Title :
Oversampled filter banks from extended perfect reconstruction filter banks
Author :
Bernardini, Riccardo ; Rinaldo, Roberto
Author_Institution :
Dipt. di Ingegneria Elettrica, Univ. di Udine
fDate :
7/1/2006 12:00:00 AM
Abstract :
Oversampled filter banks are currently being proposed for robust transmission applications. In this paper, we completely characterize multidimensional doubly finite-impulse-response (FIR) filter banks, that is, oversampled filter banks whose dual is FIR. Then, we consider the problem of extending perfect reconstruction critically sampled multidimensional filter banks in order to obtain doubly FIR (DFIR) filter banks. As a result, very simple criteria for constructing DFIR filter banks as extensions of orthogonal filter banks are obtained. This paper also analyzes the problem of constructing totally FIR filter banks, i.e., DFIR filter banks that remain DFIR even when some channels are removed. It is shown that any totally FIR filter bank can be implemented as the cascade of a critically sampled DFIR filter bank whose number of channels is equal to the subsampling factor, a redundant finite-dimensional transform, and a suitable set of delays
Keywords :
FIR filters; channel bank filters; filtering theory; signal reconstruction; signal sampling; transforms; FIR filter banks; multidimensional doubly finite-impulse-response filters; oversampled filter bank reconstruction; redundant finite-dimensional transforms; subsampling factor; Channel bank filters; Error correction codes; Filter bank; Finite impulse response filter; MIMO; Multidimensional systems; Redundancy; Robustness; Signal analysis; Wireless sensor networks; Filtering; multiple-input multiple-output (MIMO) systems; source coding; transform coding;
Journal_Title :
Signal Processing, IEEE Transactions on
DOI :
10.1109/TSP.2006.874811