On the convergence of Constant Modulus Algorithm in Adaptive Equalization. Convergence des Algorithmes du Module Constant en Égalisation Adaptative

On the convergence of Constant Modulus Algorithm in Adaptive Equalization

Convergence des Algorithmes du Module Constant en Égalisation Adaptative

Maurice Bellanger

CNAM, 292 rue Saint-Martin, 75141 Paris cedex 03

Corresponding Author Email: 
bellang@cnam.fr
Page: 
269-275
|
Received: 
17 July 2007
|
Accepted: 
N/A
|
Published: 
31 August 2007
| Citation

OPEN ACCESS

Abstract: 

A direct approach is used to determine the conditions for the minimum mean-squared error and constant modulus criteria to lead to the same coefficient vector.The condition which is obtained,namely that the normalized kurtosis of the source signal be equal to 4/3,is nearly satisfied for the QAM constellations in digital transmission,which provides a theoretical justification for the Godard conjecture.Then,the performance of the adaptive algorithms is analyzed.It turns out  that the CMA(2,2) algorithm is faster than the CMA(1,2) algorithm but much slower than the LMS algorithm.Finally, the importance of initialization for the constant modulus algorithms is underlined.

Résumé

Par une approche directe,on établit les conditions pour que le critère de l’erreur quadratique moyenne minimale et les critères du module constant conduisent au même vecteur de coefficients. La condition de kurtosis normalisé égal à 4/3 obtenue est presque satisfaite pour les constellations MAQ utilisées en transmission numérique,ce qui fournit la justification théorique de la conjecture de Godard. Ensuite,les performances des algorithmes adaptatifs sont analysées. Il apparaît que l’algorithme AMC(2,2) est plus rapide que l’algorithme AMC(1,2) mais beaucoup plus lent que l’algorithme du gradient avec signal de référence. L’importance de l’initialisation est soulignée,pour les algorithmes du module constant.

Keywords: 

Constant modulus,self adaptive,equalization,convergence.

Mots clés

Module constant,adaptatif,égalisation,convergence.

1.Introduction
2.Conditions d’Équivalence entre EQMM et MC(2,2)
3.Convergence de l’Algorithme AMC(2,2)
4.Étude de l’Algorithme AMC(1,2)
5.Initialisation des Algorithmes du Module Constant
6.Conclusion
  References

[1] D.N. GODARD, «Self-recovering Equalization and CarrierTracking in Two-dimensional Data Communication Systems», IEEE Transactions on Com., Vol.COM-28, pp.1867-75, Nov.1980. 

[2] J.R. TREICHLER and B.G. AGEE, « A New Approach to Multipath Correction of Constant Modulus Signals», IEEE Trans. ASSP, vol.31, pp.459-72,April 1983. 

[3] I. FIJALKOW, A. TOUZNI and J.R. TREICHLER, «Fractionally Spaced Equalization Using CMA: Robustness to Channel Noise and Lack of Disparity»,IEEE Trans. on Signal Processing ,vol.45,pp.5666, Jan. 1997. 

[4] C.R. JOHNSON et al., «Blind Equalization Using the Constant Modulus Criterion: a Review», Proceedings of the IEEE, Vol.86, N°10, pp.1927-1950, 1998. 

[5] H.H. ZENG, L. TONG and C.R. JOHNSON, «An Analysis of Constant Modulus Receivers», IEEE Transactions on Signal Processing, Vol.47, N°11, Nov.1999, pp.2990-2999. 

[6] P. SCHNITER and C.R. JOHNSON, «Bounds for the MMSE performance of Constant Modulus Estimators», IEEE Trans. on Information Theory,, vol.46, N°7, pp.2544-60, Nov.2000.

[7] M. BELLANGER, «On the Performance of Two Constant Modulus Algorithms in Equalization with non-CM Signals», Proceedings of IEEE-ISCAS conference, New Orleans, May 2007. 

[8] J. MAI and A.H. SAYED, «A Feedback Approach to the Steady-State Performance of Fractionally Spaced Blind Adaptive Equalizers», IEEE trans. on Signal Processing,Vol.SP-48, N°1, January 2000, pp.80-91.

[9] H. JAMALI and T. OGUNFUNMI, «Stationary points of Finite Length Constant Modulus Optimization», Signal Processing, vol. 82, 2002, pp.625-641. 

[10] Y. LI and Z. DING, «Convergence Analysis Of Finite Length Blind Adaptive Equalizers», IEEE Trans. on Signal Processing, vol.43, N°9, pp.2120-29, Sept.1995.