Processing

Please wait...

Settings

Settings

1. WO2007000709 - SIGNAL TRANSMITTING/RECEIVING APPARATUS AND METHOD FOR MULTI-INPUT MULTI-OUTPUT WIRELESS COMMUNICATION SYSTEM

Note: Text based on automatic Optical Character Recognition processes. Please use the PDF version for legal matters

SIGNAL TRANSMITTING/RECEIVING APPARATUS AND METHOD FOR MULTI- INPUT MULTI-OUTPUT WIRELESS COMMUNICATION SYSTEM

FIELD OF THE INVENTION
The present invention relates to a signal transmitting/receiving apparatus and a method for wireless communication system, and particularly to a signal transmitting/receiving apparatus and a method for multi-input multi-output wireless communication system.

BACKGROUND OF THE INVENTION

In the wireless communication system, the fading effect of the wireless channel has a significant influence on the receiving quality of the signal, while a diversity technique is an effective way for handling such fading effect. In conventional diversity techniques, multiple antennas are arranged in the receiving apparatus, and multiple duplicates of the same transmitted signal in the receiving apparatus are processed, such as maximal-ratio combination, so as to achieve a diversity gain, whereby the bit error rate performance for received signals may be improved significantly with respect to the single antenna system.

Based on the research result of the information theory, e.g., G. J. Foschini and M. J. Gans, "On limits of wireless communications in a fading environment when using multiple antennas," Wireless Personal Communications, vol. 6, no. 3, pp. 311-335, Mar. 1998, employing multi-antenna in both a transmitting apparatus and a receiving apparatus may expand the system capacity remarkably with respect to employing multi-antenna only in the receiving apparatus, which establishes the theory basis for the research on the transmission diversity technique. Meanwhile, for most of the wireless application environments, transmission diversity is a more practical solution. As for the downlink in mobile communication, employing multi-antenna at the base station is easy, and can improve the receiving quality for multiple users at the same time.
Space-time block coding is an effective transmission diversity technique, which achieves the diversity gain and the coding gain respectively through carrying out the time domain (i.e. antenna) and spatial domain (i.e. character cycle) coding to the signals in the transmitting apparatus at the same time. In 1998, Alamouti proposed a simple space-time block coding solution with excellent performance for the system which has two transmitting antennas (see Siavash M. Alamouti, "A Simple Transmission diversity Technique for Wireless Communication" IEEE Journal on Select Areas in Communications, vol. 16, No. 8, October 1998). Such solution does not need the channel information in the transmitting apparatus, i.e., the impulse response of the wireless channel between the transmitting antenna and the receiving antenna. The wireless channel may be regarded as a filter including (L+l) tags, where the tag coefficient is h - [h(0), h(ϊ), ..., h(L)] ^ and ώe channel information is h = [Zz(O), A(I), ... , h(L)] . Moreover,

Alamouti space-time block coding solution can also achieve the full-rate transmitting and the full transmission diversity gain, and the coding matrix is



where each row of Alamouti coding matrix represents the characters respectively transmitted by two antennas in the same character cycle, and is tagged as "→ antenna"; each column of Alamouti coding matrix represents the characters respectively transmitted by the same antenna in two successive character cycles, and is tagged as "J, time".
Since such coding matrix has orthogonality, and the maximum-likelihood decoding of the receiving apparatus only involves linear processing, it has much lower computation complexity. Owing to these advantages, the Alamouti's solution has been adopted by some 3G standards, such as WCDMA and CDMA2000.
Such space-time block coding solution is based upon the orthogonal matrix, the so-called space-time block coding (Space Time Block Code, STBC) solution, is then employed in the system which has Nt > 2 transmitting antennas subsequently. The so-called space-time block coding is a method of mapping the k information characters ci'c2 >-- -> c* to a PxM coding matrix C , where M is the number of the transmitting antenna, and p is the number of the element cycle occupied by the coding matrix, and each element in the coding matrix is a linear combination of C1, c2,...,ck and their conjugates.

The space-time block coding solution employing the orthogonal coding matrix is an important particular example for space-time block coding, which is characterized in that the coding matrix C satisfies C C = al } where β > 0 , 1 represents an identity matrix.

However, for the complex signals of the transmitting apparatus employing multiple antennas, only if Nt = 2 , a full-rate orthogonal coding matrix (i.e. Alamouti's solution) exists; when Nt > 2 , employing the orthogonal coding matrix will reduce the transmission

rate greatly. For example, if f ~ , the orthogonal coding matrix may be as follows:
S2 S4
S\ "S4 antenna
-S3 S4 -S2
~S4 -S2 °2
GA
\ S2 % S4
S1 S3
-^ S4 4- time
-W3 h
In the case of employing such a coding matrix, the system can only transmit A-character information in 8-character cycles, and therefore the achieved coding rate is only

1/2 (referring to the Formula (38) in V. Tarokh, H. Jafarkhani, and A. R. Calderbank,

"Space-time block codes from orthogonal designs," IEEE Transactions on Information

Theory, vol.45, no.5, pp.1456-1467, Jul.1999 for detailed reasons).
For wireless channel, if the bandwidth for transmitting signal is less than the coherence bandwidth of the channel, such a channel is called flat fading channel; otherwise, it is called frequency selective channel. When the channel is flat fading, there is no interference between elements in the receiving signals, i.e. each receiving signal is only relevant to a certain information character; and when the channel is a frequency selective channel, the receiving signal is a linear superposition of the adjacent information characters, i.e., there is interference between the elements. All of the early researches on space-time block coding and space division multiplexing system assume that the channel is flat fading, i.e., the time-delay spread of the channel is far less than the element cycle of the transmitting signal. However, in a broadband wireless communication system, the channel may present frequency selective fading, and thus it will have more practical significance to research on the transmitting solution and the corresponding receiving method employing multiple antennas in the frequency selective channel.
In order to apply the existing space-time block coding based upon the flat fading channel, the frequency selective channel can be transformed to the flat fading channel by the MTMO (Multi-Input Multi-Output) time domain equalizer in the receiving apparatus. The disadvantage of such a method is the higher computation complexity of the MIMO time domain equalizer. Moreover, because the OFDM (Orthogonal Frequency Division Multiplexing) employs multiple carriers to transmit signals, the carrier offset may result in interference between carriers, and significantly influence the detection of the transmitting signals.
In view of the above, the prior art has the disadvantages of low coding rate, complex computation, and difficulty in detecting. There is an need for a wireless communication method, which has a high coding rate, low computation complexity, and low sensibility to the carrier offset.

OBJECT AND SUMMARY OF THE INVENTION

Regarding the disadvantages in the prior art, the present invention provides a signal transmitting/receiving apparatus and method for wireless communication system, which has higher coding rate, lower computation complexity, and lower sensibility to the carrier offset.
The transmitting apparatus in the embodiment of the present invention comprises a multi-path output selector for dividing an input information sequence into multiple paths so as to obtain multiple shunt signals; multiple transmitting units, corresponding to the multiple shunt signals respectively, wherein each transmitting unit includes a serial-to-parallel converter for serial-to-parallel converting the shunt signals so as to obtain multiple block signals of a predetermined length; and a space-time block coder for space-time block coding the block signals so as to obtain multiple sets of coded block signals having orthogonality in the frequency domain.
The transmitting method in the embodiment of the present invention comprises the steps of dividing an information sequence to be transmitted into multiple paths so as to obtain multiple shunt signals; serial-to-parallel converting each of the multiple shunt signals respectively so as to obtain multiple block signals of a predetermined length; space-time block coding the multiple block signals so as to obtain multiple sets of coded block signals having orthogonality in the frequency domain; adding redundant information into the coded block signals; and transmitting the signal with the added redundant information via the transmitting antenna.
According to one embodiment of the present invention, four transmitting antennas are arranged in the transmitting apparatus, the information sequence to be transmitted is divided into signals of two independent signal paths through the serial-to-parallel conversion, and two transmitting antennas are provided for each path. Space-time block coding is performed to the signal on each path, as the transmitting antenna number is 2. Because the frequency selective channel should be taken into consideration and the frequency domain equalization is employed in the receiving apparatus, space-time block coding is carried out by taking block as a unit. Every data block output by the space-time block coder is transmitted by the corresponding antenna, after the interference-eliminated item (such as cyclic prefix) having a length of the channel order is added. The purpose of adding the interference-eliminated item (such as cyclic prefix) is to discard the receiving signal corresponding to the interference-eliminated item (such as cyclic prefix) in the receiving apparatus, thereby eliminating the interference between the adjacent data blocks induced by the frequency selective channel, and making the channel matrix into a cycle matrix. Two paths of space-time block coded signals are transmitted by four antennas at the same time, and occupy the same frequency band.
The receiving apparatus in the embodiment of the present invention comprises multiple receiving units for receiving multi-path input signals, wherein the multi-path input signals are space-time block coded block signals having orthogonality in the frequency domain; a multi-path signal separator, coupled to the multiple receiving units, for separating the input signals into multiple first separation signals; multiple output units corresponding to the multiple first separation signals respectively, each of the multiple output units comprising a linear combiner for separating the coded data blocks by applying linear combination to the first separation signals, so as to obtain second separation signals; and a frequency domain equalizer for frequency-domain equalizing the second separation signals so as to recover an information sequence.
The receiving method in the embodiment of the present invention comprises the steps of receiving multi-path input signals, wherein the input signals are the space-time block coded block signals; separating the signals transformed to the frequency domain in each path for the first time; separating the coded block signals in each path for the second time by applying linear combination to the first separation signals, and obtaining a transmitting diversity gain; and recovering an information sequence by applying frequency domain equalization to the second separation signals, respectively.
According to one embodiment of the present invention, for the receiving apparatus corresponding to a two-path 4-antenna transmitting apparatus, the transmitting signal comprises two paths of completely independent space-time block coded signals, so that the receiving apparatus needs at least two sets of receiving units including the receiving antenna to separate two paths of signals. The receiving signals, corresponding to the redundant information, on each receiving antenna are discarded, and then transformed to frequency domain by a discrete Fourier transformation. And then the two paths of signals are separated by utilizing the characteristic of space-time block coded signals and the interference suppression algorithms with low computation complexity, and finally detected respectively by the frequency domain equalization with minimum mean square error.
Base upon the above, a layer detection algorithm for sequencing according to the equalized mean square error is disclosed in one embodiment of the present invention, in which one path of signal having minimum mean square error is detected, and its contribution to the received signal will be subtracted from the received signal, and the remaining paths of signals are then detected. According to another embodiment of the present invention, the detecting method after eliminating interference can be performed iteratively to further improve the system performance.
Compared with the prior art, the embodiments of the present invention combine space division multiplexing and space-time block coding together, that is, the information sequence may be divided into multiple paths and transmitted at the same time, and each path may be space-time block coded. The present invention provides a higher transmission rate through the space division multiplexing than that of the solution employing space-time block coding only. The present invention has also achieved antenna diversity and improved the bit error rate performance of the system comparing with the solution employing the space division multiplexing only.
The present invention is applicable to the broadband wireless communication, in which the channel is frequency selective. Since the present invention may employ single carrier transmission, its advantages over the solution employing Orthogonal Frequency Division Multiplexing (OFDM) is the lower peak-to-average ratio of the transmission power, and thus the requirement for the linear range of the power amplifier at the transmitting terminal is lowered. Moreover, the sensibility to the carrier offset is lowered, whereby the requirement for the carrier synchronous device at the receiving terminal is lowered.
The conventional transmission solution only employing the space-time block coding is to perform space-time block coding to one path of signal involving all the transmitting antennas, and the space-time coded matrix obtained takes up more element cycles. In the embodiment of the present invention, the information sequence is divided into multiple paths, and each path carries out space-time block coding to one subset of the transmitting antenna array, and thus the number of the element cycles occupied up by the space-time coded matrix is reduced. Since it is assumed that the channel is constant in one coding matrix for the decoding of space-time block coded signals, the present invention has a low requirement for the time invariance of the channel, and the influence of the time- varying channels to its performance is relatively low.
Other objects and achievements of the present invention will become apparent, and the present invention will also be generally understood, from the descriptions and following claims, in conjunction with the drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic structural view of a transmitting apparatus for wireless communication system according to one embodiment of the present invention;
FIG. 2 is a flow chart of signal processing of the transmitting apparatus in the embodiment of FIG. 1;
FIG. 3 are the formats of the data blocks transmitted on the two antennas in the embodiment of FIG. 1;
FIG. 4 is a schematic structural view of a receiving apparatus for wireless communication system according to one embodiment of the present invention;
FIG. 5 is a flow chart of signal processing of the receiving apparatus in the embodiment of FIG. 4;
FIG. 6 is a flow chart of the method for recovering the space-time block coded signals according to one embodiment of the present invention;
FIG. 7 is a flow chart of the method for recovering the space-time block coded signals according to another embodiment of the present invention;
FIG. 8 is a schematic structural view of a transmitting apparatus for wireless communication system according to another embodiment of the present invention;
FIG. 9 is a schematic structural view of a receiving apparatus for wireless communication system according to another embodiment of the present invention;

FIG. 10 is a schematic structural view of a transmitting apparatus for wireless communication system according to another embodiment of the present invention; and

FIG. 11 is a schematic structural view of a receiving apparatus for wireless communication system according to another embodiment of the present invention.
In all of the above drawings, like numerals are used to indicate the same, similar or corresponding features or functions.

DETAILED DESCRIPTION OF THE INVENTION

The embodiments of the present invention combine space division multiplexing and space-time block coding together, and provide a signal transmitting/receiving apparatus and method for wireless communication system with higher coding rate and lower computation complexity.
Referring to FIGs. 1 and 2, the technical solution of the present invention is further illustrated by a preferred embodiment. As shown in FIG. 1, the transmitting apparatus for wireless communication system comprises a two-path output selector 101, and a first transmitting unit 10 and a second transmitting unit 20 respectively coupled to the two outputs of the two-path output selector 101. The first transmitting unit 10 further comprises a first serial-to-parallel (SfP) converter 102, a first space-time block coder 104, a first transmitting subunit 106 and a second transmitting subunit 108. The information sequence to be transmitted ^ is divided into two paths through the two-path output selector 101, so as to obtain two shunt signals, in which one path of signal corresponding to the first transmitting unit 10 passes through a first serial-to-parallel converter 102 so as to obtain multiple block signals having a predetermined length, and then passes through a first space-time block coder 104 in order to space-time block code the block signals to obtain the coded block signals having orthogonality in the frequency domain, sequentially. The other path of signal corresponding to the second transmitting unit 20 is operated in a similar way.
Referring to FIG. 1 and step S201 of FIG. 2, in the two-path output selector 101, the information sequence ^ is divided into two paths of shunt signals, and then be serial-to-parallel converted through the first serial-to-parallel (S/P) converter 102 to constitute block signals having a length of N: s(k) = [f(AiV),- • •, t(W + N-Ϊ)f
where s^ > are the signals in each path; k is the sequence number of the serial-to-parallel converted data block; and T is the vector or the transpose of the matrix. The following s^' denotes data block k having a length of Ν, and t(n) denotes the nth information character.
Referring to FIG. 1 and step S201 of FIG. 2, two successive blocks s(2k^ and

1^ ' stand for the input of the first space-time block coder 104, and then are processed with the space-time block coding so as to obtain the coded block signals having orthogonality in the frequency domain. According to one embodiment of the present invention, the specific coding method of the space-time block coder is marking the block k proposed to be transmitted by the transmitting antenna i (i=l,2) in the output of the space-time block coder 104 as s^ , and the coding rule is expressed as follows:
s ™ (n) = s(2kN + n) s2{lk) (n) = j((2it + I)N + n)

0,l,...,Ν-l (1)

where ^N indicates to calculate the modulus of N, and ™ indicates to conjugate.

The two equations in the first row of the Formula (1) indicate the coding method on two antennas in the first block cycle; the two equations in the second row indicate the coding method on two antennas in the second block cycle. Every coding process includes two block cycles having a length of N, which are mainly used to constitute the orthogonal coding matrix to provide transmission diversity by utilizing two transmitting antennas.

The coding rule for Formula (1) can be described as: each coding process comprises two block cycles having a length of N. The first antenna 1 transmits s^'ζ) j and the second antenna 2 transmits s^ + <* in the first block cycle; and the data block transmitted by the first antenna 1 is the result obtained by order-reversing, transferring, conjugating and negating s(2lc + 1^ , and similarly, the data block transmitted by the second antenna 2 is the result obtained by order-reversing, transferring, and conjugating ^(2/c) in the second block cycle. In this coding rule, the data blocks transmitted in one coding cycle by the two antennas have the following relation in the frequency domain:
c(2fc+l) _ e(2*0 • v(2k+l) _ «(2fc)
3I - 32 ' Λ2 ~ Λl (2) where the time domain data block is converted into the frequency domain through a normalized discrete Fourier transform matrix F by the equation < < . The normalized discrete Fourier transform matrix F is an N x N matrix, the element (k,n) is
0,l,...,N-l

5 where j is imaginary unit. As such, the coded signals have similar orthogonality in the frequency domain as

Alamouti scheme, and the difference therebetween resides in that coding is performed with character as unit in the Alamouti scheme and orthogonality is reflected in the time domain; and the coding of the present invention is performed with block as unit and orthogonality is reflected in the frequency domain.
The first transmitting subunit 106 comprises a first redundant information adder 110

(e.g., a cyclic prefix (CP) adder, referred to a cyclic prefix adder hereafter) and a first parallel-to-serial (P/S) converter 114 coupled to the first antenna 1. Likewise, the second transmitting subunit 108 comprises a second cyclic prefix adder 112 and a second parallel-to-serial (P/S) converter 116 coupled to the second antenna 2. The output S1 of the first space-time block coder 104 sequentially passes through the first cyclic prefix adder 110 and the first parallel-to-serial converter 114, and then is transmitted through the first antenna 1; while the output S2 sequentially passes through the second cyclic prefix adder 112 and the second parallel-to-serial converter 116, and then is transmitted through the second antenna 2.
Referring to FIG. 1 and step S203 of FIG. 2, each data block having a length of Ν output by the~ space-time block coder 104 is added with the redundant information (cyclic prefix (CP)) having a length of L in the first cyclic prefix (CP) adder 110, wherein channel orders between the transmitting/receiving antennas are L, i.e., the tap number of the channel is L+l; in step S204, the obtained data blocks are then transmitted by corresponding antenna after parallel-to-serial converting (P/S) by the first parallel-to-serial converter 114.
According to an embodiment of the present invention, adding CP having a length of L is to copy the last L characters of the data block to the front of the data block, e.g., the data block s(k) = [t(kN),- --,^(ZcN + N -l)f having a length of Ν will become a data block having a length of Ν + L after adding with a cyclic prefix having a length of L: u(k) = [t(kN4N-L)r , t(kN + N - 1) , t(kN),- • •, t(kN + N- l)]τ
cyclic prefix (CP)
The purpose of adding cyclic prefix is to discard the receiving signals corresponding to the cyclic prefix in the receiving apparatus, thereby eliminating the interference between the adjacent data blocks caused by the frequency selective channel, and enabling the channel matrix to be a cyclic matrix.
The cyclic prefix also can be substituted by other ways, such as adding zero data having a length of L, i.e., adding L zeros to the end of the data block. Since adding cyclic prefix is directed to eliminating the interference between the adjacent data blocks caused by the frequency selective channel, it does not matter what the added value is.
As shown in FIG. 1, the constitution of the second transmitting unit 20 is the same as that of the first transmitting unit 10. The other path of signal of the two-path output selector 101 corresponding to the second transmitting unit 20 sequentially passes through the second serial-to-parallel converter 103 and the second space-time block coder 105. The output after space-time block coding by the second space-time block coder 105 goes to the third transmitting subunit 107 and the fourth transmitting subunit 109, respectively.

For example, the output S3 of the second space-time block coder 105 sequentially passes through the third cyclic prefix adder 111 and the third parallel-to-serial converter 115, and then is transmitted by the third antenna 3; while the output S4 sequentially passes through the fourth cyclic prefix adder 113 and the fourth parallel-to-serial converter 117, and then is transmitted by the fourth antenna 4. The operational principle of the second transmitting unit 20 is the same as that of the first transmittήig unit 10 and the details are not redundantly stated.
FIG. 2 is the specific processes of the transmitting terminal in the embodiment of FIG. 1, the steps will be described in detail as follows:
In step S201, the information sequence W is divided into two paths of signals, and then each path of signal after serial-to-parallel converting by the serial-to-parallel converter constitutes a block signal having a length of N:
s(k) = IXA-V),- • •, t(kN + N- l)f .

In step S202, two successive blocks s(2k^ and s(2k + 1^ stand for the input of the first space-time block coder 104. The two paths of signals employ the same space-time block codes, so as to obtain the coded block signals having orthogonality in the frequency domain. The coding method can be referred to in the above descriptions of the coding rule

Formula (1);
In step 203, each data block having a length of N output by the space-time block coder , i.e., the coded block signals, needs to be added with the cyclic prefixes (such as the redundant information) having a length of L, wherein channel orders between the transmitting/receiving antennas are L, i.e., the tap number of the channel is L+l;
In step S204, the signals added with the redundant information are transmitted by corresponding antennas after parallel-to-serial converting.
FIG. 3 shows the formats of the data blocks transmitted on the first antenna 1 and the second antenna 2 in the above embodiment. The other path of signal is transmitted by the third antenna 3 and the fourth antenna 4, and the data blocks transmitted by them are of the same format.
According to the above embodiment of the present invention, the information sequence to be transmitted is firstly divided into two independent paths of signals to achieve the spatial multiplex coding of the information sequence. Then, each path of signal is provided with two transmitting antennas and performs two sets of independent space-time block coding, and the spectral efficiency and coding rate is improved by the frequency domain multiplexing. For example, each path of signal is serial-to-parallel converted so as to divide a path of information sequence into two independent paths of subsequences, and each subsequence after modulating is respectively transmitted by one antenna at the same time, such that the information of two characters can be transmitted only in one character cycle.
At the receiving terminal, a sequencing interference cancellation algorithm is employed to detect each path of signal, respectively. The idea of sequencing is to detect the path of signal which has the maximum signal-to-noise ratio and take the detected result as an actual value, and then subtract the contribution of this path of signal to the receiving signals from the receiving signals, and detect the remaining paths of signals sequentially in this way. The receiving antenna number desired in such a method at least equals to that of the paths of the transmitting antenna. According to the embodiment of the present invention, because each path of the signal at the transmitting terminal is the same, the bit error rate performance of the system will not be deteriorated, even when the channel presents the correlation.

Since the transmitting apparatus comprises two paths of totally independent space-time block coded signals, the receiving apparatus needs at least two receiving antennas to separate the two paths of signals and then detect thereto. The receiving apparatus may also have several receiving antennas. Along with the increase of the antenna number, multiple duplicates can be processed, such as maximal-ratio combined, so as to achieve a diversity gain, and improve the bit error rate. But for the embodiments of the present invention, the desired receiving antenna number at least equals to that of the paths of the transmitting unit, so as to complete decoding.
FIG. 4 is a schematic structural view of a receiving apparatus for a wireless communication system according to one embodiment of the present invention, in which the receiving apparatus comprises a first receiving unit 402 having a first receiving antenna 5 and a second receiving unit 403 having a second receiving antenna 6 for receiving input signals, in which the input signals are the space-time block coded block signals and have orthogonality in the frequency domain; a two-path signal separator 401 coupled to the first receiving unit 402 and the second receiving unit 403 for separating the input signals received by the first receiving unit 402 and the second receiving unit 403 into the first separation signal respectively; a first output unit 404 and a second output unit 405, wherein the first output unit 404 and the second output unit 405 correspond to the first two separation signals respectively, and respectively comprises a linear combiner 410 (411) for separating the coded data blocks by linear combining to the first separation signals, so as to obtain the second separation signal; and a frequency domain equalizer 412 (413) for frequency-domain equalizing the second separation signals so as to recover an information sequence.
For each data block having a length of N+L transmitted by the transmitting apparatus and added with a cyclic prefix, each receiving antenna receives a signal with a length of

N+L.
Referring to FIG. 4 and step S501 of FIG. 5, in taking the first receiver unit 402 as an example, for the received input signals that are space-time block coded block signals and have orthogonality in the frequency domain, the L receiving signals (redundant information) corresponding to the cyclic prefix are discarded in the cyclic prefix discarder

406, because this part of receiving signals includes the interference between two successive data blocks, whereby two data blocks having a length of N may be obtained on each antenna in one space-time block coding cycle. The two data blocks on the receiving antenna m (m=l,2) are expressed as:
yml
v


Where, mk is the zero mean; the variance is the complex white Gaussian noise vector forσ™ ; and the -'* x •** channel matrix Hnrø is a cyclic matrix constituted by the channel coefficient from the transmitting antenna n (n=l,2,3,4) to the receiving antenna m,

wherein the first row of this cyclic matrix is
; i.e., the channel coefficient vector added with N — L - I zero elements.
The data block having a length of (N+L) transmitted by the transmitting antenna n is:
u(k) = [t(kN + N - L),- - -.tCkN + N -l) , t(kN),- • •, t(kN + N - l)]τ
cyclic prefix (CP)
The corresponding receiving data block having a length of N on the receiving antenna m (with the L receiving signals corresponding to the cyclic prefix discarded) can be expressed as:


where the channel matrix H is N x N , and ΕLnm has the following important feature:

FHnwF* = ΛMm (4)

where () H denotes the conjugate transpose of the matrix, A nm is a diagonal matrix, and the diagonal element is the N-point discrete Fourier transform of the channel

coefficient vector, i.e., Anm (k, k) = ∑ hnm (/) exp 3 * , k = 0,1, • • • , N - 1.
/=o K N J
The first input unit 402 further comprises a first discrete Fourier transformer 408 (DFT). Referring to FIG. 4 and step S502 of FIG. 5, the receiving signal is discrete Fourier transformed so as to obtain the frequency domain receiving signal Ymk - FjmA:, m,k = 1,2 . Through Formulas (3) and (4), it may be:
7mi = MmSi(2i) + A2mS2(2i) + A3JS3M + JS4(2I) + Wmi (5)
Ym2 = AiJS1V+" + A2mS2(2i+1> + A3mS3(2i+1) +
+ Wm2
According to Formula (2), the relation between the input and output in the frequency domain of the system can be expressed as:

F :=





":=" refers to define a symbol for the complex quantity behind ":=", so that the quotation of this quantity may be simplified. The J and 2 are diagonal matrixes, and

the matrix like A = 1 is called as a block Alamouti matrix, because it has
LΛ2 ~ Λ1
similarity with the Alamouti coding matrix, and here the scalar in the Alamouti matrix is just substituted by the diagonal matrix. The Alamouti matrix has the following form:
"A1 A2 I
A =
A2 - Λ]_

where, «" ' ' is a diagonal matrix; ' indicates the diagonal matrix obtained after conjugating the diagonal element of A1.. Now, the block Alamouti matrix will be proved to have the following properties: the sum and product of any two block Alamouti matrixes are still a block Alamouti matrix, and the inverse matrix of any invertible block Alamouti matrix is still a block Alamouti matrix.
A1 + A1 A2 + A2
A + B =
A1 A, Λ2 + A2 -A1 -A1
I) B = is a block Alamouti matrix, then and

>2 ~ Δ1 J
obviously A + B is a block Alamouti matrix. Similarly, A-B can be proved to be a block Alamouti matrix.
A1 A1 + A2A2 A1A2 — A2A1
AB =
-A1A2 + A2A1 A1 A1 + A2A2
2) , and thus AB [s a block Alamouti matrix.

3) A is invertible, and its inverse matrix is

where ® denotes the Kronecker product of the matrix, and thus A~ 1S a block Alamouti matrix.
Therefore, Alamouti matrix is provided with the following properties:
a) The block Alamouti matrix is closed for the computations of addition, subtraction and inversion, i.e., the sum and product of any two block Alamouti matrixes are still a block Alamouti matrix, and the inverse matrix of any invertible block Alamouti matrix is still a block Alamouti matrix.
b) Λl and Λ2 are two diagonal matrix elements of the block Alamouti matrix, then

AA^ = A^A = I, ® I IA11" + IA
, where I2 represents a 2 x 2 identity matrix; ^ denotes

the Kronecker product; z is a real diagonal matrix obtained after calculating the modulus of each diagonal element. The inverse matrix of A can be expressed as:



Therefore, it requires less computation to obtain A l .
Obviously, the equivalent channel matrix A defined in Formula (6) includes four block Alamouti matrixes, and it is written as follows:

A = (7)
_A21 A22 _

where A^Z'7 1>2) are all 2N x 2^ block Alamouti matrixes.

Formula (6) establishes a relation between the space-time block coded transmitting signal and the corresponding received signal in the frequency domain, i.e., the equation of the input and output of the system in the frequency domain, which initiates the deducing of the detection algorithm at the receiving terminal. As defined in Formula (6), the symbol Y represents the received signal; Λ 1S the channel matrix; ^represents the additive noise; and s is the transmitted signal to be estimated. Formula (7) is an equivalent form of the channel matrix A defined in Formula (6) to facilitate the later deduction of the detection algorithm.
The second input unit 403 which is the same as the first input unit 402 further comprises a second cyclic prefix (CP) discarder 407 and a second Fourier transformer 409.

In a space-time block coding cycle, the information sequence received by the first receiving antenna 5 passes through the first cyclic prefix discard 406 such that the cyclic prefix is discarded, and then passes through a first discrete Fourier transformer 408 so as to obtain the frequency domain signals; and finally are input into the two-path signal separator 401. Likewise, the information sequence received by the second receiving antenna 6 passes through the second cyclic prefix discarder 407 such that the cyclic prefix is discarded, and then passes through the second discrete Fourier transformer 409 so as to obtain the frequency domain signals; and finally is input into the two-path signal separator 401.
Referring to FIG. 4 and step S503 of FIG. 5, this step is intended for interference suppression. The purpose of interference suppression is to separate the mixed two paths of signals which have been space-time block coded, i.e., to eliminate the interference of one path of space-time block coded signal to the other path of space-time block coded signal, such that each path of signal having been separated may be detected respectively to recover the information sequence. Based upon the system equation (6), the interference suppression matrix is constituted.



Multiplying C by the vector of the received signal can obtain


where p ~ Aj l Ai2A22A2i ; Q - A22 A2iAπ Ai2 . Ai2A225 A2iAπ 5 both of P and Q are the block Alamouti matrixes. Formula (9) indicates that in the signals obtained after multiplying C by the vector of the received signal, λ is only relevant to the first path of signals l s 2 , and 2 Only includes the contributions of the second path of signals 3' 4 , so that the interference between the two paths of space-time block coded signals is eliminated, thereby separating the two paths of signals.
The first output unit 404 includes a first linear combiner 410. Referring to FIG. 4 and step S504 of FIG. 5, step S504 is intended for linear combination. Through the process in step S503, the signals l J 2 transmitted by the first transmitting antenna 1 and the second transmitting antenna 2 and the signals 3 ' 4 transmitted by the third transmitting antenna 3 and the fourth transmitting antenna 4 can be detected separately. Specifically,



where, the noise component l and 2 includes the first 2N and the last 2N elements of C^ in Formula (9). It is known from the Formulas (10) and (11), same detection method may be applied to the two paths of signals. Now the detection of the first path of signal will be taken as an example for illustration.
Since the equivalent channel matrix P in Formula (10) is a block Alamouti matrix, the

method of linear combination may be applied to separate * and 2 , and the transmission diversity gain provided by space-time block coding may be obtained.
Multiplying both sides of the Formula (10) by P at the same time can obtain



where the noise component l . Since P is a block Alamouti matrix, -^ "

H "
satisfies ~~ 2 J , and ! is a -^ x -^ real diagonal matrix, 1^1 and l15'7* may be detected separately.

The first input unit 404 further comprises a first minimum mean square error (MMSE) frequency domain equalizer 412, a first determinater 414 and a second determinater 416. Referring to FIG. 4 and step S505 of FIG. 5, step S505 is intended for the minimum mean square error frequency domain equalization. Taking the detection of * as an example, extracting the portion relevant to the detection of l in Formula (12) can obtain
Z1 = P1S1 + V1 = P1Fs1 + V1 (13)
7 V 7 V
where l and J are the first N elements of and . The diagonal elements of the p
diagonal matrix 1 in Formula (13) are all the quadratic sum of the modulus of two numerals, so that the diversity achieved by the system is 2. Based upon Formula (13), ! may be recovered by the linear equalizer, which mainly comprises a zero-forcing (ZF) equalizer and a minimum mean square error (MMSE) equalizer. Because the zero-forcing equalizer does not take the noise into account, noise amplification may be possible. Therefore, the MMSE equalizer having a better performance is employed in this

embodiment, i.e., the matrix R is designed in order to minimize
,

E(-)
where ^ ' represents the mathematical expectation. Due to the orthogonality rule, this equals to design R to satisfy:



Due to D " =~ AΆ1U0AΆT221 5 D is a block Alamouti matrix, and it satisfies

DD -D D - 12 (X) D1 After computation, mean value of the noise component V1 in
( V1V1H \ \ - σw 2 (ϊN +O1)P1 . If the

autocorrelation matrix of the information sequence J is EyS1S1 ) = σsϊN , and the

information sequence is irrelevant to the noise, s2(2k) (n) - s((2k + Ϊ)N + ri) may be

obtained through Formula (14), and the minimum mean square error estimation to ! may be obtained by determining the output of the following MMSE equalizer.


Since both Au and Az2 are block Alamouti matrixes in the interference cancellation matrix (Formula 8), its inversion computation has lower computation complexity, and the product of the Alamouti matrix can be analyzed into the product of the diagonal matrix.

Therefore, both the matrix inversion and matrix multiplication involved in the computation for the interference cancellation matrix C have lower computation complexity. The matrix

P1 +^(IN +D1)
σs needs to be inversed in frequency domain MMSE equalization (Formula

15) is a diagonal matrix, and when the block length N is the power of 2, the computation of multiplying F may be completed by the Inverse Fast Fourier Transform (IFFT), which reduces the computation complexity.
In view of the above analysis, for the two-path multiplexing single-carrier space-time coding system in the present invention, both interference cancellation and frequency domain equalization of the receiving apparatus are achieved with lower complexity.
One output from the two-path signal separator 401 sequentially passes through the first linear combiner 410 and the first minimum mean square error (MMSE) frequency domain equalizer 412, so as to obtain two outputs, which pass through the first determinater 414 and the second determinater 416 respectively, and finally output from the first determinater 414 and the second determinater 416. Likewise, since the second output unit 405 is the same as the first output unit 404, the other output from the two-path signal separator 401 sequentially passes through the second linear combiner 411 and the second minimum mean square error (MMSE) frequency domain equalizer 413 so as to obtain two outputs which pass through the third determinater 415 and the fourth determinater 417 respectively, and finally output from the third determinater 415 and the fourth determinater 417.
FIG.5 is a flow chart of signal processing of the receiving apparatus, and the steps will be described in detail as follows:
In step S501, firstly, for the received input signals, the space-time block coded block signals, in the cyclic prefix, the L receiving signals (redundant information) corresponding to the cyclic prefix are discarded, because such part of the received signals includes the interference between two successive data blocks.
In step S502, the blocks in each path of signal are transformed from the time domain to the frequency domain by the discrete Fourier transformation.

In step S503, interference suppression is performed to separate the mixed two paths of signals which have been space-time block coded, i.e., to eliminate the interference of one path of space-time block coded signal to the other path of space-time block coded signals, such that each path of signal having been separated may be detected respectively to recover the information sequence.
In step S504, the linear combination is performed, so that the signals l 5 2 transmitted by the first transmitting antenna 1 and the second transmitting antenna 2 and the signals 3' 4 transmitted by the third transmitting antenna 3 and the fourth transmitting antenna 4 can be detected separately to achieve the second separation of the coded blocks in each path of signal, thereby obtaining the transmission diversity gain.

In step S505, the minimum mean square error (MMSE) frequency domain equalization is performed to recover the signals.
The method of signal detection in steps S501-S505 may be referred to in the above descriptions of the embodiment in FIG. 4 for detail. In view of the above analysis, for the two-path multiplexing single-carrier space-time coding system in the present invention, both interference cancellation and frequency domain equalization of the receiving apparatus are achieved with lower complexity.
In the above mentioned methods, two paths of the spatial multiplexing signals which have been space-time block coded are detected by completely the same methods, such as that two paths of signals may be detected at the same time by parallel processing. But this method only utilizes transmission diversity, rather than the receiving diversity provided by the two receiving antennas. In order to utilize the receiving diversity to improve the detection performance of the receiver, a layer and iterative algorithm is proposed based upon the algorithms of the joint interference cancellation and the frequency domain equalization according to one embodiment of the present invention. In the layer detection algorithm, a path of signal having minimum mean square error after the MMSE frequency domain equalization will be detected firstly, and the resulting estimated value is taken as an actual value, such that the contribution to the received signal will be subtracted from the received signal. If the estimated value of the path of signal detected firstly is right, then its contribution to the received signal will be eliminated totally, and at this time the system equals to the conventional 2-transmitting antennas 2-receiving antenna single-carrier space-time block coding system, and therefore, the detection of the remaining path of signal may use the receiving diversity provided by the two receiving antennas.
Because the detection quality of the first detected path of signal determines the following quality of eliminating the interference for such path of signal in the layer detection, the path of signal with a high detection reliability will be detected first. Based upon the MMSE frequency domain equalization, a method for determining the detection sequence according to the equalized mean square error (MSE) is deduced. Taking the

detection of * as an example, the mean square error autocorrelation matrix corresponding to the output of MMSE equalizer in Formula (15) is as follows:



The matrix G defined in Formula (16) is a real diagonal matrix. The diagonal element of e corresponds to the mean square error of each character detected through MMSE in

the data block 1 . Due to the property of the discrete Fourier matrix, the diagonal elements of e are equal and equal to

N (17)

Where ' ' represents the trace of the matrix, and thus l corresponds to the mean square error of the first path of signal detected through MMSE. The mean square error of the other path of signal detected through MMSE may be calculated in the same way.
According to one embodiment of the present invention, as shown in FIG. 6, the method for recovering the space-time block coded signals comprises the following steps:

In step S601, comparing the mean square errors of the two paths of signals detected through MMSE;
In step S602, firstly detecting a path of signal with smaller mean square error;

In step S 603, subtracting the contribution of this path of signal to the received signal from the received signal so as to utilize the receiving diversity;
In step S604, detecting the other path of signal.
Before detection, the mean square errors are firstly compared. The detection begins at the path of signal with smaller mean square error (more generally, the first path of signals svs2y The algorithms of joint interference cancellation and the minimum mean square error equalization may be employed so as to obtain the estimations lJ 2 for 15 2 , and the estimated results will be transformed to the frequency domain through DFT. The contribution of the first path of signal to the received signal is subtracted from the received signal, as shown in the following Formula



Assuming the functions of l' 2 have been eliminated for detection of the remaining path of signals si's* } the system is simplified to be a conventional 2-transmittmg antenna and 2-receiving antenna single-carrier space-time block encoding system which does not employ spatial multiplexing. Therefore, the detection thereof does not only achieve the transmission diversity provided by space-time block encoding, and also obtain the receiving diversity provided by the two receiving antennas.
In this layer detection algorithm, the path of signal firstly detected does not utilize the receiving diversity. Instead, the path of signal later detected utilizes the receiving diversity, and thus the path of signal later detected will have a better bit error rate performance.

Employing the same method of interference elimination, the function of the path of signal later detected can also be eliminated from the received signal, and then the path of signal firstly detected will be re-estimated, such that the detection of both two paths of signals utilizes the receiving diversity. This method of interference elimination and re-estimation may be performed iteratively to further improve the detection quality of these two paths of signals, until the overall performance is optimized. For example, according to one embodiment of the present invention in FIG. 7, the method for recovering the space-time block coded signals comprises the following steps:
In step S701, comparing the mean square errors of the two paths of signals detected through MMSE;

In step S 702, firstly detecting a path of signal with smaller mean square error;
In step S703, subtracting the contribution of this path of signal to the received signal from the received signal so as to utilize the receiving diversity;
In step S704, detecting the other path of signal;
In step S705, subtracting the contribution of this path of signal to the received signals detected in step S704 from the received signal; and
In step S706, repeating steps S704 to S705 until the overall performance is optimized. According to one embodiment of the present invention, the steps after the above S706 will not be iterative, and the result will be output directly.
According to one embodiment of the present invention, the method for recovering the space-time block coded signals can be described as:
firstly detecting a path of signal with smaller mean square error; and then subtracting the contribution of this path of signal to the received signal from the received signal so as eliminate the interference; detecting the remaining path of signal by the minimum mean square error frequency domain equalization method for one-path space-time block coding transmission so as to obtain the receiving diversity provided by two receiving antennas and then subtracting the contribution of this path of signal to the received signal from the received signal to eliminate the interference; re-estimating the firstly detected path of signal through the minimum mean square error frequency domain equalization method for one-path space-time block coding transmission, so that the detection of the former path of signal also achieves the receiving diversity provided by two receiving antennas. The method of interference elimination and re-estimation may be performed in an iterative way in order to improve the detection quality of the two paths of signals, until the overall performance is optimized.
As described above, a preferred embodiment of the present invention is illustrated in detail through a 4-transmitting antenna and 2-receiving antenna system. It will be understood that the application of the present invention is not limited to the abovementioned system, but includes other wireless communication (with different numbers of antenna) systems.
FIG. 8 shows one embodiment of a transmitting system with 8 transmitting antennas.

Such transmitting system comprises a 4-path output selector and four transmitting units. Elements included by each transmitting unit are the same as those included by the transmitting unit of the embodiment of FIG. 1. In this embodiment, the inputs of the space-time block coder included in each transmitting unit are two successive blocks

1^ ' and 1^ c ' . As such, the coded blocks may be transmitted in the same way as that of the embodiment in FIG. 1. Since 4 paths of signals are employed for the independent coding, the coding rate is improved.
FIG. 9 is a schematic structural view of a receiving apparatus according to another embodiment of the present invention. Such a receiving apparatus comprises 4 receiving units and 4 output units. Elements of each receiving unit and output unit are the same as those of the receiving unit and output unit of the embodiment of FIG. 3. In the receiving apparatus, each antenna may receive two data blocks with a length of N in a space-time block coding cycle. Along with the increase of the number of the antenna, the channel vector between the integral transmitting and receiving system can be indicated as a vector matrix with the transmitting antenna number multiplying the receiving antenna number.

In the receiving apparatus as shown in FIG. 9, the received signals can also be processed using the method in FIG. 5. The processing method depicted in FIG. 5 completely corresponds to the process completed in the system in FIG. 9, and are not redundantly stated.
FIG. 10 and FIG. 11 illustrate another embodiment of the present invention. The transmitting apparatus of this embodiment has 8 antennas, and is divided into two paths. Each path comprises 4 transmitting antennas, and the receiving apparatus has 2 antennas. Each path of the transmitting apparatus in the embodiment of the present invention employs the space-time block coding with 4 transmitting antennas. Such a coding mode may achieve a better coding effect, and the bit error rate after the decoding is decreased greatly.
In view of the above, the present invention is provided with higher coding rate and time-varying channels which are more applicable to fast changes, as well as less computation complexity, compared with the prior art.
For 4-transmitting antenna system according to one embodiment of the present invention, the coding rate achieved by the conventional orthogonal design without employing spatial multiplexing is only 1/2, while the coding rate obtained by the combination of spatial multiplexing and orthogonal space-time block coding is 2, thus it is suitable for the application with a high requirement for the spectral efficiency. The conventional orthogonal design without employing spatial multiplexing requires the channel to keep constant within 8 successive blocks. However, the present invention only requires the channel to keep constant within 2 successive blocks, thus it is more suitable for the applications in the time-varying channel. The receiving method of the present invention utilizes the orthogonality of the space-time block coded signals, such that both the interference suppression method for separating two paths of signals and the frequency domain equalization method for detecting the signals have lower computation complexity, thereby facilitating practical application.
The present invention also provides the layer detection algorithm and the iterative detection algorithm for improving the detection performance, which reduce the bit error rate with a lower computation complexity. The computer stimulation result demonstrates that the receiving method of the present invention has robustness for the channel estimation error, and under most circumstances, only one iterative step is needed to obtain the optimum system performance.
While the technical contents and features of the present invention have been described above, it will be understood that substitutions and modifications of the present invention may be made by the persons skilled in the art from the teachings and disclosure of the present invention. Therefore, the scope of the present invention should involve various substitutions and modifications without departing from the spirit and scope of the present invention and should be contemplated by the following claims, rather than being limited to the content disclosed by the embodiments.