Evaluation of performance of BICM for high-order modulations was usually limited by the lack of formal description of the metrics used in the decision process. This problem was partially palliated by bounding techniques, e.g., [1] or the so-called expurgated bounds in [2]. However, these bounding techniques are either very loose, or when tight, they must be algorithmic.
To analyze the performance of any linear code under maximum likeli- hood (ML) decoding, union bound techniques are often used. The union bound (UB) on the BER for a convolutional code—which represents an upper bound on the BER of the code Pb—is given by [33, Sec. 4.4]
Pb≤ UB = ∞ X
d=dfree
βd· PEP(d), (23)
where dfree is the free distance of the code, βd is the weight distribution spectrum of the code, and PEP(d) is the pairwise error probability, which represents the probability of detecting a codeword with Hamming weight d instead of the transmitted all-zero codeword.
The weight distribution spectrum of the code βdis the number of bit er- rors (the information error weight) for error events of distance d. Although this concept was first used for convolutional codes, it is also possible to cal- culate the weight distribution spectrum of a turbo code using the uniform interleaver concept introduced in [53, 54]. Different algorithms for calculat- ing spectrums can be used, for example the recursive algorithm presented in [55].
For a symmetric and memoryless channel, the pairwise error probability in (23) can be written as the tail probability of a sum of d L-values in the nonzero path [19, 56], i.e.,
PEP(d) = Pr d X j=1 L(j) > 0 = Z ∞ 0 pΣd(λ) dλ, (24) where pΣ
d(λ) represents the PDF of the sum of d L-values. Since the L- values are assumed to be independent (due to the ideal interleaver), the PDF of their sum is a convolution of the individual PDFs. The pairwise
22 Introduction
error probability is then calculated as an integral of this convolution over the positive values of the argument λ.
Pairwise error probability calculations are straightforward for BPSK or QPSK, however, for high-order QAM modulations, they become nontrivial. Important contributions addressing this issue have been published recently by A. Martinez, A. Guill´en i F`abregas and G. Caire who presented in [57] a method to approximate the BISO BICM channel by a BPSK channel with scaled SNR. Another approach presented by the same authors to tackle this problem is the so-called saddlepoint approximation [19], which has also been used in [56] for computing bounds on the BER for BPSK over fading channels. The main drawback of these approaches is that both rely on numerical integration.
6
Purpose and Contributions
The purpose of this thesis is to develop methods to characterize the perfor- mance of BICM when high-order modulation schemes are used. The main contribution is the development of analytical expressions for the PDF of L- values for BICM and BICM-ID. Based on these results the BICM capacity is efficiently calculated and the coded performance is predicted using union bound techniques. Moreover, BICM with a packet retransmission scheme is analyzed, and the EXIT functions of the demapper for BICM-ID are also calculated.
The use of histograms (obtained through Monte-Carlo simulation) or other estimation methods (for example the cumulant method or the Gaus- sian mixture models) as estimates of the PDFs do not allow for an analytical treatment and are computationally costly alternatives when compared to the use of closed-form expressions. This provides the motivation to derive analytical expressions for the PDF.
The three contributions included in this thesis are summarized in the following list.
Paper A – On the Distribution of Extrinsic L-values in
Gray-mapped 16-QAM
In this paper the issue of the probabilistic modeling of the extrinsic L-values for BICM-ID is addressed. Starting with a simple piece-wise linear model of the L-values obtained via the max-log approximation, expressions for the cumulative distribution functions (CDFs) of the L-values are found. The de- sired forms of the PDFs for Gray-mapped 16-QAM are found differentiating the CDF. The developed analytical expressions are then used to efficiently compute the so-called EXIT functions of the demapper for different values
7 Future Work 23
of SNR. The proposed analytical expressions are also compared with the histograms of the L-values obtained through numerical simulations.
Paper B – Distribution of L-values in Gray-mapped M
2-
QAM: Closed-form Approximations and Applications
In this paper closed-form expressions for the PDFs of the L-values in BICM for QAM constellations with Gray mapping are developed. Based on these expressions, two simple Gaussian mixture approximations that are analyti- cally tractable are also proposed. The developments are used to efficiently calculate the BICM capacity, and to develop bounds on the coded bit-error rate when a convolutional code is used. The coded performance of an hybrid automatic repeat request (HARQ) based on constellation rearrangement is also evaluated.
Paper C – Distribution of Max-Log Metrics for QAM-
based BICM in Fading Channels
In this paper closed-form expressions for the PDFs of the L-values in BICM transmissions over fully-interleaved fading channels are derived. The expres- sions are valid for the relevant case of QAM schemes with Gray mapping when the metrics are calculated via the so-called max-log approximation. The results presented are particularized for a Rayleigh fading channel, how- ever, developments for the general case of a Nakagami-m case are also in- cluded. Using the developed expressions, the performance of BICM trans- missions using convolutional and turbo codes is efficiently evaluated, i.e., without resorting to otherwise required two-dimensional numerical integra- tion. The BICM capacity for different fading channels and constellation sizes is also evaluated.
7
Future Work
Some open issues related to the work presented in this thesis are described in the following list:
• It is still not clear which mappings are the ones which maximize the BICM capacity. As discussed previously, a Gray mapping was con- jectured to be the optimal, however, this has been been recently dis- proved. Consequently, an open problem is to find the optimum map- ping for any constellation size. Based on the results available in the literature, which are available only for small constellation sizes where it is still possible to do a full search, it seems that there is no unique optimum mapping. However, the BRGC is a good candidate since
24 Introduction
numerical results have shown that this mapping is the optimum for most of the analyzed cases and SNR values of interest.
• Although good mappings for BICM-ID are well-known in the liter- ature, there are no systematic methods to construct them. All the good mappings are found based on some algorithmic search which has obvious limitations when the constellation size increases. It is known that mappings for BICM-ID can be characterized by its EDS which is analogous to the distance spectrum of a convolutional code. In this sense, it would be possible to calculate the minimum free distance of the mapping for a given constellation size, and eventually come up with a method to systematically construct good mappings based on optimum EDS.
• It is well-known that the optimum BICM-ID design can be achieved if the code and the mapping are jointly designed. In the literature however, this approach is considered too difficult, and in general the design of mappings is tackled for a given code and vice versa. More- over, the interleaver design in this case is also crucial, but in most of the cases is completely ignored.
• Since the binary mapping of high-order modulation schemes produces unequal error protection, the code and the interleaver can be designed to take into account this property. With a proper selection of code and interleaver design, a performance improvement should be expected when compared to the standard codes and the (pseudo)random inter- leaver used up to now.
8
Related Contributions
Other related publications by the author, which are not in included in this thesis, are:
• “On Adaptive BICM with Finite Block-Length and Simplified Metrics Cal- culation”, A. Alvarado, H. Carrasco, and R. Feick, IEEE Vehicular Tech- nology Conference 2006, VTC-2006 Fall, Montreal, Canada, Sep. 2006. • “Distribution of L-values in Gray-mapped M2
-QAM Signals: Exact Expres- sions and Simple Approximations”, A. Alvarado, L. Szczecinski, R. Feick, and L. Ahumada, IEEE Global Telecommunications Conference GLOBE- COM 2007, Washington, USA, Nov. 2007.
• “Probability Density Functions of Reliability Metrics for 16-QAM-Based BICM Transmission in Rayleigh Channel”, L. Szczecinski, A. Alvarado, and R. Feick, IEEE International Conference on Communications, ICC 2007, Glasgow, UK, June 2007.
References 25
• “Closed-form approximation of Coded BER in QAM-based BICM Faded Transmission”, L. Szczecinski, A. Alvarado, and R. Feick, IEEE Sarnoff Symposium 2008, Princeton, NJ, USA, Apr. 2008.
References
[1] E. Zehavi, “8-PSK trellis codes for a Rayleigh channel,” IEEE Trans. Com- mun., vol. 40, no. 3, pp. 873–884, May 1992.
[2] G. Caire, G. Taricco, and E. Biglieri, “Bit-interleaved coded modulation,” IEEE Trans. Inf. Theory, vol. 44, no. 3, pp. 927–946, May 1998.
[3] A. Goldsmith, Wireless Communications. New York, NY: Cambridge Uni- versity Press, 2005.
[4] G. Ungerboeck, “Channel coding with multilevel/phase signals,” IEEE Trans. Inf. Theory, vol. 28, no. 1, pp. 55–67, Jan. 1982.
[5] IEEE 802.11, “Wireless LAN medium access control (MAC) and physical layer (PHY) specifications: High-speed physical layer in the 5GHz band,” IEEE Std 802.11a-1999(R2003), Tech. Rep., Jul. 1999.
[6] I. Koffman and V. Roman, “Broadband wireless access solutions based on OFDM access in IEEE 802.16,” IEEE Commun. Mag., vol. 40, no. 4, pp. 96–103, Apr 2002.
[7] A. Batra et al., “Multi-band OFDM physical layer proposal for IEEE 802.15 task group 3a,” Mar. 2004, Document IEEE P802.15-03/268r3, available at http://grouper.ieee.org/groups/802/15/.
[8] T. Lestable et al., “D2.2.3 modulation and coding schemes for the WINNER II system,” WINNER II, Tech. Rep. IST-4-027756, November 2007, available at https://www.ist-winner.org.
[9] E. Biglieri, “Coding and modulation for a horrible channel,” IEEE Commun. Mag., vol. 45, no. 5, pp. 92–98, May 2003.
[10] F. Schreckenbach, “Iterative decoding of bit-interleaved coded modulation,” Ph.D. dissertation, Munich University of Technology, Munich, Germany, 2007.
[11] P. ¨Ormeci, X. Liu, D. L. Goeckel, and R. D. Wesel, “Adaptive bit-interleaved coded modulation,” IEEE Trans. Commun., vol. 49, no. 9, pp. 1572–1581, Sep. 2001.
[12] X. Li and J. A. Ritcey, “Bit-interleaved coded modulation with iterative decoding,” IEEE Commun. Lett., vol. 1, no. 6, pp. 169–171, Nov. 1997. [13] S. ten Brink, J. Speidel, and R.-H. Yan, “Iterative demapping for QPSK
modulation,” IEE Electronics Letters, vol. 34, no. 15, pp. 1459–1460, Jul. 1998.
[14] A. Chindapol and J. A. Ritcey, “Design, analysis, and performance evaluation for BICM-ID with square QAM constellations in Rayleigh fading channels,” IEEE J. Sel. Areas Commun., vol. 19, no. 5, pp. 944–957, May 2001.
26 Introduction
[15] M. T¨uchler, “Design of serially concatenated systems depending on the block length,” IEEE Trans. Commun., vol. 52, no. 2, pp. 209–218, Feb. 2004. [16] F. Schreckenbach, N. G¨ortz, J. Hagenauer, and G. Bauch, “Optimization of
symbol mappings for bit-interleaved coded modulation with iterative decod- ing,” IEEE Commun. Lett., vol. 7, no. 12, pp. 593–595, Dec. 2003.
[17] L. Szczecinski, H. Chafnaji, and C. Hermosilla, “Modulation doping for itera- tive demapping of bit-interleaved coded modulation,” IEEE Commun. Lett., vol. 9, no. 12, pp. 1031–1033, Dec. 2005.
[18] X. Li, A. Chindapol, and J. A. Ritcey, “Bit-interlaved coded modulation with iterative decoding and 8PSK signaling,” IEEE Trans. Commun., vol. 50, no. 6, pp. 1250–1257, Aug. 2002.
[19] A. Martinez, A. Guill´en i F`abregas, and G. Caire, “Error probability analysis of bit-interleaved coded modulation,” IEEE Trans. Inf. Theory, vol. 52, no. 1, pp. 262–271, Jan. 2006.
[20] C. Berrou, A. Glavieux, and P. Thitimajshima, “Near Shannon limit error- correcting coding and decoding: Turbo codes,” in International Conference on Communications, ICC 1993, Geneva, Switzerland, May 1993, pp. 1064– 1070.
[21] A. J. Viterbi, “An intuitive justification and a simplified implementation of the MAP decoder for convolutional codes,” IEEE J. Sel. Areas Commun., vol. 16, no. 2, pp. 260–264, Feb. 1998.
[22] M. K. Simon and R. Annavajjala, “On the optimality of bit detection of certain digital modulations,” IEEE Trans. Commun., vol. 53, no. 2, pp. 299– 307, Feb. 2005.
[23] B. Classon, K. Blankenship, and V. Desai, “Channel coding for 4G sys- tems with adaptive modulation and coding,” IEEE Wireless Commun. Mag., vol. 9, no. 2, pp. 8–13, Apr. 2002.
[24] A. Alvarado, H. Carrasco, and R. Feick, “On adaptive BICM with finite block-length and simplified metrics calculation,” in IEEE Vehicular Technol- ogy Conference 2006, VTC-2006 Fall, Montreal, Canada, Sep. 2006. [25] Ericsson, Motorola, and Nokia, “Link evaluation methods for high speed
downlink packet access (HSDPA),” TSG-RAN Working Group 1 Meeting #15, TSGR1#15(00)1093, Tech. Rep., Aug. 2000.
[26] K. Hyun and D. Yoon, “Bit metric generation for Gray coded QAM signals,” IEE Proc.-Commun., vol. 152, no. 6, pp. 1134–1138, Dec. 2005.
[27] M. S. Raju, R. Annavajjala, and A. Chockalingam, “BER analysis of QAM on fading channels with transmit diversity,” IEEE Trans. Wireless Commun., vol. 5, no. 3, pp. 481–486, Mar. 2006.
[28] S. Lin and J. D. J. Costello, Error Control Coding, 2nd ed. Upper Saddle River, New Jersey 07458, USA: Prentice-Hall, Inc., 2003.
[29] R. H. Morelos-Zaragoza, The Art of Error Correcting Codes, 1st ed. New York, NY: John Wiley & Sons, 2002.
References 27
[30] P. Robertson, E. Villebrun, and P. Hoeher, “A comparison of optimal and sub-optimal MAP decoding algorithms operating in the log domain,” in IEEE International Conf. on Commun., ICC 95, Jun. 1995, pp. 1009–1013. [31] M. Valenti and J. Sun, “The UMTS turbo code and an efficient decoder
implementation suitable for software defined radios,” International Journal on Wireless Information Networks, vol. 8, no. 4, pp. 203–216, Oct. 2001. [32] W. Gross and P. Gulak, “Simplified MAP algorithm suitable for implemen-
tation of turbo decoders,” Electronics Letters, vol. 34, no. 16, pp. 1577–1578, Aug. 1998.
[33] A. J. Viterbi and J. K. Omura, Principles of Digital Communications and Coding. McGraw-Hill, 1979.
[34] C. E. Shannon, “A mathematical theory of communications,” Bell System Technical Journal, vol. 27, pp. 379–423 and 623–656, July and Oct. 1948. [35] T. Cover and J. Thomas, Elements of Information Theory. New York, USA:
Wiley series in telecommunications, John Wiley & Sons, 1991.
[36] C. Stierstorfer and R. Fischer, “Adaptive interleaving for bit-interleaved coded modulation,” in 7th International ITG Conference on Source and Channel Coding (SCC), Ulm, Germany, Jan. 2008.
[37] A. Abedi, E. Thompson, and A. K. Khandani, “Application of cumulant method in performance evaluation of turbo-like codes,” IEEE Trans. Com- mun., vol. 55, no. 11, pp. 2037–2041, Nov. 2007.
[38] R. Redner and H. Walker, “Mixture densities, maximum likelihood and the EM algorithm,” SIAM Review, vol. 26, pp. 195–239, Apr. 1994.
[39] J. Lindblom and J. Samuelsson, “Bounded support Gaussian mixture mod- eling of speech spectra,” IEEE Trans. Speech Audio Process., vol. 11, no. 1, pp. 88–99, Jan. 2003.
[40] S. ten Brink, “Convergence behaviour of iteratively decoded parallel concate- nated codes,” IEEE Trans. Commun., vol. 49, no. 10, pp. 1727–1737, Oct. 2001.
[41] S. ten Brink, M. T¨uchler, and J. Hagenahuer, “Measures for tracing conver- gence of iterative decoding algorithms,” in 4th International ITG Conference on Source and Channel Coding (SCC), Berlin, Germany, Jan. 2002.
[42] J. Tan and G. St¨uber, “Analysis and design of interleaver mappings for iter- atively decoded BICM,” in IEEE International Conference on Communica- tions, ICC 2002, New York, USA, May 2002.
[43] L. Zhao, L. Lampe, and J. Huber, “Study of bit-interleaved coded space-time modulation with different labeling,” in IEEE Information Theory Workshop, Paris, France, Mar. 2003, pp. 199–202.
[44] T. Clevorn, S. Godtmann, and P. Vary, “Optimized mappings for iteratively decoded BICM on rayleigh channels with IQ interleaving,” in IEEE Vehicular Technology Conference 2006, VTC-2006 Spring, vol. 5, Melbourne, Australia, May 2006, pp. 2083–2087.
28 Introduction
[45] J. Tan and G. St¨uber, “Analysis and design of symbol mappers for iteratively decoded BICM,” IEEE Trans. Wireless Commun., vol. 4, no. 2, pp. 662–672, Mar. 2005.
[46] E. Agrell, J. Lassing, E. G. Str¨om, and T. Ottosson, “On the optimality of the binary reflected Gray code,” IEEE Trans. Inf. Theory, vol. 50, no. 12, pp. 3170–3182, Dec. 2004.
[47] E. Agrell, J. Lassing, E. G. Str¨om, and T. Ottosson, “Gray coding for mul- tilevel constellations in Gaussian noise,” IEEE Trans. Inf. Theory, vol. 53, no. 1, pp. 224–235, Jan. 2007.
[48] C. Stierstorfer and R. Fischer, “(Gray) Mappings for bit-interleaved coded modulation,” in IEEE Vehicular Technology Conference 2007, VTC-2007 Spring, Dublin, Ireland, Apr. 2007.
[49] F. Schreckenbach and G. Bauch, “Bit-interleaved coded irregular modula- tion,” Eur. Trans. on Telecommun., vol. 17, no. 2, pp. 269–282, mar.-apr. 2006.
[50] M. Benjillali, L. Szczecinski, and S. Aissa, “Probability density functions of logarithmic likelihood ratios in rectangular QAM,” in Twenty-Third Biennial Symposium on Communications, Kingston, Canada, May 2006, pp. 283–286. [51] L. Szczecinski and M. Benjillali, “Probability density functions of logarithmic likelihood ratios in phase shift keying BICM,” in IEEE Global Telecommuni- cations Conference GLOBECOM 2006, San Francisco, USA, Nov. 2006. [52] L. Szczecinski, R. Bettancourt, and R. Feick, “Probability density func-
tion of reliability metrics in BICM with arbitrary modulation: Closed-form through algorithmic approach,” in IEEE Global Telecommunications Confer- ence GLOBECOM 2006, San Francisco, USA, Nov. 2006.
[53] S. Benedetto and G. Montorsi, “Average performance of parallel concate- nated block codes,” Electronics Letters, vol. 31, no. 3, pp. 156–158, Feb. 1995.
[54] ——, “Unveiling turbo codes: Some results on parallel concatenated coding schemes,” IEEE Trans. Inf. Theory, vol. 42, no. 2, pp. 409–428, Mar. 1996. [55] D. Divsalar, S. Dolinar, R. J. McEliece, and F. Pollara, “Transfer function
bounds on the performance of turbo codes,” JPL, Cal. Tech., TDA Progr. Rep. 42-121, Aug. 1995.
[56] A. Martinez, A. Guill´en i F`abregas, and G. Caire, “A closed-form approx- imation for the error probability of BPSK fading channels,” IEEE Trans. Wireless Commun., vol. 6, no. 6, pp. 2051–2054, Jun. 2007.
[57] A. Guill´en i F`abregas, A. Martinez, and G. Caire, “Error probability of bit- interleaved coded modulation using the Gaussian approximation,” in 38th Conference on Information Sciences and Systems CISS 2004, Princeton Uni- versity, NJ, USA, Mar. 2004.
Part II
Paper A
On the Distribution of Extrinsic L-values in
Gray-mapped 16-QAM
Alex Alvarado, Leszek Szczecinski, and Rodolfo Feick
Published in Proceedings of the 2007 International Conference on Wireless Communications and Mobile Computing, IWCMC’07 ISBN:978-1-59593-695-0, pp. 329–336 Honolulu, Hawaii, USA
c
Paper B
Distribution of L-values in Gray-mapped M
2-QAM:
Closed-form Approximations and Applications
Alex Alvarado, Leszek Szczecinski, Rodolfo Feick, and Luciano Ahumada
Accepted for publication in IEEE Transactions on Communications.
Paper C
Distribution of Max-Log Metrics for QAM-based
BICM in Faded Channels
Leszek Szczecinski, Alex Alvarado, and Rodolfo Feick
Accepted for publication in IEEE Transactions on Communications.