CN101163121B

CN101163121B - Communication system and character code selecting method thereof

Info

Publication number: CN101163121B
Application number: CN2006101317989A
Authority: CN
Inventors: 何从廉; 李大嵩
Original assignee: Industrial Technology Research Institute ITRI
Current assignee: Industrial Technology Research Institute ITRI
Priority date: 2006-10-12
Filing date: 2006-10-12
Publication date: 2011-06-01
Anticipated expiration: 2026-10-12
Also published as: CN101163121A

Abstract

The invention provides a code selection method. The code selection method can be applied to a multiple-input and multiple-output communication system. The method consists of the following steps. First, a transmitter provides plural code form. Then a receiver receives the code form, calculates or looks up tables to obtain corresponding bit error rate for each code form according to a coding mode. Second, the receiver selects a group code format with the smallest bit error rate, and returns to the transmitter. Finally, the transmitter decides the group code used for transmitting data according to the code form with the smallest bit error rate.

Description

Communication system and character code selection method thereof

Technical Field

The present invention relates to a mimo communication system and a space-time (ST) code selection method thereof, and more particularly, to a group space-time (ST) code selection method and related system using a group space-time block code (G-STBC) coding scheme design and under limited feedback information.

Background

Since wireless communication systems require more and more data traffic, it is an important objective to improve the quality and performance of wireless communication by efficiently encoding, modulating, and processing signals. In order to effectively improve the data transmission rate and the communication link quality, more and more wireless communication systems adopt a multiple-input multiple-output (MIMO) multiple-antenna system design, i.e., multiple antennas are used at the transmitting end and the receiving end, so as to achieve a strong service quality requirement of the next generation wireless communication system. For encoding of multiple antenna systems, space-time codes (STC) are generally used. The coding is a transmission design spanning time and space, and utilizes the relation between transmission antennas and transmission time to achieve maximum diversity (full diversity), and even provide coding gain (coding gain). The main operation modes of space-time codes can be divided into Spatial Diversity (SD) and Spatial Multiplexing (SM). High link quality can be achieved by transmitting redundant (redundant) signals over different transmit antennas in a spatially diverse manner, such as space-time block code (STBC) and space-time trellis code (STTC). The high spectral efficiency can be achieved by simultaneously transmitting different data on different transmit antennas via a space-multiplexing scheme, such as multi-layered space-time code (LSTC) or the well-known Bell Labs layered ST (BLAST) technique.

Fig. 1 is a schematic diagram of a conventional space-time block coding system, and it is apparent that two symbols transmitted span space and time (space-time code). Fig. 2 shows another conventional coding technique for combining Q-2 Alamouti's space-time codes at the transmitting end, which is called double space-time transmit diversity (DSTTD) technique. The space-time character code is

X : = [\frac{\begin{matrix} S_{1,1} & - S_{1,2}^{*} \\ S_{1,2} & S_{1,1}^{*} \end{matrix}}{\begin{matrix} S_{2,1} & S_{2,2}^{*} \\ S_{2,2} & S_{2,1}^{*} \end{matrix}}]

The coding rate in this case is R ═ 2, and the diversity gain obtained is 2. However, although the encoding method of this encoding technique is simpler, the decoding method is more complicated than the STBC, and the codeword structure suffers from many limitations, so that the encoding design is less flexible and therefore the performance is poor.

However, although the spatial diversity can improve the quality of the communication link, the spectrum utilization efficiency is low; in contrast, although spatial multiplexing can improve data transmission rate, it has poor resistance to channel fading. Therefore, in order to obtain the maximum benefit, an optimal profit and loss point between spatial diversity and spatial multitasking must be obtained.

In addition, the receiver must be simpler and better to compute to achieve fast decoding, and the receiver design can also be simplified. The receiver must also have space-time code decoding capability relative to the space-time code at the transmitter, and interference cancellation and signal detection are required. One commonly used optimal decoding method for a receiver uses a joint Maximum Likelihood (ML) detection method, which uses a probability statistics method to decode the received signal. However, although the joint maximum likelihood estimation method has better performance, the required computational complexity is higher. Another decoding method adopts ordered progressive interference cancellation (OSIC) detection, which uses ordered and previous iteration results to perform interference cancellation and signal detection on received signals one by one. For example, assuming that the receiver receives signals from mobile subscriber 1 and mobile subscriber 2, the OSIC detection method can first perform signal detection for mobile subscriber 1 and then deduct (eliminate) the interference of mobile subscriber 1 to obtain the signal of mobile subscriber 2. The OSIC detection method is a less computationally efficient but more efficient technique than the joint ML method, however, it does not further utilize the orthogonal property of space-time codes to simplify the signal detector.

In addition, the wireless communication environment is a time-varying response, and the transmitting and receiving end need a suitable adaptive transmission mechanism to handle signal distortion and achieve the best quality of service. Therefore, a good adaptive transmission scheme adjustment is also needed for MIMO systems.

Disclosure of Invention

The invention provides a MIMO communication system and a character code selection method thereof, which can utilize the algebraic characteristics of orthogonal space-time block codes (O-STBC) to carry out G-STBC coding design and select an optimal character code structure for transmission based on the condition of minimum Bit Error Rate (BER) under limited return information.

The invention provides a character code (codeword) selection method. The group character selection method is suitable for a communication system with multiple transmitting antennas and multiple receiving antennas. The method comprises the following steps. First, the transmitter provides a complex digital code form. Then, the receiver receives the word formats and calculates or looks up the table according to a decoding method to obtain a corresponding Bit Error Rate (BER) of each word. Next, the receiver selects a codeword with the minimum bit error rate and transmits the codeword back to the transmitter. Finally, the transmitter determines a space-time codeword for data transmission based on the codeword format with the minimum bit error rate.

The invention also provides a MIMO communication system. The transmitter first provides a complex digital code form. The receiver receives the codeword formats and calculates or looks up the table to obtain a corresponding bit error rate for each codeword format according to a decoding method. The receiver selects a code form with the minimum bit error rate and transmits the code form back to the transmitter, and the transmitter determines a space-time code for data transmission according to the code form with the minimum bit error rate.

In order to make the aforementioned and other objects, features, and advantages of the present invention comprehensible, preferred embodiments accompanied with figures are described in detail below.

Drawings

FIG. 1 shows a schematic diagram of a prior art space-time block coding system.

FIG. 2 shows a schematic diagram of another prior art space-time block coding system.

Fig. 3 is a diagram illustrating a communication system according to an embodiment of the invention.

FIG. 4 is a flow chart of word selection according to an embodiment of the present invention.

Fig. 5 and fig. 6 are schematic diagrams illustrating a flow of the codeword selection method according to fig. 4 at a transmitting end and a receiving end, respectively.

FIG. 7 is a schematic diagram of a Matched Filter Channel Matrix (MFCM) in accordance with an embodiment of the present invention.

Fig. 8a-8e are schematic diagrams illustrating 5 possible grouped space-time codewords with a total transmit antenna number of 10 according to an embodiment of the invention.

Description of the figures

300-a communication system;

310-emitter;

312-solution multitasking module;

314-modulation module;

316-G-STBC coder;

318 to a controller;

320-receiver;

322-space-time code matched filter;

324-group OSIC detector;

326 to a multiplexer;

328-selecting the signal;

T1-TN-transmit antenna;

R1-RM-receiving antenna;

S410-S460;

S510-S550;

S610-S650;

(K) -an orthogonal matrix;

n-total number of transmit antennas;

k to symbol time.

Detailed Description

Fig. 3 is a diagram illustrating a communication system according to an embodiment of the invention. The communication system 300 includes at least one transmitter 310 and one receiver 320. Please note that the communication system of the present invention is a transceiver architecture including a receiver and a transmitter design, which can be used in a MIMO system. The transmitter 310 includes a de-multiplexing module 312, a modulation module 314, a group-wise space-time block code (G-STBC) encoder 316, a controller 318, and N transmit antennas T1-TN. The N transmitting antennas T1-TN are divided into Q antenna groups, and each antenna group is composed of 2-4 antennas. The receiver 320 includes a space-time code channel Matched Filter (MF) 322, a group OSIC detector 324, a multiplexer 326, and M receive antennas R1-RN, where M is greater than or equal to Q.

First, at the transmitter, the input data string is demultiplexed by the demultiplexing module 312 to generate a plurality of sub-data strings, modulated into data symbols (symbols) by the modulator 314, space-time coded by the group space-time block code encoder 316 according to the selected codeword structure, and finally transmitted to the receiver 328 by the antennas T1-TN. In this embodiment, the codeword structure includes the number of antennas used and the required transmission symbol time.

At the receiver end, the receiver 320 receives the transmitted signal and performs signal demodulation via the space-time channel matched filter 322. The space-time channel matched filter 322 is a matched filter channel matrix MFCM, and the space-time code channel matched filter 322 can reduce the spatial dimension of the received signal according to the MFCM for further decoding by the subsequent group OSIC detector 324. The decoding principle and form of the matched filter channel matrix MFCM will be explained in detail as follows. The group OSIC detector 324 then receives the signal after passing through the space-time code channel matched filter 322, and performs interference cancellation and signal detection according to the OSIC detection rule. The above detection method corresponds to the modulation method used by the transmitting end, for example, when the transmitting end uses BPSK modulation, the detection method of the receiving end is the detection for BPSK, and when the transmitting end uses QPSK modulation, the detection method of the receiving end is the detection for QPSK. In other words, different modulation schemes will correspond to different detection methods. Since BPSK and QPSK decoding belong to the detection of real and complex symbols, respectively, the following discussion will also be directed to the detection of real and complex symbols, respectively. For example, for the detection of real symbols, the detection can be performed by using the antenna group OSIC detection method directly, that is, all symbol data of an antenna group can be detected simultaneously in each iteration. For complex symbol detection, it is necessary to perform detection by matching with antenna group detection rule, two-stage detection rule and recursive (recursive) method. This is because for the detection of complex symbols, only 2L are used_qThe real symbols are half the symbols (i.e. of the unit block.，L_qOne) can be detected simultaneously in a certain iteration of the OSIC, and the real part (Re) of the complex symbols belonging to a certain antenna group is the imaginary part (Im) of the complex symbols. The antenna group detection method for complex symbols means that the results of the real part (Re) and the imaginary part (Im) of the same antenna group are averaged to generate a matrix structure similar to real symbols, and the detection can be performed by the antenna group OSIC detection method. The two-stage detection method is to detect the antenna group (N) with higher link capability first_q4 or N_q3), re-detecting antenna groups with lower link capability (N)_q2) to reduce its computational complexity. The recursive method combines the two-stage detection rule and utilizes the previous operation result to obtain the detection code element of the next stage, thereby effectively reducing the operation complexity. By combining the above decoding method, the computational complexity of the receiver is reduced, and the design complexity of the receiver is simplified.

In addition, in this embodiment, the group OSIC detector 324 calculates or looks up the table to obtain the Bit Error Rate (BER) corresponding to each antenna group according to the given environmental parameters. The group OSIC detector 324, after calculating, finds the antenna group with the lowest BER among all the BERs and transmits this information back to the transmitter 310 using the selection signal 328.

Based on the low complexity group-based OSIC detection architecture, the present invention will develop a set of group-based space-time codes without total transmit antenna number limitation. Under the required transmission power and data transmission rate limitation, a space-time word selection criterion is provided to properly select the best word according to the minimum BER judgment condition.

It is noted that the optimal space-time codeword structure for transmission in the present invention depends on the feedback information of the receiver. Therefore, initially, the transmitter may first send a training symbol (training symbol) with a codeword format to the receiver, and after the receiver calculates the codeword with the minimum BER, the transmitter and the receiver perform actual data transmission with the codeword structure corresponding to the minimum BER. In addition, since the transmitter has different codeword structures for transmission selection, the transmitter is designed to send a reselection signal when the transmission performance is low or the environment changes, so as to request the receiver to select another codeword structure with a smaller BER for transmission. Therefore, the invention can provide an adaptive transmission adjustment mechanism.

FIG. 4 shows a flowchart of word selection according to an embodiment of the present invention. First, the transmitter 310 is divided into Q antenna groups according to the number N of transmit antennas, providing possible space-time codeword forms (step S410). In the invention, because the signals adopt O-STBC coding, the number of the antennas in each antenna group can only be 2, 3 or 4 to meet the orthogonal characteristic. The selection of the code form includes different antenna configuration modes. For example, assuming a total number of transmit antennas N of 10, the antenna configuration specified by one codeword may be a set of 2 antennas, and thus may be divided into 5 antenna groups of Σ₁And (2, 2, 2, 2, 2). Similarly, assuming another antenna configuration with 3-and 4-antennas as a group, the antenna configuration may be sigma₂And (3, 3, 4). Suppose that the total number of T1-T10 antennas is according to sigma₁The code format of (1), T1 and T2 are used as a group to transmit data simultaneously, and T3 and T4 are used as a group to transmit data simultaneously; if according to sigma₂The code format of (1), then T1, T2 and T3 are used as a group to transmit data simultaneously. The Q antenna groups are then transmitted to the receiver 320. The receiver 320 receives the codewords (in antenna groups) (step S420), and calculates or looks up the table to obtain the BER value corresponding to each space-time codeword (step S430). The above-mentioned calculation or table lookup will be described in detail as follows. Then, the receiver 320 finds the codeword with the minimum BER and sends back the selection signal containing this information to the transmitter 310 (step S440). Then, the transmitter 310 determines the codeword structure for transmission according to the codeword structure specified in the selection signal (step S450). Finally, the transmitter and the receiver use the codeword structure to transmit data (step S460).

Please refer to fig. 5 and fig. 6, which illustrate a flowchart of the character code selecting method according to fig. 4 at the transmitting end and the receiving end, respectively. As shown in fig. 5, the transmitter 310 divides the antenna groups into Q groups according to the total number N of transmit antennas to generate possible candidate codewords (step S510). The characteristics of space-time codewords, including coding rate, diversity gain, and receiver computational complexity, are affected by the antenna group structure (i.e., the codeword structure). For N total transmit antenna numbers, different antenna group configurations may provide different coding rates even with the same number of antenna groups. Furthermore, different antenna group configurations may provide different diversity gains, resulting in different communication link performance. In addition, different antenna group configurations will also result in different receiver computation complexities based on the group-wise OSIC detection algorithm. Therefore, the selection of the optimal space-time codeword should take into account the three characteristics mentioned above simultaneously. Table three lists all possible space-time codeword selections L from total antenna number N2 to N16^NThe representation and the corresponding coding rate R^N. In TABLE III, J^NRepresenting the number of possible words, S, in the total number of antennas N^NAntenna configuration, L, representing each codeword^NThe number of iterations required for detection of the OSIC is indicated. Thus, different total transmit antenna numbers N will have different possible word structures.

N	J^N	Possible Tian line group bridge, S^N	Coding rate	Maximum generationsNumber of times
					2	1	(2)	1	1
3	1	(3)	0.5	1
					4	2	(4)，(2，2)	0.5，2	1，2
5	1	(2，3)	1.5	4
					6	3	(3，3)，(2，4)，(2，2，2)	1，1.5，3	2，4，3
7	2	(3，4)，(2，2，3)	1，2.5	2，6
					8	4	(4，4)，(2，3，3)，(2，2，4)，(2，2，2，2)	1，2，2.5，4	2，6，6，4
9	3	(3，3，3)，(2，3，4)，(2，2，2，3)	1.5，2，3.5	6，3，8
					10	5	(3，3，4)，(2，4，4)，(2，2，3，3)，(2，2，2，4)，(2，2，2，2，2)	1.5，2，3，3.5，5	3，6，8，8.5
11	4	(3，4，4)，(2，3，3，3)，(2，2，3，4)，(2，2，2，2，4)	1.5，2.5，3，4.5	3，8，8，10
					12	7	(4，4，4)，(3，3，3)，(2，3，3，4)，(2，2，4，4)，(2，2，2，3，3)，(2，2，2，2，4)，(2，2，2，2，2，2)	1.5，2，2.5，3，4，4.5，6	3，4，8，8，1010，6
13	5	(3，3，3，4)，(2，3，4，4)，(2，2，3，3，3)，(2，2，2，3，4)，(2，2，2，2，2，3)	2，2.5，3.5，4，5.5	4，8，10，10，12
					14	8	(3，3，3，4)，(2，4，4，4)，(2，3，3，3，3)，(2，2，3，3，4)，(2，2，2，4，4)，(2，2，2，2，3，3)，(2，2，2，2，2，4)，(2，2，2，2，2，2，2)	2，2.5，3，3.5，4，5，5.5，7	4，8，10，10，1012，12，7
15	7	(3，4，4，4)，(3，3，3，3，3)，(2，3，3，3，4)，(2，2，3，4，4)，(2，2，2，3，3，3)，(2，2，2，2，3，4)，(2，2，2，2，2，2，3)	2，2.5，3，3.5，4.5，5，6.5	4，5，10，10，1212，14

16

10

(4，4，4，4)，(3，3，3，3，4)，(2，3，3，4，4)，(2，2，4，4，4)，(2，2，3，3，3，3)，(2，2，2，3，3，4)，(2，2，2，2，4，4)，(2，2，2，2，2，3，3)，(2，2，2，2，2，2，4)(2，2，2，2，2，2，2，2)

2，2.5，3，3.5，4，4.5，5，6，6.5，8

4，5，10，10，1212，12，14，14，8

Watch III

Fig. 8a-8e are schematic diagrams of 5 possible codewords (3, 3, 4), (2, 4, 4), (2, 2, 3, 3), (2, 2, 2, 4), and (2, 2, 2, 2, 2, 2) respectively, when the total number of transmit antennas is 10 according to an embodiment of the invention. For example, fig. 8c is a schematic diagram of a G-STBC codeword structure when the total number of transmit antennas N is 10, the antennas are divided into 4 antenna groups, and the codeword (antenna group structure) structure is (2, 2, 3, 3), where 2 antennas are used for a part using 2 × 2 STBC, 3 antennas are used for a part using 3 × 8 STBC, and the STBC used in each group is an orthogonal STBC (O-STBC), and the codeword structure is based on a symbol time length N being 8. As can be seen from fig. 8a-8e, the antenna group structure and the codeword structure are determined under the condition that the space-time codes in each group must be orthogonal.

These possible space-time codeword forms are then transmitted to the receiver 320 (step S520). Then, wait for the receiver 320 to return a message. Then, it is determined whether the receiver 320 has returned the information (step S530). If the receiver 320 has not returned the message (no in step S530), the process returns to step S530 to continue the determination. If the receiver 320 has returned the message (yes in step S530), the transmitter 310 determines the best code word to be the space-time code word for transmission according to the best code word in the returned message (step S540). Finally, the data is encoded by the optimal space-time character code and then transmitted to the receiving end for data transmission (step S550).

At the receiving end, as shown in fig. 6, the receiver 320 receives the space-time word form sent by the transmitter 310 (step S610). The code format includes different antenna configurations. For example, assuming that the total number of transmit antennas N is 10, it can be seen from table three that 5 possible codeword structures are included. Next, according to the decoding method, a BER value corresponding to each space-time codeword is calculated (table lookup) (step S620). Referring to table three, when the total number of transmitting antennas N is 10, 5 BER values BER 1-BER 5 are obtained after step S620 is executed, in the above example, the codeword Σ₁(2, 2, 2, 2, 2) will have a corresponding BER value BER1, codeword Σ₂(3, 3, 4) has a corresponding BER value BER2, and so on. Next, the codeword format having the minimum BER value is found (step S630). Spectral efficiency h (M) of demand for a system_j ^N) If a bit transmission rate R is given_b(i.e., the modulation format) then the corresponding space-time code will be defined. Thus, at a given total transmit power P_TAnd the required data transmission rate R_bUnder the limitation of (1), if the jth character code is selected, the requirement is satisfied

<math><mrow><mi>j</mi><mo>=</mo><mi>arg</mi><munder><mi>min</mi><mrow><mo>&ForAll;</mo><mi>j</mi></mrow></munder><msub><mi>P</mi><mi>e</mi></msub><mrow><mo>(</mo><msubsup><mi>S</mi><mi>j</mi><mi>N</mi></msubsup><mo>|</mo><mo>{</mo><msub><mi>P</mi><mi>T</mi></msub><mo>,</mo><msub><mi>R</mi><mi>b</mi></msub><mo>}</mo><mo>)</mo></mrow><mo>.</mo></mrow></math>

Wherein P is_eRepresents the overall BER value for the jth codeword, which can be calculated as follows:

in the above formula, P_e，q(j) Is the BER of the q antenna group in the j mode, which is defined as the combination of a particular space-time codeword and the modulation format used. Suppose that the error propagation effect (P) is neglected in the iterative operation process of OSIC_e，q(j) Can be approximated as a signal-to-interference-plus-noise ratio (SINR) gamma after detection_qAnd a transmission data rate R_b，qFunction of (2)

Wherein

While

The detected noise power and symbol mean square error for each of the q-th antenna group. Wherein

Is defined as H_c，i(H_c，iAs will be described later), e_lThen is at

The l-th unit standard vector (unit standard vector). The invention assumes that M-ary QAM modulation is used, the BER of the q-th antenna group can be approximated as follows

Wherein,

<math><mrow><msup><mn>2</mn><msub><mi>R</mi><mrow><mi>b</mi><mo>,</mo><mi>q</mi></mrow></msub></msup><mo>=</mo><mover><mi>M</mi><mo>&OverBar;</mo></mover><mo>,</mo></mrow></math>

erfc () is a complementary error function (complementary error function).

From the above formula, at the required transmission power P_TAnd a data transmission rate R_bUnder the limitation, according to the BER performance, the word code corresponding to the minimum BER performance can be selected. It is assumed that BER2 is the minimum BER value among all the calculated BER values. The code format corresponding to the minimum BER value is regarded as the optimal space-time code format, so the optimal code format is the corresponding code format sigma₂That is, (3, 3, 4) will be the best codeword format. Thus, the receiver 320 generates a selection signal containing the optimal codeword format as ∑₂The selection signal is sent back to the transmitter 310 (step S640). Finally, the transmitter 310 will be based on this optimal codeword format (Σ)₂) Encoding is performed and then data transmission is performed with the receiver (step S650).

In summary, the transmission system and the transmission method thereof according to the present invention can improve the link quality and increase the data rate, and can also have a better balance point between the two. In addition, because the transmitting end uses the O-STBC coding mode, the operation complexity of the receiver can be effectively reduced.

In order to explain the determination of the antenna group and the relationship and influence between various correlation operations, mathematical formulas are described below. Please note that the following doctor's paper "Space-time signal processing for MIMO wireless communications" by the same inventor as the present invention is cited: for the results of Space-time signaling and interference suppression (hereinafter referred to as "document"), detailed derivation processes can be referred to the inventor's paper, and only some of the results are summarized and improved to assist the description.

First, a format of data is defined, and a G-STBC system under a Rayleigh flat-fading environment is considered, as shown in fig. 3, where N antennas are placed at a transmitting end and M antennas are placed at a receiving end. The N transmit antennas are divided into Q antenna groups (N)₁，...，N_Q) Each antenna group uses2 to 4 antennas so that N₁+...+N_QN. For the qth group, B will be consecutive_qThe data code elements are coded in an O-STBC coding mode and are coded at K_qWithin one code element time by N_qThe antennas are transmitted.

If define K ═ max { K₁，...，K_QK ═ K }, then_qK_qWherein k is_q＝K/K_q. L will be transmitted out of each antenna group in K symbol times_q＝k_qB_qIndependent symbols, so that a common symbol is transmitted by Q antenna groups in K symbol times

A data symbol. The code word for each antenna group, called group code, may be represented by N_qSpace-time character code matrix (codeword) X of xK_qIs fully described. The q group data code element s_q(k) Divided into blocks of data strings such as

s_{q, l} (k) = s_{q} (L_{q} k + \tilde{l} 1), l = 1, . . ., L_{q}

The space-time code of the qth group can be written as

Wherein A is_q，lIs a space-time modulation matrix (modulation matrix). For the convenience of subsequent analysis, the following definitions are provided

{\tilde{s}}_{q, l} (k) = Re {s_{q, l} (k)}, l = 1, . . ., L_{q},

And

{\tilde{s}}_{q, l} (k) = Im {s_{q, l - L_{q}} (k)}, 1 = L_{q} + 1, . . ., 2 L_{q} .

assuming that M (Q) antennas are used at the receiving end, the signals received by the M antennas in K symbol times are

Wherein

H_{q} = \sqrt{P_{q}} C_{q},

P_qIs the transmission power of the qth group and satisfies P₁+...+P_Q＝P_TAnd P is_TIs the total transmit power; in addition, C_qFor the q group to receiver MIMO channel matrix, finally V (k) C^M×KIs a noise matrix. The following assumptions will be used in the following discussion:

(a1) data symbol s_q(k) Q is 1., Q is i.i.d., the average value is 0(zero-mean), the variance is 1(unit-variance), and the same modulation technique is employed.

(a2) Each antenna group transmitting the same power, i.e. P₁＝...＝P_Q＝P_T/Q。

(a3)C_qQ1.., Q, each element in the matrix is an i.i.d. complex gaussian random variable and its average value is 0, the variance is 1, and it is assumed to remain unchanged for K symbol times.

(a4) V (k) is space-time white noise and has an average value of 0 and a variance of s_v ²。

(a5) According to O-STBC, when real symbols are used and 2 ≦ N_q4 or less, or using complex symbols and N_qIn case of 2, an orthogonal space-time block code with a coding rate of 1(unit-rate) is used; when using complex symbols and 3 ≦ N_qUnder the condition of less than or equal to 4, an orthogonal space-time zone block code with the code rate of 1/2(half-rate) is adopted.

Real number vector model

For analytical convenience, (2) was adapted to the following 2KM X1 linear vector model

y_{c} (k) : = {[{\tilde{y}}^{T} (k), {\tilde{y}}^{T} (k + 1), . . ., {\tilde{y}}^{T} (k + K - 1)]}^{T} = H_{c} s_{c} (k) + v_{c} (k)

(3)

Wherein

<math><mrow><mover><mi>y</mi><mo>~</mo></mover><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>:</mo><mo>=</mo><msup><mrow><mo>[</mo><mi>Re</mi><mo>{</mo><msup><mi>y</mi><mi>T</mi></msup><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>}</mo><mi>Im</mi><mo>{</mo><msup><mi>y</mi><mi>T</mi></msup><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>}</mo><mo>]</mo></mrow><mi>T</mi></msup><mo>&Element;</mo><msup><mi>R</mi><mrow><mn>2</mn><mi>M</mi></mrow></msup></mrow></math>

<math><mrow><mrow><msub><mover><mi>s</mi><mo>~</mo></mover><mi>q</mi></msub><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>:</mo><mo>=</mo><msup><mrow><mo>[</mo><mi>Re</mi><mo>{</mo><msubsup><mi>s</mi><mi>q</mi><mi>T</mi></msubsup><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>}</mo><mi>Im</mi><mo>{</mo><msubsup><mi>s</mi><mi>q</mi><mi>T</mi></msubsup><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>}</mo><mo>]</mo></mrow><mi>T</mi></msup><mo>&Element;</mo><msup><mi>R</mi><mrow><mn>2</mn><msub><mi>L</mi><mi>q</mi></msub></mrow></msup></mrow></math>

s_{q} (k) : = {[s_{q, 1} (k), s_{q, 2} (k), . . ., s_{q, L_{q}} (k)]}^{T}

H_c∈R^2KM×2LT

Is an equivalent channel matrix, and

<math><mrow><msub><mi>s</mi><mi>c</mi></msub><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>:</mo><mo>=</mo><msup><mrow><mo>[</mo><msubsup><mover><mi>s</mi><mo>~</mo></mover><mn>1</mn><mi>T</mi></msubsup><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>,</mo><msubsup><mover><mi>s</mi><mo>~</mo></mover><mn>2</mn><mi>T</mi></msubsup><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>,</mo><mo>.</mo><mo>.</mo><mo>.</mo><msubsup><mover><mi>s</mi><mo>~</mo></mover><mi>Q</mi><mi>T</mi></msubsup><mrow><mo>(</mo><mi>k</mi><mo>)</mo></mrow><mo>]</mo></mrow><mi>T</mi></msup><mo>&Element;</mo><msup><mi>R</mi><mrow><mn>2</mn><msub><mi>L</mi><mi>T</mi></msub></mrow></msup><mo>,</mo></mrow></math>

v_c(k)∈^R2KMis a noise vector. Will y_c(k) The left and right of the equation are each multiplied by H_cWill produce a matched-filtered (MF) data vector

z (k) : = H_{c}^{T} y_{c} (k) = F s_{c} (k) + v (k) - - - (4)

Wherein

<math><mrow><mi>F</mi><mo>:</mo><mo>=</mo><msubsup><mi>H</mi><mi>c</mi><mi>T</mi></msubsup><msub><mi>H</mi><mi>c</mi></msub><mo>&Element;</mo><msup><mi>R</mi><mrow><mn>2</mn><msub><mi>L</mi><mi>T</mi></msub><mo>×</mo><msub><mrow><mn>2</mn><mi>L</mi></mrow><mi>T</mi></msub></mrow></msup></mrow></math>

To match the filtered channel matrix (MFCM),

v (k) : = H_{c}^{T} v_{c} (k) .

the signal is detected based on the model (4) as follows.

OSIC detection method using real symbols

The transmitted signal is detected using the algorithm proposed in the document one. By using the special structure of the matched filter channel matrix F, the OSIC detector can simultaneously detect all L of an antenna group in each iteration_qThe "antenna group OSIC detection" method is called "K symbol data.

A. Matched filter channel matrix

In order for the OSIC algorithm to effectively detect the transmitted signal, the structure of F must be analyzed. Since the detection rule is proposed in the literature and has a complete analysis, it is applicable to the real symbol case, and therefore only the result is stated here.

Defining O (K) as a set of K real orthogonal matrices with K independent variables; (K, L) is a set of all K real orthogonal matrices with L independent variables.

Results II.1: considering real symbols, and 2 ≦ N_p，N_qLess than or equal to 4. From O-STBC, K ∈ {2, 4 }. Definition F_p，qThe (p, q) th K × K block matrix of F, where F is defined as (4). Then F can be known_q，q＝α_qI_KAnd F_p，qE (K), if p ≠ q. ,

the results of result II.1 are collated in Table I, where F_p，q ^(s，t)Is F_q，qThe (s, t) -th block matrix of (a). The structure of F is shown in FIG. 7.

Watch 1

B. Group type OSIC detection algorithm

F will be found next^-1In a rough sense, will have the same structure as F. First define phi_KL(L) is a set of all inverse (invertible) KL real symmetric (symmetric) matrices such that for X ∈ Φ_KL(L) in which X_k，lThe (K, l) th K × K sub-matrix of X can obtain X_l，l＝β_qI_KAnd X_k，le.O (K) when k ≠ l.

Fact II.1: if F ∈ Φ_KL(L) then F^-1The same result is obtained (see the literature I). ,

according to fact II.1, the following result II.2 is obtained:

results II.2: considering real symbols, and 2 ≦ N_p，N_qAnd (4) is less than or equal to 4, and K is equal to {2, 4 }. F is belonged to R^LT×LTIs defined in (4). Then each F^-1The K × K block diagonal (diagonals) submatrix of (a) is a constant identity matrix; each F^-1The off-diagonal (off-diagonals) K × K block sub-matrix of (a) belongs to o (K). From this, F is known^-1All KL diagonal elements of (1) have L ═ Q different levels (levels)β _l1.. L, i.e., L

diag(F^-1)＝{β₁，...，β₁，β₂，...，β₂，...，β_L，...，β_L}(5)，

Wherein each level in (5)β _lThe number of (2) is K. As shown in (5), based on zero-forcing (ZF) criterion (or Minimum Mean Square Error (MMSE) criterion), the OSIC detector can simultaneously detect K symbols in the initial iteration.

Based on the detection-subtraction procedure of OSIC, it can be known that in the ith iteration of OSIC, i is 1_i ^-1Wherein

F_{i} = H_{c, i}^{T} H_{c, i} - - - (6)

The matched filter channel matrix in the ith iteration of OSIC can be obtained by using the sum of the coefficients of H and H_cWherein i blocks (i.e., H) in units of K columns (columns) are deleted_c，i) (corresponding to the signal detected in the previous iteration). From this, F is known_i∈F(L-i)。

Results II.3: from fact II.1, it can be seen that

And is

Whereinβ _i，lTo be distributed in F_i ^-1Diagonal on-line 1 different levels, eachβ _lK in each case. ,

from the above analysis, the algorithm can perform the antenna group OSIC detection.

OSIC detection method using complex symbols

Although antenna group based OSIC detection can be performed using real symbols. However, antenna group based OSIC detection is not possible with complex symbol usage. Only by 2L_qOne real symbol is a half symbol (i.e., L._qOne) can be detected simultaneously in a certain iterative process of the OSIC, and the symbols are imaginary parts if they are not real parts of complex symbols belonging to a certain antenna group.

A. Matched filter channel matrix

At complex symbols, there are the following results that differ from result ii.1.

Results III.1: considering complex symbols, and 2 ≦ N_p，N_qLess than or equal to 4. From O-STBC, K ∈ {2, 8 }. Definition F_p，q2L of (p, q) th of F_q×2L_qThe block matrix, wherein F is defined in (4). The results as in table two can be obtained.

Watch two

B. Group type OSIC detection algorithm

From the result III.1, F can be presumed_i ^-1Will no longer have a similar F_iThe structure of (1). In analysis F_i ^-1Before the structure of (2), the following parameters need to be defined. First, in the ith iteration of the OSIC, the real number symbol of half of a certain antenna group is defined as a decision group (decision group) Γ_i，g，g＝1，...，G_iWherein G is_iIs the total number of decision groups in the ith iteration. Thus, each antenna group will have two decision groups, each with L_qA real symbol. Further define I_i，1And I_i，1/2The decision group index (index) sets for the i-th iteration are the code rate 1 and code rate 1/2 space-time codes, respectively, and thus G_i，1＝|I_i，1I and G_i，2＝|I_i，2And | is the number of parameters included in each set. Last definition of I_i：＝I_i，1∪I_i，1/2Set of total decision group labels in the ith iteration, therefore G_i＝|I_iL. According to the above definition, it can be seen that

(8)

In addition, if F is to be_iIs divided into G_iA decision group, each decision group having the same non-zero value, i.e. a value

Suppose { a_i，g}_g＝1 ^GiHaving D_i(G_i) At different levels, it can be known

Is defined as follows

Set of all inverse J x J real symmetric matrices such that

As a matter of fact, it is possible to obtain: (1) each block diagonal submatrix of X is a constant unit matrix

G1, wherein M_g ∈{2，4，8}。(2){a_g}_g＝1 ^GHas D different levels, wherein

(3) M for i, j ≠ j, X_i×M_jEach of the 4 x 4 block sub-matrices in the block off-diagonal sub-matrix belongs to (4) or a zero matrix. Wherein, when D ═ G, then

Will be rewritten as

Results III.2: consider a complex symbol. Assume that code rate 1 and code rate 1/2 for space-time codes exist in both OSIC's i iterations, i.e.And is

Definition of i (G), G1_iIs the element of g in i, and F_i∈R^Ji×JiFor matching filter channel matrix in the ith iteration of OSIC, wherein

And L is_g∈{2，4，8}。，

Therefore, ifThenFrom the result III.2, F_i ^-1Having G_i(≥D_i) A different level { beta_i，g}_g＝1 ^Gi(wherein F_iOnly D_iA different level), i.e., a

As shown in (11), only 2L of the total amount of the active carbon is required_qOne real symbol is a half symbol (i.e., L._qOne) can be detected simultaneously in some iterative process of the OSIC, and the symbols are imaginary parts of complex symbols belonging to an antenna group, if not real parts. Therefore, such a detection characteristic will cause a computational burden. To solve the above problems, techniques for reducing the complexity of the operation will be developed later.

Implementation of group-based OSIC detection techniques using complex symbols

Some implementation issues of group-wise OSIC detection will be discussed, including antenna group-wise detection rule, two-stage (two-stage) detection rule, and recursive (recursive) implementation. Since the implementation of the group-based OSIC detection using real symbols can be achieved by the method proposed in the document, only how to implement the group-based OSIC detection using complex symbols is discussed below.

A. Antenna group type detection method

Distribution in F as described in result III.2_i ^-1In a certain antenna group on diagonal

An element will have two different levels

And

. To implement antenna group detection, one of the simplest methods is to search for F directly_i ^-1The labels of the diagonal online decision group are checked to see which label corresponds to the smallest level. This method is simple but has poor performance. Thus, a two level hierarchy would correspond to a certain antenna groupAnd

carry out averaging

And F is_i ^-1xIs on-line and opposite angles of

An element is rewritten as

Can be according to beta_i，1，β_i，2，...，β_i，Q-iSearch the labels and check which label has the smallest corresponding hierarchy. This method of searching for labels is not the best method and therefore performance is degraded, but computer simulations show that performance degradation is not severe.

B. Two-stage detection method

Theoretically, antenna groups with higher diversity gain (i.e., with higher number of antennas or lower coding rate) are more robust against channel fading and have better performance. In view of the above, the present invention provides a two-stage detection method. By first detecting the antenna group (N) with the lower code rate_q4 or N_q3) re-detection of antenna group (N) with higher coding rate_q2) to reduce its computational complexity. Similarly, the label search method is not the best method, and thus the performance is degraded, but computer simulations show that the performance degradation is not severe.

C. Recursive formal implementation

To further slow down the receiver computation, a recursive rule like that proposed in the document one is used to implement the group-wise OSIC detector. The method proposed in the document is not directly and completely applicable to the present invention, and some modifications are required to be used.

From the results III.2, it is clear that

The smallest dimension in which orthogonal matrices can be formed is 4 x 4, where,is F_i ^-1The (p, q) th sub-matrix of (1). Since this recursive rule based implementation can only process one block orthogonal matrix at a time, when F is the case_iIf the dimension of (a) is large, implementing the group-type OSIC detector in a recursive manner directly needs more recursion times, resulting in larger calculation amount. But fortunately, if the two-stage detection method is combined, the calculation amount can be reduced. Suppose in F_i ^-1Elements that belong to an antenna group diagonally online have two different levels. If F_iIs the matched filtering channel matrix in the ith iteration operation. Then F_i1 ^-Can be divided into

(13)

Wherein

<math><mrow><msub><mi>B</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msub><mo>&Element;</mo><msup><mi>R</mi><mrow><mn>2</mn><mrow><mo>(</mo><msub><mi>L</mi><mi>T</mi></msub><mo>-</mo><msubsup><mi>Σ</mi><mrow><mi>j</mi><mo>=</mo><mn>1</mn></mrow><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msubsup><msub><mi>L</mi><msub><mi>q</mi><mi>j</mi></msub></msub><mo>)</mo></mrow><mo>×</mo><msub><mrow><mn>2</mn><mi>L</mi></mrow><msub><mi>q</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msub></msub><mo>,</mo></mrow></msup></mrow></math>

D_{i - 1} = d_{i - 1} I_{2 L_{q_{i - 1}}},

And d_i-1Is a constant. Further, if it is assumed that

F of dimension_i-1Is F_i-1 ^-1A primary sub-matrix of

F_{0}^{- 1} = F^{- 1} .

From (13) using inversion lemma for block matrix, F_i-1Can be expressed as follows

<math><mrow><msub><mover><mi>F</mi><mo>&OverBar;</mo></mover><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msub><mo>=</mo><msup><mrow><mo>(</mo><msub><mi>F</mi><mi>i</mi></msub><mo>-</mo><msub><mi>B</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msub><msubsup><mi>D</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow><mrow><mo>-</mo><mn>1</mn></mrow></msubsup><msubsup><mi>B</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow><mi>T</mi></msubsup><mo>)</mo></mrow><mrow><mo>-</mo><mn>1</mn></mrow></msup><mo>-</mo><mo>-</mo><mo>-</mo><mrow><mo>(</mo><mn>14</mn><mo>)</mo></mrow></mrow></math>

From (14), it is known

<math><mrow><msub><mi>F</mi><mi>i</mi></msub><mo>=</mo><msubsup><mover><mi>F</mi><mo>&OverBar;</mo></mover><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow><mrow><mo>-</mo><mn>1</mn></mrow></msubsup><mo>+</mo><msub><mi>B</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msub><msubsup><mi>D</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow><mrow><mo>-</mo><mn>1</mn></mrow></msubsup><msubsup><mi>B</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow><mi>T</mi></msubsup><mo>-</mo><mo>-</mo><mo>-</mo><mrow><mo>(</mo><mn>15</mn><mo>)</mo></mrow></mrow></math>

Obtained by using matrix inversion lemma and deriving

<math><mrow><msubsup><mi>F</mi><mi>i</mi><mrow><mo>-</mo><mn>1</mn></mrow></msubsup><mo>=</mo><msub><mover><mi>F</mi><mo>&OverBar;</mo></mover><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msub><mo>-</mo><msub><mi>E</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msub><msubsup><mi>C</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow><mrow><mo>-</mo><mn>1</mn></mrow></msubsup><msubsup><mi>E</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow><mi>T</mi></msubsup><mo>-</mo><mo>-</mo><mo>-</mo><mrow><mo>(</mo><mn>16</mn><mo>)</mo></mrow></mrow></math>

Wherein,

<math><mrow><msub><mi>E</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msub><mo>:</mo><mo>=</mo><msub><mover><mi>F</mi><mo>&OverBar;</mo></mover><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msub><msub><mi>B</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msub></mrow></math>

and is

<math><mrow><msub><mi>C</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msub><mo>=</mo><msubsup><mi>B</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow><mi>T</mi></msubsup><msub><mover><mi>F</mi><mo>&OverBar;</mo></mover><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msub><msub><mi>B</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msub><mo>+</mo><msub><mi>D</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msub><mo>=</mo><mfenced open='[' close=']'><mtable><mtr><mtd><msub><mi>c</mi><mrow><mn>1</mn><mo>,</mo><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msub><msub><mi>I</mi><msub><mi>L</mi><msub><mi>q</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msub></msub></msub></mtd><mtd><msub><mn>0</mn><msub><mi>L</mi><msub><mi>q</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msub></msub></msub></mtd></mtr><mtr><mtd><msub><mn>0</mn><msub><mi>L</mi><msub><mi>q</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msub></msub></msub></mtd><mtd><msub><mi>c</mi><mrow><mn>2</mn><mo>,</mo><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msub><msub><mi>I</mi><msub><mi>L</mi><msub><mi>q</mi><mrow><mi>i</mi><mo>-</mo><mn>1</mn></mrow></msub></msub></msub></mtd></mtr></mtable></mfenced></mrow></math>

Wherein,

c_{j, i - 1}, j = 1,2,

is a constant.

The derivation described above utilizes F_i-1And F_i-1 ^-1Under the information of (2), a simple recursive formula is provided to calculate F_i ^-1Without any direct inverse matrix operation, the complexity of the operation can be effectively reduced.

The above description provides several different embodiments or embodiments applying different features of the present invention. The particular devices and methods in the examples are provided to help explain the principal spirit and objects of the invention, although the invention is not limited thereto.

Therefore, although the present invention has been described with reference to the preferred embodiments, it should be understood that various changes and modifications can be made therein by those skilled in the art without departing from the spirit and scope of the invention as defined by the appended claims.

Claims

1. A method for selecting a codeword for a multiple-input multiple-output (MIMO) communication system having multiple transmit antennas and multiple receive antennas, the method comprising:

a transmitter providing complex digital code form;

a receiver receives the word forms and calculates or looks up the table to obtain a corresponding bit error rate of each word form according to a decoding mode;

the receiver selects a codeword with the minimum bit error rate and transmits the codeword back to the transmitter; and

the transmitter determines a codeword for data transmission based on the codeword type with the minimum bit error rate,

wherein each of the codewords is encoded using orthogonal space-time codes, and the decoding is a sequenced progressive interference cancellation detection method and wherein the receiver further performs the sequenced progressive interference cancellation detection by averaging different levels of elements belonging to an antenna group on a diagonal of an inverse matrix of a matched filter channel matrix for a plurality of symbols.

2. The codeword selection method of claim 1 wherein the transmitter performs the steps of:

generating the codeword form according to the number of the transmitting antennas;

receiving a codeword form of the receiver; and

according to the received character code form, the character code for transmission is determined, and the determined character code is used for transmission.

3. The codeword selection method of claim 1 wherein the receiver performs the steps of:

receiving the codeword form of the transmitter;

calculating the corresponding bit error rate of each of the codeword types according to the decoding method; and

transmitting a selection signal to the transmitter for determining the codeword to be transmitted by the transmitter,

wherein the selection signal includes information of the codeword with the minimum bit error rate.

4. The method of claim 1, wherein the codeword selection is a space-time codeword selection.

5. The method for selecting words according to claim 1, wherein each of the words comprises N_qA root antenna, and N_qIs 2, 3 or 4.

6. The codeword selection method according to claim 5, wherein the number of antennas N within each of the codewords_qDifferent.

7. The method of claim 1, wherein the progressive interference cancellation detection method further comprises a two-stage detection method.

8. The method of claim 7 wherein the two-stage detection method comprises detecting groups of words having 3 or 4 antennas first, and then detecting groups of words having 2 antennas.

9. The method of claim 8, wherein the two-stage detection method comprises performing two-stage detection using recursive detection.

10. The codeword selection method as claimed in claim 1, wherein the receiver performs antenna group-wise detection by using the ordered progressive interference cancellation method under detection of real symbols.

11. A communication system having multiple transmit antennas and multiple receive antennas, comprising:

a transmitter for providing complex digital code form; and

a receiver for receiving the word forms and calculating or looking up the table to obtain a corresponding bit error rate of each word form according to a decoding method,

wherein the receiver selects a codeword format with the minimum bit error rate and transmits the codeword format back to the transmitter, the transmitter determines a codeword for data transmission according to the codeword format with the minimum bit error rate,

wherein each of the codewords is encoded using orthogonal space-time codes, and the decoding is a progressive interference cancellation detection method and wherein the receiver further performs the progressive interference cancellation detection by averaging different levels of elements belonging to an antenna group on a diagonal of an inverse matrix of a matched filter channel matrix for a plurality of symbols.

12. The communication system of claim 11, wherein the transmitter further comprises:

a de-multiplexing module for generating multiple subdata strings from the input data;

a modulation module for modulating the sub data string into complex modulation code elements;

a group space-time block code encoder for performing group space-time block coding on the modulated modulation symbols; and

a controller for providing the word type and determining the word for data transmission according to the word type with the minimum bit error rate returned by the receiver.

13. The communication system of claim 12, wherein the receiver further comprises:

a space-time code channel matched filter for reducing the spatial dimension of the encoded signal transmitted by the transmitter;

a detector for detecting the encoded signal; and

a multiplexer for restoring the detected signal to the original data.

14. The communication system of claim 13 wherein each of said codewords comprises N_qA root antenna, and N_qIs 2, 3 or 4.

15. The communication system of claim 14 wherein the number of antennas in each of said codewords is N_qIs different.

16. The communication system of claim 11 wherein the sequenced progressive interference cancellation detection method further performs detection using a two stage detection method.

17. The system of claim 16 wherein the two-stage detection method comprises decoding a codeword set having 3 or 4 antennas prior to decoding a codeword set having 2 antennas.

18. The communication system of claim 17 wherein the two-stage detection method comprises detection using recursive detection.

19. The system of claim 13 wherein the detector is a detector that performs progressive interference cancellation for antenna group detection in decoding real symbols.

20. The communication system of claim 13 wherein the detector is a detector that performs averaging of a real part and an imaginary part of each of the codewords for antenna group-wise detection under complex symbols.