Structure adaptation of linear MMSE adaptive filters

The tap-length, or the number of the taps, is an important structural parameter of the linear MMSE adaptive filter. Although the optimum tap-length that balances performance and complexity varies with scenarios, most current adaptive filters fix the tap-length at some compromise value, making them inefficient to implement especially in time-varying scenarios. A novel gradient search based variable tap-length algorithm is proposed, using the concept of the pseudo-fractional tap-length, and it is shown that the new algorithm can converge to the optimum tap-length in the mean. Results of computer simulations are also provided to verify the analysis.

[1]  Desmond P. Taylor,et al.  Blind channel order estimation based on second-order statistics , 2003, IEEE Signal Processing Letters.

[2]  Alan Weiss,et al.  Digital adaptive filters: Conditions for convergence, rates of convergence, effects of noise and errors arising from the implementation , 1979, IEEE Trans. Inf. Theory.

[3]  Jean Pierre Delmas,et al.  Blind channel approximation: effective channel order determination , 1998, Conference Record of Thirty-Second Asilomar Conference on Signals, Systems and Computers (Cat. No.98CH36284).

[4]  J. Nagumo,et al.  A learning method for system identification , 1967, IEEE Transactions on Automatic Control.

[5]  B. Farhang-Boroujeny,et al.  Adaptive Filters: Theory and Applications , 1999 .

[6]  D.G.M. Cruickshank,et al.  Linear equalisers, with dynamic and automatic length selection , 2001 .

[7]  S. Haykin,et al.  Adaptive Filter Theory , 1986 .

[8]  Iven M. Y. Mareels,et al.  LMS estimation via structural detection , 1998, IEEE Trans. Signal Process..

[9]  Huijuan Cui,et al.  Convergence analysis of a deficient-length LMS filter and optimal-length sequence to model exponential decay impulse response , 2003, IEEE Signal Process. Lett..

[10]  Yu Gong,et al.  Adaptive least mean square CDMA detection with gram-schmidt pre-processing , 2001 .

[11]  Bo Wahlberg,et al.  Improved LMS estimation via structural detection , 1995, Proceedings of 1995 IEEE International Symposium on Information Theory.

[12]  Bernard Widrow,et al.  Adaptive Signal Processing , 1985 .

[13]  Y. K. Won,et al.  Variable LMS Algorithms Using the Time Constant Concept , 1994, IEEE International Conference on Consumer Electronics.

[14]  L. Greenstein,et al.  Tap-selectable decision-feedback equalization , 1997, IEEE Trans. Commun..

[15]  Arie Feuer,et al.  Variable length stochastic gradient algorithm , 1991, IEEE Trans. Signal Process..