathrm{to}}(n,l)={-x}_3^{t{e}^{*}}({(-n)}_{N/2},l),\\ x_4^{\mathrm{te}}(n,l+1)=x_3^{{\mathrm{to}}^{*}}({(-n)}_{N/2},l+1),\\ x_4^{\mathrm{to}}(n,l+1)=-x_1^{{\mathrm{te}}^{*}}({(-n)}_{N/2},l+1),\end{array}\right.0\&le;n\&le;N/2-1;$

Obtain time-domain signal x at last _Tx4(n, l), x _Tx4(n l+1) is

$\left\{\begin{array}{c}{x}_{\mathrm{Tx}4}(n,l)=[x_4^{\mathrm{te}}(n,l)+x_4^{\mathrm{to}}(n,l)W_N^{-n}]/2,\\ x_{\mathrm{Tx}4}(n+N/2,l)=[x_4^{\mathrm{te}}(n,l)-x_4^{\mathrm{to}}(n,l)W_N^{-n}]/2,\\ x_{\mathrm{Tx}4}(n,l+1)=[x_4^{\mathrm{te}}(n,l+1)+x_4^{\mathrm{to}}(n,l+1)W_N^{-n}]/2,\\ x_{\mathrm{Tx}4}(n+N/2,l+1)=[x_4^{\mathrm{te}}(n,l+1)-x_4^{\mathrm{to}}(n,l+1)W_N^{-n}]/2,\end{array}0\&le;n\&le;N/2-1;\right.$

Step 5. generates the PN sequence of four different respective length according to the length 420 or 945 of time-domain synchronous orthogonal frequency-division multiplexing system signal frame frame head;

Step 6. is according to the channel frame structure of time-domain synchronous orthogonal frequency-division multiplexing system; in the TDS-OFDM of transmitting antenna Tx1, Tx2, four links of Tx3 and Tx4 protection at interval, insert above-mentioned different PN sequence respectively, the frame x that frame head PN sequence and step (4) are obtained as frame head _Tx1(n, l), x _Tx2(n, l), x _Tx3(n, l), x _Tx4(n l) forms four signal frames that transmitting chain is complete separately respectively;

Above-mentioned complete TDS-OFDM signal is formed filtering to step 7. and digital to analog conversion is handled, then through comprising frequency up-converted and power amplifier in interior front-end processing, launch in predetermined channel bandwidth by antenna Tx1, Tx2, Tx3 and Tx4 respectively at last, finish transmission antenna diversity.

2, the emission diversity method of time-domain synchronous orthogonal frequency-division multiplexing system according to claim 1 is characterized in that, the equivalent space-time block code structure of described time domain and frequency domain combined coding is expressed as with matrix G:

$G=\left[\begin{array}{cccc}X(2k,l)& X^{*}(2k+1,l)& X^{*}(2k,l+1)& X(2k+1,l+1)\\ X(2k+1,l)& {-X}^{*}(2k,l)& X^{*}(2k+1,l+1)& -X(2k,l+1)\\ X(2k,l+1)& X^{*}(2k+1,l+1)& -X^{*}(2k,l)& -X(2k+1,l)\\ X(2k+1,l+1)& {-X}^{*}(2k,l+1)& {-X}^{*}(2k+1,l)& X(2k,l)\end{array}\right],0\&le;k\&le;N/2-1;$

Wherein, (k l) is the frequency domain list entries to X, and k represents the subcarrier sequence number, and l represents the OFDM frame number, and this is the space-time block code structure of an accurate orthogonal design.

3, the emission diversity method of time-domain synchronous orthogonal frequency-division multiplexing system according to claim 1, it is characterized in that, when the link of any one transmitting antenna broke down, the link of its excess-three transmitting antenna was still formed the space-time block code structure of accurate orthogonal design.

4, the emission diversity method of time-domain synchronous orthogonal frequency-division multiplexing system according to claim 1 is characterized in that, when the link that has only two transmitting antennas is working properly, then has:

(1) Tx1 and Tx3, perhaps Tx2 and Tx4, coded OFDM structure when forming the orthogonal space of one 2 antenna respectively;

(2) Tx1 and Tx2, perhaps Tx3 and Tx4, the orthogonal space of forming one 2 antenna respectively is the coded OFDM structure frequently.

Images