Exponentially-modulated filter bank transmultiplexer with fine-coarse adaptive filtering

In this paper we explore filter bank based multicarrier transmission, which has received increasing interest also in wireless communications system development as an alternative to the DFT-based OFDM technique. The odd-stacked low delay exponentially-modulated filter bank (EMFB) transmultiplexer is described. The subfilter biorthogonality condition is distorted in practical transmission channels and we use oversampled (by factor of two) per-subband equalization to recover the data symbols. The fine-coarse normalized least mean square (LMS) algorithm is applied with EMFB-specific modifications. The adaptation speed of the fine-subequalizers are further improved with simple re-initialization procedure. Numerical studies exhibit fast convergence and relatively small residual error in semiblind, decision-directed operation mode.

[1]  Woo-Jin Song,et al.  A complementary pair LMS algorithm for adaptive filtering , 1997, 1997 IEEE International Conference on Acoustics, Speech, and Signal Processing.

[2]  J.E. Mazo,et al.  Digital communications , 1985, Proceedings of the IEEE.

[3]  M. Renfors,et al.  Efficient Implementation of 2$times$Oversampled Exponentially Modulated Filter Banks , 2006, IEEE Transactions on Circuits and Systems II: Express Briefs.

[4]  M. Renfors,et al.  Prototype filter design for arbitrary delay exponentially-modulated filter bank , 2008, 2008 3rd International Symposium on Communications, Control and Signal Processing.

[5]  Tanja Karp,et al.  Modified DFT filter banks with perfect reconstruction , 1999 .