7972

IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 57, NO. 12, DECEMBER 2011

Maximum Rate of Unitary-Weight, Single-Symbol Decodable STBCs Sanjay Karmakar, Student Member, IEEE, K. Pavan Srinath, and B. Sundar Rajan, Senior Member, IEEE

Abstract—It is well known that the space-time block codes (STBCs) from complex orthogonal designs (CODs) are single-symbol decodable/symbol-by-symbol decodable (SSD). The weight matrices of the square CODs are all unitary and obtainable from the unitary matrix representations of Clifford Algebras when the number of transmit antennas is a power of 2. The rate of the square CODs for has been shown to be complex symbols per channel use. However, SSD codes having unitary-weight matrices need not be CODs, an example being the minimum-decoding-complexity STBCs from quasi-orthogonal designs. In this paper, an achievable upper bound on the rate complex of any unitary-weight SSD code is derived to be symbols per channel use for antennas, and this upper bound is larger than that of the CODs. By way of code construction, the interrelationship between the weight matrices of unitary-weight SSD codes is studied. Also, the coding gain of all unitary-weight SSD codes is proved to be the same for QAM constellations and conditions that are necessary for unitary-weight SSD codes to achieve full transmit diversity and optimum coding gain are presented. Index Terms—Anticommuting matrices, complex orthogonal designs, minimum-decoding-complexity codes, quasi-orthogonal designs, space-time block codes.

I. INTRODUCTION

S

PACE-TIME block codes from complex orthogonal designs [1], [2] are popular because they offer full transmit diversity for any arbitrary signal constellation and also are singlesymbol decodable. In fact, CODs are single-real-symbol decodable for constellations such as the rectangular QAM, which can be expressed as a Cartesian product of two PAM constellations, while for constellations such as PSK, CODs are single-complex-symbol decodable. The weight matrices, also called linear dispersion matrices [3] (refer to Section II-A for a definition of

Manuscript received September 16, 2006; revised November 16, 2009; accepted December 02, 2010. Date of current version December 07, 2011. This work was supported in part by the DRDO-IISc Program on Advanced Research in Mathematical Engineering and by the Council of Scientific & Industrial Research (CSIR), India, through Research Grant (22(0365)/04/EMR-II) to B. S. Rajan. The material in this paper was presented in part at the 2006 IEEE International Symposium on Information Theory. S. Karmakar was with the Indian Institute of Science, Bangalore 560012, India. He is now with the Department of Electrical, Computer, and Energy Engineering, University of Colorado, Boulder, CO 80309-0425 USA (e-mail: [email protected]). K. P. Srinath and B. S. Rajan are with the Department of Electrical Communication Engineering, Indian Institute of Science, Bangalore 560012, India (e-mail: [email protected]; [email protected]). Communicated by S. W. McLaughlin, Associate Editor for Coding Techniques. Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TIT.2011.2170106

weight matrices), of the square CODs are all unitary, and a detailed construction method to obtain these weight matrices from irreducible matrix representations of Clifford Algebras has been presented in [4] for transmit antennas. It has also been shown that the maximum rate of the square CODs for transmit antennas is complex symbols per channel use. Although rectangular CODs [5] offer a higher rate, they are not delay efficient, making square CODs more attractive in practice. In general, single-complex-symbol decodable codes need not be CODs. Throughout this paper, unless otherwise mentioned, SSD codes refer to single-complex-symbol decodable codes. The co-ordinate interleaved orthogonal designs (CIODs) [6] have been shown to be SSD codes, while offering full transmit diversity for specific complex constellations only. However, the CIODs have non-unitary-weight matrices. SSD codes that include unitary-weight codes and rectangular designs have been reported in [7], [8] and are popularly known as minimum-decoding-complexity codes from quasi-orthogonal designs (MDCQODs). The rates of both the CIODs and the transmit antennas have class of codes reported in [8] for been shown to be complex symbols per channel use. In [8], the maximum rate of the MDCQODs has been reported, and this rate includes that for rectangular designs. However, the maximum rate of general SSD codes has not been reported so far in the literature, to the best of our knowledge. In this paper, we make the following contributions. • We derive an upper bound on the rate of unitary-weight transmit antennas. This upper bound is SSD codes for complex symbols per channel use. found to be • We give a general construction method to obtain codes that meet this upper bound and further show the interrelationship between the weight matrices of general unitary-weight SSD codes. All known unitary-weight SSD codes including square MDCQODs are special cases of this construction. • We prove that all unitary-weight SSD codes have the same coding gain and specifically for QAM constellations, we provide the angle of rotation that ensures full transmit diversity and optimum coding gain for all unitary-weight SSD codes. The organization of the paper is as follows. Section II gives the system model and relevant definitions. Section III introduces the notion of normalization and its use in the analysis of unitary-weight SSD codes. Section IV provides the upper bound on the rate of unitary-weight SSD codes and the structure of general unitary-weight SSD codes. Full diversity conditions, coding gain calculations for QAM, and simulation results are given in Section V, Sections V-A and V-B, respectively. Discussions on the direction for future research constitute Section VI.

0018-9448/$26.00 © 2011 IEEE

KARMAKAR et al.: MAXIMUM RATE OF UNITARY-WEIGHT, SINGLE-SYMBOL DECODABLE STBCS

Notations: and denote the field of real and complex numbers, respectively and represents . denotes the with complex engroup of invertible matrices of size tries. For any complex matrix and represent the trace, the Frobenius norm, the Hermitian and the and represent the determinant of , respectively. identity matrix and the zero matrix, respectively. For a complex random variable denotes that has a complex normal distribution with mean 0 and variance . For a complex variable and represent the real and the imaginary parts of , respectively, and denotes the absolute value denotes the cardinality of . of . For a set

7973

where matrix

, are the non-zero eigenvalues of the and is the minimum of the rank of for all possible codeword pairs . If the code offers full-diversity for a constellation , then, the coding gain is , where is the minimum of the determinant of the matrix among all possible codeword matrix pairs , with . A. Single-Symbol Decodable Codes In this subsection, we formally define and classify linear SSD codes. Any codeword matrix of a linear dispersion STBC with complex information symbols can be expressed as

II. SYSTEM MODEL We consider Rayleigh quasi-static flat-fading MIMO channel with full channel state information (CSI) at the receiver and no CSI at the transmitter. We assume a MIMO system with transmit antennas and receive antennas. Since we are considering only square STBCs in this paper, the number of time slots is also . The channel model is

where is the codeword matrix, transmitted over channel uses, is a complex white Gaussian noise matrix with i.i.d. entries is the channel matrix with the entries assumed to be i.i.d. circularly symmetric Gaussian random variables and is the received matrix. Definition 1: (STBC) A space-time block code is a set of complex matrices called codeword matrices. For a system with transmit antennas, a codeword matrix is a matrix, where is the number of time slots ( in this paper) and the th entry of the codeword matrix refers to the signal transmitted by the th transmit antenna in the th time slot. Definition 2: (Code rate) If there are independent complex information symbols in the codeword which are transmitted over channel uses, then, the code rate is defined to be complex symbols per channel use. For instance, for the Alamouti code, and . So, its code rate is 1 complex symbol per channel use. Definition 3: (Full-Diversity Code) An STBC encoding symbols chosen from a constellation is said to offer full-diversity iff for every possible codeword pair , with , the codeword difference matrix is full-ranked [10]. In general, whether a code offers full-diversity or not depends on the constellation that it employs. A code can offer full diversity for a certain complex constellation but not for another complex constellation. The CODs are special in this aspect since they offer full-diversity for any arbitrary complex constellation. Definition 4: (Coding Gain) The coding gain of an STBC is defined as

(1) take values from a complex where constellation . Then, , i.e., the number of codewords, is . The set of complex matrices , called weight matrices, define . Notice that in (1), all the weight matrices are required to form a linearly independent set over , since we are transmitting independent information symbols. Assuming that perfect channel state information (CSI) is available at the receiver, the maximum likelihood (ML) decision rule minimizes the metric (2) Since there are different codewords, in general, ML decoding requires computations, one for each codeword. Suppose the set of weight matrices are chosen such that the decoding metric (2) could be decomposed as

which is a sum of positive terms, each involving exactly complex variables only, where . Then, decoding requires computations and the code is called a -symbol decodable code [11]. The case corresponds to codes that include the well known CODs as a proper subclass, and have been extensively studied [1], [4], [6]–[9]. The codes corresponding to are called double-symbol-decodable (DSD) codes. The quasi-orthogonal designs studied in [12]–[14] are proper subclasses of DSD codes. Definition 5: [1] A square complex orthogonal design for transmit antennas is a set of codeword matrices of size , with each codeword matrix satisfying the following conditions: • the entries of are complex linear combination of and their complex conjugates . • (Orthonormality:)

holds for any complex values for

.

7974

IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 57, NO. 12, DECEMBER 2011

A set of necessary and sufficient conditions for is [1], [4]

to be a COD

(3) (4a)

same as that of gain the same for both the STBCs.

, making the coding

The STBCs and are said to be equivalent. To simplify our analysis of unitary-weight codes, we make use of normalization as described below. Let be a unitary-weight STBC and let its codeword matrix be expressed as

(4b) (4c) for

, and (5)

STBCs obtained from CODs [1], [4] are SSD like the well known Alamouti code [9], and satisfy (3), (4), and (5). For to be SSD, it is not necessary that it satisfies (3) and (5), i.e., it is sufficient that it satisfies only (4)—this result was shown in [6]. Since then, different classes of SSD codes that are not CODs have been studied by several authors, [6]–[8]. SSD codes can be systematically classified as follows. 1) Linear STBCs satisfying (3), (4), and (5) are CODs. 2) Linear STBCs satisfying (4) are called SSD codes. These may or may not satisfy (3) and (5). 3) Linear STBCs satisfying (3) and (4) and not satisfying (5) are called unitary-weight SSD codes. 4) Linear STBCs satisfying (4) and not satisfying (3) are called non-unitary-weight SSD codes. These may or may not satisfy (5). The codes discussed in [6] which are called CIODs, constitute an example class of non-unitary-weight SSD codes. The classes of codes studied in [8] are unitary-weight SSD codes. The classes of codes studied in [7], called minimum decoding complexity codes from quasi-orthogonal designs (MDCQOD codes), include some unitary-weight SSD codes as well as nonunitary-weight SSD codes. The notion of SSD codes has been extended to coding for MIMO-OFDM systems in [15] and [16], and recently, low-decoding complexity codes called 2-group and 4-group decodable codes [17]–[20] and SSD codes [21] in particular have been studied for use in cooperative networks as distributed STBCs.

Consider the code . Clearly, from Lemma 1, is equivalent to . The weight matrices of are

With this, a codeword matrix

of

can be written as

We call the code to be the normalized code of . In general, any unitary-weight SSD code with one of its weight matrices being the identity matrix is called normalized unitary-weight SSD code. Studying unitary-weight SSD codes becomes simpler by studying normalized unitary-weight SSD codes. Also, an upper bound on the rate of unitary-weight SSD codes is the same as that of the normalized unitary-weight SSD codes. For the normalized unitary-weight SSD code transmitting symbols in channel uses, the conditions presented in (3) and (4) can be rewritten as (6) (7) (8) (9) for

, and (10) (11)

III. UNITARY-WEIGHT SSD CODES In this section, we analyze the structure of the weight matrices of unitary-weight SSD codes. We make use of the following lemma in our analysis. Lemma 1: Let be a unitary-weight STBC and consider the STBC , where is any unitary matrix. Then, is SSD iff is SSD. Further, both the codes have the same coding gain for the constellation . Proof: The proof is straightforward. For the STBC , the weight matrices are . It is easy to verify that if the matrices satisfy (4), then, the matrices also satisfy (4) and vice-versa. Further, for any pair of distinct codeword matrices and , the eigenvalues of are the

(12) . So, every weight matrix except for and should square to . Shown below is the grouping of weight matrices (We will later show that ).

Except , the elements in the first row should mutually anticommute and also square to . From (8) and (9), it is clear that if , then, should anticommute with all the weight matrices except . So, the upper bound on the rate of a unitary-weight SSD code is determined by the number of mutually anticommuting unitary matrices. The following section deals with determining the upper bound.

KARMAKAR et al.: MAXIMUM RATE OF UNITARY-WEIGHT, SINGLE-SYMBOL DECODABLE STBCS

IV. AN UPPER BOUND ON THE RATE UNITARY-WEIGHT SSD CODES

OF

In this section, we determine the upper bound on the rate of unitary-weight SSD codes and also give a general construction scheme to obtain codes meeting the upper bound. To do so, we make use of the following lemmas regarding matrices of size . matrices with complex entries. Lemma 2: [22] Consider 1) If , with odd, then there are elements of that anticommute pairwise if and only if . 2) If and invertible matrices anticommute pairwise, then the set of products with along with forms a basis for the dimensional space of all matrices over . In each case is a scaled identity matrix. Proof: Available in [22]. Let be anticommuting, anti-Hermitian, unitary matrices (so that ). The following two lemmas are applicable for such matrices. Lemma 3: The product squares to Proof:

with .

This proves the lemma. Lemma 4: Let

and

7975

From Lemma 2, the maximum number of pairwise anticommuting matrices of size is . Hence, the maximum possible value of , i.e., the number of complex information symbols of an SSD code, is , since we also consider as a weight matrix. In order to provide an achievable upper bound on the rate of unitary-weight SSD codes, we first assume that the case is a possibility. Denoting the anticommuting matrices by , we note from Lemma 2 that the set is a basis for over . Therefore, is a basis for over . It can be checked by applying Lemma 4 that the only product matrix that anticommutes with and is , where . So, it must be that . For our construction, we need anticommuting, anti-Hermitian, unitary matrices (so that they square to ). An excellent treatment of irreducible matrix representations of Clifford algebras is given in [4] and the same paper also presents an algorithm to obtain pairwise anticommuting matrices that all square to . In fact, these matrices are precisely the weight matrices (except ) of square CODs. As mentioned before, we denote them by , with , and

It must be noted that the matrices obtained from [4] are not unique, i.e., these are not the only set of mutually anticommuting, anti-Hermitian, unitary matrices of size . It can be noted by applying Lemma 3 that when is odd. We are now ready to prove the main result of the paper. Theorem 1: The rate of a unitary-weight SSD code is upper bounded as

with and . Let . Then the product matrix commutes with if exactly one of the following is satisfied, and anticommutes otherwise. 1) and are all odd. 2) The product is even and is even (including 0). Proof: If , we note that

and if

Now,

Case 1). Since and are all odd, . Case 2). The product is even and is even (including 0). So, .

Proof: We prove the theorem in three parts as follows. Claim 1: . To prove this, let us first suppose that , in which case, we have the following grouping scheme:

Let and . This is possible because of Lemma 2. Considering , since anticommutes with and , every individual term of must anticommute with and . So, we look for all possible candidates from the set which anticommute with and . By applying Lemma 4, the only possible choice is . Since the weight matrices are required to be independent over and in view of the condition in (7), there is no valid possibility for . As a result, a unitary-weight SSD code with independent complex symbols does not exist. Claim 2: . To prove this, we assume that

7976

IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 57, NO. 12, DECEMBER 2011

is a possibility, in which case, we have the following grouping of weight matrices.

Considering , each of the terms that is a linear combination of should anticommute with and . The only possibilities from the set are and or . Therefore, . Next, considering , the only elements anticommuting with and are and . Therefore, . Since should also anticommute with , either or . So, either and or and . In either case, the assignment violates the rule that the weight matrices are linearly independent over . As a result, we can’t have any valid elements as the weight matrices and hence, . Claim 3: . Consider the following grouping scheme of weight matrices:

In the above grouping scheme, if is odd, and if is even. It can be noted that . Clearly, the weight matrices are linearly independent over and satisfy (6)–(12). Hence, an SSD code transmitting complex symbols in channel uses exists. This completes the proof. We observe that for 2 transmit antennas, . So, the rate of a unitary-weight SSD code for 2 transmit antennas can be at most 1 complex symbol per channel use, which is also the rate of the well-known Alamouti code, which is single-real-symbol decodable and offers full diversity for all complex constellations. So, the unitary-weight SSD code for 2 transmit antennas offers no advantage compared to the Alamouti code. So, in the subsequent analysis, we only consider codes for transmit antennas, . Theorem 2: Any maximal rate, normalized unitary-weight SSD code must satisfy the following in addition to satisfying (6)–(12).

1) Let . We prove that this is not a possi, then bility. If it were true, i.e., should anticommute with and , as also seen in (8). The only matrices from the set that anticommute with and are and . So, let , with (13) where if and if Next, considering anticommute with and Further, since

( is odd) ( is even). , the only matrices from the set that and are . As a result, . , we have

, with as mentioned before. Since and are linearly independent over , either or . Suppose . Since anticommutes with , we see that cannot be . This is because both and commute with . Therefore, . By a similar argument, . Considering that and anticommute pairwise, we must have (14) (15) (16) From (14), (15), and (16), we obtain

Hence,

, which contradicts (13). So, cannot be anti-Hermitian. . In this case, we first look for possi2) Let bilities for . As argued before, either or , with if is even and if is odd. Assuming that , we have and

for

. Proof: Consider the following grouping of weight matrices: Since

can have two possibilities. Either .

,

or (17)

KARMAKAR et al.: MAXIMUM RATE OF UNITARY-WEIGHT, SINGLE-SYMBOL DECODABLE STBCS

with

Further, considering that and anticommute with each other, we have the following equalities:

7977

that the elements in the first row except are all mutually anticommuting matrices and all of them also anticommute with . So, and are pairwise anticommuting matrices. Hence, instead of , if we were to choose and as the anticommuting matrices to begin with, we would end up with the weight matrices as shown in the table at the top of the page. So, (19)

with as mentioned before. From the above set of equations, we obtain

So, if obtain,

and

, with if . So, from (17), we . By a Similar argument, we obtain, .

. This furFrom (18) and (19), ther implies that must be a unitary, Hermitian matrix, because of the choice of . So, the weight matrices of the normalized unitary-weight SSD code for transmit antennas are

Therefore, This completes the proof of the theorem. For 4 transmit antennas, by applying the procedure outlined in [4], we obtain the following pairwise anticommuting, antiHermitian matrices. It is easy to see that the above assignment of matrices violates the linear independence of the matrices over . Therefore, the assumption that is not valid. So, let

where,

Now,

the

only

possibility

is

that

. Similarly, by to , we assigning see that the conditions in (10), (11) and (12) are satisfied and from the discussion made above, this is the only assignment possible. Now, we only need to find a valid assignment for . Firstly, we note that (18) From Lemma 1, multiplying all the weight matrices by (i.e., ) will result in another unitary-weight SSD code with the weight matrices grouped as shown at the top of the page, after interchanging the first and the second columns. It should be noted in the above grouping scheme

For constructing a maximal rate, unitary weight SSD code for 4 transmit antennas, we only need 3 pairwise anticommuting, anti-Hermitian matrices. Hence, choosing and and applying the construction method described above, we obtain a maximal rate, unitary-weight SSD code for 4 transmit antennas, a codeword matrix of which is shown in (20), at the bottom of the next page. In general, for transmit antennas, we need unitary, antiHermitian, pairwise anticommuting matrices. If there are exactly pairwise anticommuting matrices of size , then, any matrix among them is a scaled product of the other . So, the following observations can be made about any maximal rate, normalized unitary-weight SSD code. 1) Either are among pairwise anticommuting matrices and are exactly pairwise anticommuting matrices, or are among pairwise

7978

IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 57, NO. 12, DECEMBER 2011

anticommuting matrices and are exactly pairwise anticommuting matrices. 2) is a Hermitian matrix and if are among pairwise anticommuting matrices, or if are among pairwise anticommuting matrices. V. DIVERSITY

AND

CODING GAIN SSD CODES

OF

where is unitary and is a diagonal matrix with the diagonal entries being . Therefore,

UNITARY-WEIGHT and

We have seen in Lemma 1 that the coding gain of a unitaryweight SSD codes does not change when normalized. In this section, we obtain a common expression for the coding gain of all unitary-weight SSD codes and identify the conditions on QAM constellations that will allow unitary weight SSD codes to have full transmit diversity and high coding gain. Let and be two distinct codewords of any normalized unitary-weight SSD code . Let

where,

with The minimum of the determinant, denoted by for all possible non-zero is given as Let

, of

and . Then, Since the expression inside the bracket in the above equation is a product of the sum of squares of real numbers, its minimum occurs when all but one among are zeros. So, (21) be the algebraic multiplicity of 1 as the eigenvalue of and be that of . We make use of the following lemma to conclude that .

Let

Since are and

is unitary and Hermitian, the eigenvalues of is unitarily diagonalizable. Let

,

and be unitary, Lemma 5: Let pairwise anticommuting matrices. Then, the product matrix

(20)

KARMAKAR et al.: MAXIMUM RATE OF UNITARY-WEIGHT, SINGLE-SYMBOL DECODABLE STBCS

, with the exception of , is traceless. Proof: It is well known that two matrices and . Let and anticommuting matrices. So,

for any be two invertible,

(22) Similarly, it can be shown that . By applying Lemma , 4, it can be seen that any product matrix anticommutes with some other product matrix from the set . Therefore, from the result obtained in (22), we can say that every product matrix except is traceless. Since is a scaled product of matrices among unitary, pairwise anticommuting matrices, is traceless. Hence, . So, (21) becomes (23) From the above expression, it is clear that for maximal-rate, unitary-weight SSD codes to offer full transmit diversity, the difference set , where is the constellation employed, should not have any points that lie on lines that are at degrees in the complex plane from the origin. Further, since the analysis leading up to the expression in (23) is not specific to any particular unitary-weight SSD code, we can infer that for any particular constellation , all maximal-rate, unitary-weight SSD codes have the same coding gain. A. Diversity, Coding Gain Calculations for QAM In this subsection we show that all maximal-rate, unitaryweight SSD codes have the same coding gain as the CIODs [6] for QAM constellations. Let , be the information symbols that take values from a constellation . Consider the following unitary rotation.

The above operation is equivalent to rotating to obtain . Then, from (23), we have

by

radians (24)

The above expression is the same as the one for CIOD, obtained in [6]. It is to be noted that (24) holds even when the angle of rotation is radians. In order to maximize , the minimum of the product , called the product distance, must be maximized. This has been done for QAM in [6], by rotating QAM constellations by an angle of . So, , should take values from a rotated QAM constellation, with the angle of rotation being . So, the

7979

TABLE I COMPARISON OF THE MINIMUM DETERMINANTS OF A FEW SSD CODES FOR 4 TRANSMIT ANTENNAS

original information symbols , should take values from a rotated QAM constellation, the angle of rotation being . Since the coding gain for CIOD has been maximized in [6] by using a radian rotated QAM constellation, the coding gain for all unitary-weight SSD codes when the symbols take values from a radian rotated QAM constellation is also maximized. Table I gives a comparison of the minimum determinants for the CIOD, MDCQOD and the unitary-weight SSD code presented in (20), all the codes designed for 4 transmit antennas. In the calculations, all the codes have the same average energy but the constellation energy has been allowed to increase with the increase in constellation size. As analytically shown, the minimum determinants are the same for all the three codes. B. Simulation Results for Our SSD Codes In this subsection, we provide some simulation results for 4 transmit antennas. All simulations are done assuming a quasi-static Rayleigh fading channel. The number of receive antennas is 1. Fig. 1 shows the codeword error rate (CER) performances of the CIOD for 4 transmit antennas, the MDCQOD for 4 transmit antennas [7], and the new design whose codeword matrix is as in (20), at 2 bits and 3 bits per channel use (bpcu). For transmission at 2 bpcu, the constellation employed is the radian rotated 4-QAM for the CIOD and the radian rotated 4-QAM for the MDCQOD and the new design. For transmission at 3 bpcu, the constellations employed are the rotated rectangular 8-QAM and the rotated square-derived 8-QAM, the angle of rotation being the same as in the case of 4-QAM. A squared-derived 8-QAM constellation is obtained by removing the signal point with the highest energy from a 9-QAM. Specifically, it is the set . The simulation results support the fact that for QAM constellations, the coding gain of the SSD codes is the same as that of the CIOD. VI. DISCUSSION AND CONCLUDING REMARKS In this paper, we have provided an achievable upper bound on the rate of unitary-weight SSD codes for transmit antennas. The upper bound has been shown to be complex symbols per channel use. We also have completely characterized the structure of the weight matrices of the codes meeting the upper bound. We have further shown that all unitary-weight SSD codes that meet the upper bound have the same coding gain as that of the CIODs and we have also identified the angle of rotation for QAM constellations that allow the codes to have optimum coding gain. The analysis done in this paper throws open the following questions.

7980

IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 57, NO. 12, DECEMBER 2011

Fig. 1. Comparison of the CER performance of the SSD codes and the CIOD.

1) What is the upper bound on the rate of square, non-unitaryweight SSD codes? Further, what are the conditions on the signal constellation that allow non-unitary-weight SSD codes to achieve full-diversity and optimum coding gain? 2) What is the upper bound on the rate of rectangular SSD codes, the class of which the rectangular MDCQODs presented in [7] and [8] are a subclass? Further, the analysis in this paper can be used to study the rates of multi-symbol decodable codes, the upper bounds of which has never been reported in literature. These questions provide some directions for future research. ACKNOWLEDGMENT This work was partly supported by the DRDO-IISc Program on Advanced Research in Mathematical Engineering and by the Council of Scientific & Industrial Research (CSIR), India, through Research Grant (22(0365)/04/EMR-II) to B. S. Rajan. We thank X.-G. Xia for sending the preprint of [7]. We would also like to thank the anonymous Reviewers for pointing out [22], which helped in simplifying the proof for the upper bound on the rate of unitary-weight SSD codes. REFERENCES [1] V. Tarokh, H. Jafarkhani, and A. R. Calderbank, “Space-time block codes from orthogonal designs,” IEEE Trans. Inf. Theory, vol. 45, no. 5, pp. 1456–1467, Jul. 1999.

[2] V. Tarokh, H. Jafarkhani, and A. R. Calderbank, “Correction to ‘Spacetime block codes from orthogonal designs’,” IEEE Trans. Inf. Theory, vol. 46, no. 1, p. 314, Jan. 2000. [3] B. Hassibi and B. Hochwald, “High-rate codes that are linear in space and time,” IEEE Trans. Inf. Theory, vol. 48, no. 7, pp. 1804–1824, Jul. 2002. [4] O. Tirkkonen and A. Hottinen, “Square-matrix embeddable space-time block codes for complex signal constellations,” IEEE Trans. Inf. Theory, vol. 48, no. 2, pp. 384–395, Feb. 2002. [5] X. B. Liang, “Orthogonal designs with maximal rates,” IEEE Trans. Inf. Theory, vol. 49, no. 10, pp. 2468–2503, Oct. 2003. [6] Md. Z. A. Khan and B. S. Rajan, “Single-symbol maximum-likelihood decodable linear STBCs,” IEEE Trans. Inf. Theory, vol. 52, no. 5, pp. 2062–2091, May 2006. [7] H. Wang, D. Wang, and X.-G. Xia, “On optimal quasi-orthogonal space-time block codes with minimum decoding complexity,” IEEE Trans. Inf. Theory, vol. 55, no. 3, pp. 1104–1130, Mar. 2009. [8] C. Yuen, Y. L. Guan, and T. T. Tjhung, “Quasi-orthogonal STBC with minimum decoding complexity,” IEEE Trans. Wireless Comm., vol. 4, no. 5, pp. 2089–2094, Sep. 2005. [9] S. M. Alamouti, “A simple transmit diversity scheme for wireless communications,” IEEE J. Sel. Areas Comm., vol. 16, no. 8, pp. 1451–1458, Oct. 1998. [10] V. Tarokh, N. Seshadri, and A. R. Calderbank, “Space time codes for high date rate wireless communication: Performance criterion and code construction,” IEEE Trans. Inf. Theory, vol. 44, pp. 744–765, Oct. 1998. [11] S. Karmakar and B. S. Rajan, “Multigroup-decodable STBCs from clifford algebras,” IEEE Trans. Inf. Theory, vol. 55, no. 01, pp. 223–231, Jan. 2009. [12] H. Jafarkhani, “A quasi-orthogonal space-time block code,” IEEE Trans. Commun., vol. 49, no. 1, pp. 1–4, Jan. 2001. [13] W. Su and X. G. Xia, “Signal constellations for QOSTBC with full diversity,” IEEE Trans. Inf. Theory, vol. 50, no. 10, pp. 2331–2347, Oct. 2004.

KARMAKAR et al.: MAXIMUM RATE OF UNITARY-WEIGHT, SINGLE-SYMBOL DECODABLE STBCS

[14] N. Sharma and C. B. Papadias, “Improved quasi-orthogonal codes through constellation rotation,” IEEE Trans. Commun., vol. 51, no. 3, pp. 332–335, 2003. [15] R. V. J. R. Doddi, V. Shashidhar, Md. Z. A. Khan, and B. S. Rajan, “Low-complexity, full-diversity space-time-frequency block codes for MIMO-OFDM,” in Proc. IEEE GLOBECOM, Dallas, TX, Nov. 29–Dec. 3 2004, pp. 204–208, Communication Theory Symposium. [16] S. Gowrisankar and B. S. Rajan, “A rate-one full-diversity low-complexity space-time-frequency block code (STFBC) for 4-Tx MIMOOFDM,” in Proc. IEEE Int. Symp. Inf. Theory (ISIT), Adelaide, Australia, Sept. 2–9, 2005, pp. 2090–2094. [17] T. Kiran and B. S. Rajan, “Distributed space-time codes with reduced decoding complexity,” in Proc. IEEE Int. Symp. Inf. Theory (ISIT 2006), Seattle, WA, Jul. 09–14, 2006, pp. 542–546. [18] T. Kiran and B. S. Rajan, “Partially-coherent distributed space-time codes with differential encoder and decoder,” in Proc. IEEE Int. Symp. Inf. Theory (ISIT), Seattle, WA, Jul. 09–14, 2006, pp. 547–551. [19] T. Kiran and B. S. Rajan, “Partially-coherent distributed space-time codes with differential encoder and decoder,” IEEE J. Sel. Areas Commun., vol. 25, no. 2, pp. 426–433, Feb. 2007. [20] G. S. Rajan and B. S. Rajan, “A non-orthogonal distributed space-time coded protocol, Part-I: Signal model and design criteria, Part-II: Code construction and DM-G tradeoff,” in Proc. IEEE Inf. Theory Workshop (ITW), Chengdu, China, Oct. 22–26, 2006. [21] Z. Yi and I.-M. Kim, “Single-symbol ML decodable distributed STBCs for cooperative networks,” IEEE Trans. Inf. Theory, vol. 53, no. 8, pp. 2977–2985, Aug. 2007. [22] D. B. Shapiro and R. Martin, “Anticommuting matrices,” Amer. Math. Monthly, vol. 105, no. 6, pp. 565–566, Jun.–Jul. 1998. Sanjay Karmakar (S’08) received the Bachelors degree in electronics and telecommunications engineering from Jadavpur University, Kolkata, India, in 2003 and the Masters degree in telecommunication engineering from the Indian Institute of Science, Bangalore, in 2006. He joined the University of Colorado at Boulder in Fall 2007, where he currently is working towards the Ph.D. degree in the electrical and computer science Department. From 2003 to 2004, he worked at the DRDO. He also worked at Beceem Communications for a year. His research interests include cooperative communication, information theory, and code design.

7981

K. Pavan Srinath received the Bachelor of Engg. degree in electronics and communication from B. M. Sreenivasiah College of Engg., Bangalore, India and the Master of Engg. degree in telecommunication from the Indian Institute of Science, Bangalore, in 2005 and 2008, respectively. From September 2005 to June 2006, he was with Robert Bosch India limited, Bangalore. Currently, he is a working towards his Ph.D. in the Department of Electrical Communication Engineering, Indian Institute of Science, Bangalore. His primary research interests include wireless communication, space-time coding and coding for wireless relay networks.

B. Sundar Rajan (S’84–M’91–SM’98) was born in Tamil Nadu, India. He received the B.Sc. degree in mathematics from Madras University, India, the B.Tech. degree in electronics from Madras Institute of Technology and the M.Tech. and Ph.D. degrees in electrical engineering from the Indian Institute of Technology, Kanpur, in 1979, 1982, 1984, and 1989 respectively. He was a faculty member with the Department of Electrical Engineering at the Indian Institute of Technology in Delhi, from 1990 to 1997. Since 1998, he has been a Professor in the Department of Electrical Communication Engineering at the Indian Institute of Science, Bangalore. His primary research interests include space-time coding for MIMO channels, distributed space-time coding and cooperative communication, coding for multiple-access, relay channels and network coding with emphasis on algebraic techniques. Dr. Rajan is an Associate Editor of the IEEE TRANSACTIONS ON INFORMATION THEORY, an Editor of the IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, and an Editorial Board Member of the International Journal of Information and Coding Theory. He served as a Technical Program Co-Chair of the IEEE Information Theory Workshop (ITW’02), held in Bangalore, in 2002. He is a Fellow of Indian National Academy of Engineering, a Fellow of the National Academy of Sciences, India, and a recipient of the IETE Pune Center’s S.V.C Aiya Award for Telecom Education in 2004. Dr. Rajan is a Member of the American Mathematical Society.

Maximum Rate of Unitary-Weight, Single-Symbol ... - IEEE Xplore

Dec 7, 2011 - is a power of 2. The rate of the square CODs for has been shown to be complex symbols per channel use. However, SSD codes having.

3MB Sizes 1 Downloads 241 Views

Recommend Documents

Optimized Software Implementation of a Full-Rate IEEE ... - IEEE Xplore
Hardware implementations are often used to meet the high-data- rate requirements of 802.11a standard. Although software based solutions are more attractive ...

Maximum principle for optimal control of sterilization of ... - IEEE Xplore
Feb 19, 2007 - BING SUN†. Department of Mathematics, Bohai University, Jinzhou, Liaoning 121000,. People's Republic of China. AND. MI-XIA WU. College of Applied Sciences, Beijing University of Technology, Beijing 100022,. People's Republic of China

Distributed Sum-Rate Maximization Over Finite Rate ... - IEEE Xplore
of a wired backhaul (typically an x-DSL line) to exchange control data to enable a local coordination with the aim of improving spectral efficiency. Since the backhaul is prone to random (un- predictable) delay and packet drop and the exchanged data

Rate-Dependent Hysteresis Modeling and Control of a ... - IEEE Xplore
a Piezostage Using Online Support Vector Machine and Relevance .... model and suppress the rate-dependent hysteretic behavior of a piezostage system, and ...

Effects of Heterogenous Mobility on Rate Adaptation ... - IEEE Xplore
rate adaptation and user scheduling (JRAUS) policy for cellular networks and compare it with the conventional and reference. JRAUS policies. We also evaluate ...

Blind Maximum Likelihood CFO Estimation for OFDM ... - IEEE Xplore
The authors are with the Department of Electrical and Computer En- gineering, National University of .... Finally, we fix. , and compare the two algorithms by ...

The Geometry of the MIMO Broadcast Channel Rate ... - IEEE Xplore
Telephone: +49 89 289-28508, Fax: +49 89 289-28504, Email: {hunger,joham}@tum.de ... dirty paper coding is applied, we show that the analogon to different ...

Failure Rate Modeling Using Equipment Inspection Data - IEEE Xplore
Page 1 ... of customizing failure rates using equipment inspection data. This ... that the incorporation of condition data leads to richer reliability models.

IEEE Photonics Technology - IEEE Xplore
Abstract—Due to the high beam divergence of standard laser diodes (LDs), these are not suitable for wavelength-selective feed- back without extra optical ...

Proto-Object Based Rate Control for JPEG2000: An ... - IEEE Xplore
Abstract—The JPEG2000 system provides scalability with respect to quality, resolution and color component in the transfer of images. However, scalability with ...

Ordered Statistics based rate allocation scheme for ... - IEEE Xplore
We propose a new rate allocation algorithm for closed loop MIMO-OFDM system. The new scheme utilizes or- dered statistics of channel matrix's singular value ...

Distributive Subband Allocation, Power and Rate ... - IEEE Xplore
in an orthogonal frequency-division multiple-access (OFDMA) cellular system with ... systems with imperfect CSIT in order to take into account of potential packet ...

Ad Hoc Networking With Rate-Limited Infrastructure ... - IEEE Xplore
Computer Science and Engineering. Dankook University. Yongin, Republic of Korea. Email: [email protected]. Abstract—Capacity scaling of a large hybrid ...

Proto-Object Based Rate Control for JPEG2000: An ... - IEEE Xplore
Constraints inherent in a modern visual data transmission system, such as heterogeneous network, varying connection quality, or the need to operate on a variety of devices with a wide range of capabilities, motivate an intense worldwide research effo

seam carving with rate-dependent seam path information - IEEE Xplore
7-1-2 Yoto, Utsunomiya, Tochigi, 321-8585 Japan email: {tanaka, madoka, kato}@is.utsunomiya-u.ac.jp. ABSTRACT. In this paper, we present a seam carving, ...

wright layout - IEEE Xplore
tive specifications for voice over asynchronous transfer mode (VoATM) [2], voice over IP. (VoIP), and voice over frame relay (VoFR) [3]. Much has been written ...

Device Ensembles - IEEE Xplore
Dec 2, 2004 - time, the computer and consumer electronics indus- tries are defining ... tered on data synchronization between desktops and personal digital ...

wright layout - IEEE Xplore
ACCEPTED FROM OPEN CALL. INTRODUCTION. Two trends motivate this article: first, the growth of telecommunications industry interest in the implementation ...

Evolutionary Computation, IEEE Transactions on - IEEE Xplore
search strategy to a great number of habitats and prey distributions. We propose to synthesize a similar search strategy for the massively multimodal problems of ...

I iJl! - IEEE Xplore
Email: [email protected]. Abstract: A ... consumptions are 8.3mA and 1.lmA for WCDMA mode .... 8.3mA from a 1.5V supply under WCDMA mode and.

Gigabit DSL - IEEE Xplore
(DSL) technology based on MIMO transmission methods finds that symmetric data rates of more than 1 Gbps are achievable over four twisted pairs (category 3) ...

IEEE CIS Social Media - IEEE Xplore
Feb 2, 2012 - interact (e.g., talk with microphones/ headsets, listen to presentations, ask questions, etc.) with other avatars virtu- ally located in the same ...

Grammatical evolution - Evolutionary Computation, IEEE ... - IEEE Xplore
definition are used in a genotype-to-phenotype mapping process to a program. ... evolutionary process on the actual programs, but rather on vari- able-length ...

SITAR - IEEE Xplore
SITAR: A Scalable Intrusion-Tolerant Architecture for Distributed Services. ∗. Feiyi Wang, Frank Jou. Advanced Network Research Group. MCNC. Research Triangle Park, NC. Email: {fwang2,jou}@mcnc.org. Fengmin Gong. Intrusion Detection Technology Divi