A Survey of Kurtosis Optimization Schemes for MISO Source Separation and Equalization

Abstract

Blind source separation and equalization aim at recovering a set of unknown source signals from their linearly distorted mixtures observed at a sensor array output, with little or no prior knowledge about the sources or the distorting channel. This fundamental signal processing problem arises in a broad range of applications such as multiuser digital communications, biomedical data analysis, and seismic exploration. Put forward over three decades ago, the normalized fourth-order cumulant, also known as kurtosis, has arguably become one of the most popular blind source separation and equalization criteria. Using multiple-input single-output (MISO) filter structures for single source extraction combined with suitable deflation procedures, the kurtosis contrast yields separation algorithms free of spurious extrema in ideal system conditions. The lack of closed-form solutions, however, calls for numerical optimization schemes. The present chapter reviews some of the iterative algorithms most widely used for MISO source separation and equalization based on kurtosis. These include gradient and Newton search methods, algorithms with optimal step-size selection, as well as techniques based on reference signals. Their main features are briefly summarized and their performance is illustrated by some numerical experiments in digital communications and biomedical signal processing.

References

  1. 1.
    Amari, S.: Natural gradient works efficiently in learning. Neural Comput. 10(2), 251–276 (1998) MathSciNetCrossRefGoogle Scholar
  2. 2.
    Amari, S., Cichocki, A.: Adaptive blind signal processing—neural network approaches. Proc. IEEE 86(10), 2026–2048 (1998) CrossRefGoogle Scholar
  3. 3.
    Belouchrani, A., Abed-Meraim, K., Cardoso, J.F., Moulines, E.: A blind source separation technique using second-order statistics. IEEE Trans. Signal Process. 45(2), 434–444 (1997) CrossRefGoogle Scholar
  4. 4.
    Benveniste, A., Métivier, M., Priouret, P.: Algorithmes Adaptatifs et Approximations Stochastiques. Masson, Paris (1987). English translation by S.S. Wilson, Adaptive Algorithms and Stochastic Approximations. Springer, Berlin (1990) Google Scholar
  5. 5.
    Bermejo, S.: Finite sample effects in higher order statistics contrast functions for sequential blind source separation. IEEE Signal Process. Lett. 12(6), 481–484 (2005) CrossRefGoogle Scholar
  6. 6.
    Bermejo, S.: Finite sample effects of the fast ICA algorithm. Neurocomputing 71(1–3), 392–399 (2007) CrossRefGoogle Scholar
  7. 7.
    Bingham, E., Hyvärinen, A.: A fast fixed-point algorithm for independent component analysis of complex valued signals. Int. J. Neural Syst. 10(1), 1–8 (2000) Google Scholar
  8. 8.
    Cardoso, J.F.: On the performance of orthogonal source separation algorithms. In: Proc. EUSIPCO-94, VII European Signal Processing Conference, Edinburgh, UK, pp. 776–779 (1994) Google Scholar
  9. 9.
    Cardoso, J.F., Laheld, B.H.: Equivariant adaptive source separation. IEEE Trans. Signal Process. 44(12), 3017–3030 (1996) CrossRefGoogle Scholar
  10. 10.
    Castella, M., Moreau, E.: Generalized identifiability conditions for blind convolutive MIMO separation. IEEE Trans. Signal Process. 57(7), 2846–2852 (2009) MathSciNetCrossRefGoogle Scholar
  11. 11.
    Castella, M., Moreau, E.: A new optimization method for reference-based quadratic contrast functions in a deflation scenario. In: Proc. IEEE Int. Conf. on Acoustics, Speech and Signal Processing (ICASSP), Taipei, Taiwan, R.O.C, pp. 3161–3164 (2009) Google Scholar
  12. 12.
    Castella, M., Moreau, E.: Reference based contrast functions in a semi-blind context. In: Proc. of ICA’09, 8th International Conference on Independent Component Analysis and Signal Separation, Paraty-RJ, Brazil. LNCS, vol. 5441, pp. 9–16 (2009) CrossRefGoogle Scholar
  13. 13.
    Castella, M., Moreau, E.: New kurtosis optimization schemes for MISO equalization. IEEE Trans. Signal Process. 60(3), 1319–1330 (2012) MathSciNetCrossRefGoogle Scholar
  14. 14.
    Castella, M., Bianchi, P., Chevreuil, A., Pesquet, J.C.: A blind source separation framework for detecting CPM sources mixed by a convolutive MIMO filter. Signal Process. 86(8), 1950–1967 (2006) MATHCrossRefGoogle Scholar
  15. 15.
    Castella, M., Chevreuil, A., Pesquet, J.C.: Mélanges convolutifs. In: Comon, P., Jutten, C. (eds.) Séparation de Sources, Tome 1: Concepts de Base et Analyse en Composantes Indépendantes, pp. 231–272. Hermès, Paris (2007). Chap. 7 Google Scholar
  16. 16.
    Castella, M., Rhioui, S., Moreau, E., Pesquet, J.C.: Quadratic higher-order criteria for iterative blind separation of a MIMO convolutive mixture of sources. IEEE Trans. Signal Process. 55(1), 218–232 (2007) MathSciNetCrossRefGoogle Scholar
  17. 17.
    Castella, M., Chevreuil, A., Pesquet, J.C.: Convolutive mixtures. In: Comon, P., Jutten, C. (eds.) Handbook of Blind Source Separation, Independent Component Analysis and Applications, pp. 281–324. Academic Press, New York (2010). Chap. 8 CrossRefGoogle Scholar
  18. 18.
    Castells, F., Rieta, J.J., Millet, J., Zarzoso, V.: Spatiotemporal blind source separation approach to atrial activity estimation in atrial tachyarrhythmias. IEEE Trans. Biomed. Eng. 52(2), 258–267 (2005) CrossRefGoogle Scholar
  19. 19.
    Comon, P.: Independent component analysis, a new concept? Signal Process. 36(3), 287–314 (1994). Special Issue on Higher-Order Statistics MATHCrossRefGoogle Scholar
  20. 20.
    Comon, P.: Independent component analysis, contrasts, and convolutive mixtures. In: Proc. 2nd IMA International Conference on Mathematics in Communications, Lancaster, UK, pp. 10–17 (2002) Google Scholar
  21. 21.
    Comon, P.: Contrasts, independent component analysis, and blind deconvolution. Int. J. Adapt. Control Signal Process. 18(3), 225–243 (2004). Special Issue on Blind Signal Separation MATHCrossRefGoogle Scholar
  22. 22.
    Comon, P., Jutten, C. (eds.): Handbook of Blind Source Separation, Independent Component Analysis and Applications. Academic Press, Oxford (2010) Google Scholar
  23. 23.
    Comon, P., Moreau, E.: Improved contrast dedicated to blind separation in communications. In: Proc. ICASSP-97, 22nd IEEE International Conference on Acoustics, Speech and Signal Processing, Munich, Germany, pp. 3453–3456 (1997) Google Scholar
  24. 24.
    De Lathauwer, L., Callaerts, D., De Moor, B., Vandewalle, J.: Fetal electrocardiogram extraction by source subspace separation. In: Proc. IEEE/ATHOS Signal Processing Conference on Higher-Order Statistics, Girona, Spain, pp. 134–138 (1995) Google Scholar
  25. 25.
    De Lathauwer, L., De Moor, B., Vandewalle, J.: Fetal electrocardiogram extraction by blind source subspace separation. IEEE Trans. Biomed. Eng. 47(5), 567–572 (2000). Special Topic Section on Advances in Statistical Signal Processing for Biomedicine CrossRefGoogle Scholar
  26. 26.
    Delfosse, N., Loubaton, P.: Adaptive blind separation of independent sources: a deflation approach. Signal Process. 45(1), 59–83 (1995) MATHCrossRefGoogle Scholar
  27. 27.
    Ding, Z., Nguyen, T.: Stationary points of a kurtosis maximization algorithm for blind signal separation and antenna beamforming. IEEE Trans. Signal Process. 48(6), 1587–1596 (2000) CrossRefGoogle Scholar
  28. 28.
    Donoho, D.: On minimum entropy deconvolution. In: Proc. 2nd Applied Time Series Analysis Symposium, Tulsa, OK, pp. 565–608 (1980) Google Scholar
  29. 29.
    Douglas, S.C.: On the convergence behavior of the FastICA algorithm. In: Proc. ICA-2003, 4th International Symposium on Independent Component Analysis and Blind Signal Separation, Nara, Japan, pp. 409–414 (2003) Google Scholar
  30. 30.
    Douglas, S.C.: Fixed-point algorithms for the blind separation of arbitrary complex-valued non-Gaussian signal mixtures. EURASIP J. Adv. Signal Process. (2007). doi:10.1155/2007/36525 Google Scholar
  31. 31.
    Dubroca, R., De Luigi, C., Castella, M., Moreau, E.: A general algebraic algorithm for blind extraction of one source in a MIMO convolutive mixture. IEEE Trans. Signal Process. 58(5), 2484–2493 (2010) MathSciNetCrossRefGoogle Scholar
  32. 32.
    Gesbert, D., Shafi, M., Shan-Shiu, D., Smith, P.J., Naguib, A.: From theory to practice: an overview of MIMO space-time coded wireless systems. IEEE J. Sel. Areas Commun. 21(3), 281–302 (2003) CrossRefGoogle Scholar
  33. 33.
    Godard, D.N.: Self-recovering equalization and carrier tracking in two-dimensional data communication systems. IEEE Trans. Commun. 28(11), 1867–1875 (1980) CrossRefGoogle Scholar
  34. 34.
    Hyvärinen, A.: One-unit contrast functions for independent component analysis: a statistical analysis. In: Proc. IEEE Neural Networks for Signal Processing Workshop, Amelia Island, FL, pp. 388–397 (1997) Google Scholar
  35. 35.
    Hyvärinen, A.: Fast and robust fixed-point algorithms for independent component analysis. IEEE Trans. Neural Netw. 10(3), 626–634 (1999) CrossRefGoogle Scholar
  36. 36.
    Hyvärinen, A., Oja, E.: A fast fixed-point algorithm for independent component analysis. Neural Comput. 9(7), 1483–1492 (1997) CrossRefGoogle Scholar
  37. 37.
    Johnson, C.R., Schniter, P., Fijalkow, I., Tong, L., Behm, J.D., et al.: The core of FSE-CMA behavior theory. In: Haykin, S.S. (ed.) Unsupervised Adaptive Filtering, Vol. II: Blind Deconvolution, pp. 13–112. Wiley, New York (2000). Chap. 2 Google Scholar
  38. 38.
    Li, H., Adali, T.: A class of complex ICA algorithms based on the kurtosis cost function. IEEE Trans. Neural Netw. 19(3), 408–420 (2008) CrossRefGoogle Scholar
  39. 39.
    Loubaton, P., Regalia, P.: Blind deconvolution of multivariate signals: a deflation approach. In: Proceedings of ICC, Geneva, Switzerland, pp. 1160–1164 (1993) Google Scholar
  40. 40.
    Moreau, E.: Criteria for complex sources separation. In: Proc. EUSIPCO-96, VII European Signal Processing Conference, Trieste, Italy, vol. II, pp. 931–934 (1996) Google Scholar
  41. 41.
    Moreau, E., Comon, P.: Contrasts. In: Comon, P., Jutten, C. (eds.) Handbook of Blind Source Separation, Independent Component Analysis and Applications, pp. 65–105. Academic Press, Oxford (2010). Chap. 3 CrossRefGoogle Scholar
  42. 42.
    Moreau, E., Macchi, O.: Complex self-adaptive algorithms for source separation based on high order contrasts. In: Proc. EUSIPCO-94, VII European Signal Processing Conference, Edinburgh, UK, pp. 1157–1160 (1994) Google Scholar
  43. 43.
    Moreau, E., Macchi, O.: A one stage self-adaptive algorithm for source separation. In: Proc. ICASSP-94, 19th IEEE International Conference on Acoustics, Speech and Signal Processing, Adelaide, Australia, vol. 3, pp. 49–52 (1994) Google Scholar
  44. 44.
    Moreau, E., Macchi, O.: High order contrasts for self-adaptive source separation. Int. J. Adapt. Control Signal Process. 10(1), 19–46 (1996) MathSciNetMATHCrossRefGoogle Scholar
  45. 45.
    Moreau, E., Pesquet, J.C.: Generalized contrasts for multichannel blind deconvolution of linear systems. IEEE Signal Process. Lett. 4(6), 182–183 (1997) CrossRefGoogle Scholar
  46. 46.
    Moreau, E., Thirion-Moreau, N.: Non symmetrical contrasts for sources separation. IEEE Trans. Signal Process. 47(8), 2241–2252 (1999) CrossRefGoogle Scholar
  47. 47.
    Moreau, E., Pesquet, J.C., Thirion-Moreau, N.: Convolutive blind signal separation based on asymmetrical contrast functions. IEEE Trans. Signal Process. 55(1), 356–371 (2007) MathSciNetCrossRefGoogle Scholar
  48. 48.
    Novey, M., Adali, T.: Complex ICA by negentropy maximization. IEEE Trans. Neural Netw. 19(4), 596–609 (2008) CrossRefGoogle Scholar
  49. 49.
    Novey, M., Adali, T.: On extending the complex FastICA algorithm to noncircular sources. IEEE Trans. Signal Process. 56(5), 2148–2154 (2008) MathSciNetCrossRefGoogle Scholar
  50. 50.
    Papadias, C.B.: Globally convergent blind source separation based on a multiuser kurtosis maximization criterion. IEEE Trans. Signal Process. 48(12), 3508–3519 (2000) MathSciNetCrossRefGoogle Scholar
  51. 51.
    Pesquet, J.C., Moreau, E.: Cumulant based independence measures for linear mixtures. IEEE Trans. Inf. Theory 47(5), 1947–1956 (2001) MathSciNetMATHCrossRefGoogle Scholar
  52. 52.
    Press, W.H., Teukolsky, S.A., Vetterling, W.T., Flannery, B.P.: Numerical Recipes in C. The Art of Scientific Computing, 2nd edn. Cambridge University Press, Cambridge (1992) MATHGoogle Scholar
  53. 53.
    Regalia, P.A.: A finite-interval constant modulus algorithm. In: Proc. ICASSP-2002, 27th International Conference on Acoustics, Speech and Signal Processing, Vol. III, Orlando, FL, pp. 2285–2288 (2002) Google Scholar
  54. 54.
    Rieta, J.J., Zarzoso, V., Millet-Roig, J., García-Civera, R., Ruiz-Granell, R.: Atrial activity extraction based on blind source separation as an alternative to QRST cancellation for atrial fibrillation analysis. In: Proc. Computers in Cardiology, Boston, MA, vol. 27, pp. 69–72 (2000) Google Scholar
  55. 55.
    Rieta, J.J., Castells, F., Sánchez, C., Zarzoso, V., Millet, J.: Atrial activity extraction for atrial fibrillation analysis using blind source separation. IEEE Trans. Biomed. Eng. 51(7), 1176–1186 (2004) CrossRefGoogle Scholar
  56. 56.
    Ristaniemi, T., Joutsensalo, J.: Advanced ICA-based receivers for block fading DS-CDMA channels. Signal Process. 82(3), 417–431 (2002) MATHCrossRefGoogle Scholar
  57. 57.
    Sato, Y.: A method of self-recovering equalization for multi-level amplitude modulation. IEEE Trans. Commun. 23, 679–682 (1975) CrossRefGoogle Scholar
  58. 58.
    Shalvi, O., Weinstein, E.: New criteria for blind deconvolution of nonminimum phase systems (channels). IEEE Trans. Inf. Theory 36(2), 312–321 (1990) MathSciNetMATHCrossRefGoogle Scholar
  59. 59.
    Simon, C., Loubaton, P., Jutten, C.: Separation of a class of convolutive mixtures: a contrast function approach. Signal Process. 4(81), 883–887 (2001) CrossRefGoogle Scholar
  60. 60.
    Stuart, A., Ord, K.: Kendall’s Advanced Theory of Statistics, vol. 1, 6th edn. Hodder Arnold, Sevenoaks (1994) Google Scholar
  61. 61.
    Thirion-Moreau, N., Moreau, E.: Generalized criteria for blind multivariate signal equalization. IEEE Signal Process. Lett. 9(2), 72–74 (2002) CrossRefGoogle Scholar
  62. 62.
    Tichavský, P., Koldovský, Z., Oja, E.: Performance analysis of the FastICA algorithm and Cramér-Rao bounds for linear independent component analysis. IEEE Trans. Signal Process. 54(4), 1189–1203 (2006) CrossRefGoogle Scholar
  63. 63.
    Treichler, J.R., Agee, B.G.: A new approach to multipath correction of constant modulus signals. IEEE Trans. Acoust. Speech Signal Process. 31(2), 459–472 (1983) CrossRefGoogle Scholar
  64. 64.
    Tugnait, J.K.: Identification and deconvolution of multichannel linear non-Gaussian processes using higher order statistics and inverse filter criteria. IEEE Trans. Signal Process. 45(3), 658–672 (1997) MathSciNetCrossRefGoogle Scholar
  65. 65.
    Widrow, B., Glover, J.R., McCool, J.M., et al.: Adaptive noise cancelling: principles and applications. Proc. IEEE 63(12), 1692–1716 (1975) CrossRefGoogle Scholar
  66. 66.
    Wiggins, R.A.: Minimum entropy deconvolution. Geoexploration 16, 21–35 (1978) CrossRefGoogle Scholar
  67. 67.
    Zarzoso, V.: Extraction of ECG characteristics using source separation techniques: exploiting statistical independence and beyond. In: Naït-Ali, A. (ed.) Advanced Biosignal Processing, pp. 15–47. Springer, Berlin (2009). Chap. 2 CrossRefGoogle Scholar
  68. 68.
    Zarzoso, V., Comon, P.: Blind and semi-blind equalization based on the constant power criterion. IEEE Trans. Signal Process. 53(11), 4363–4375 (2005) MathSciNetCrossRefGoogle Scholar
  69. 69.
    Zarzoso, V., Comon, P.: Blind channel equalization with algebraic optimal step size. In: Proc. EUSIPCO-2005, XIII European Signal Processing Conference, Antalya, Turkey (2005) Google Scholar
  70. 70.
    Zarzoso, V., Comon, P.: Semi-blind constant modulus equalization with optimal step size. In: Proc. ICASSP-2005, 30th International Conference on Acoustics, Speech and Signal Processing, Vol. III, Philadelphia, PA, pp. 577–580 (2005) CrossRefGoogle Scholar
  71. 71.
    Zarzoso, V., Comon, P.: Comparative speed analysis of FastICA. In: Proc. ICA-2007, 7th International Conference on Independent Component Analysis and Signal Separation, London, UK, pp. 293–300 (2007) CrossRefGoogle Scholar
  72. 72.
    Zarzoso, V., Comon, P.: Optimal step-size constant modulus algorithm. IEEE Trans. Commun. 56(1), 10–13 (2008) CrossRefGoogle Scholar
  73. 73.
    Zarzoso, V., Comon, P.: Automated extraction of atrial fibrillation activity from the surface ECG using independent component analysis in the frequency domain. In: Proc. Medical Physics and Biomedical Engineering World Congress, pp. 395–398 (2009). Invited Google Scholar
  74. 74.
    Zarzoso, V., Comon, P.: Robust independent component analysis by iterative maximization of the kurtosis contrast with algebraic optimal step size. IEEE Trans. Neural Netw. 21(2), 248–261 (2010) CrossRefGoogle Scholar
  75. 75.
    Zarzoso, V., Nandi, A.K.: Noninvasive fetal electrocardiogram extraction: blind separation versus adaptive noise cancellation. IEEE Trans. Biomed. Eng. 48(1), 12–18 (2001) CrossRefGoogle Scholar
  76. 76.
    Zarzoso, V., Comon, P., Kallel, M.: How fast is FastICA? In: Proc. EUSIPCO-2006, XIV European Signal Processing Conference, Florence, Italy (2006) Google Scholar
  77. 77.
    Zarzoso, V., Comon, P., Slock, D.: Semi-blind methods for communications. In: Comon, P., Jutten, C. (eds.) Handbook of Blind Source Separation, Independent Component Analysis and Applications, pp. 593–638. Academic Press, Oxford (2010). Chap. 15 CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2013

Authors and Affiliations

  1. 1.CNRS UMR 5157 SAMOVARInstitut Mines-Télécom/Télécom SudParisEvry CedexFrance
  2. 2.ISITV, LSIS UMR-CNRS 7296University of the South Toulon-VarLa Valette du Var CedexFrance
  3. 3.I3S LaboratoryUniversity of Nice Sophia Antipolis, CNRSSophia Antipolis CedexFrance

Personalised recommendations