A block floating-point realization of the gradient adaptive lattice filter
Authors:
- 21 March 2005
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Letters
- Vol. 12 (4) , 265-268
Abstract
We present a novel scheme to implement the gradient adaptive lattice (GAL) algorithm using block floating point (BFP) arithmetic that permits processing of data over a wide dynamic range at a cost significantly less than that of a floating point (FP) processor. Appropriate formats for the input data, the prediction errors, and the reflection coefficients are adopted, taking care so that for the prediction errors and the reflection coefficients, they remain invariant to the respective order and time update processes. Care is also taken to prevent overflow during prediction error computation and reflection coefficient updating by using an appropriate exponent assignment algorithm and an upper bound on the step-size mantissa.Keywords
This publication has 8 references indexed in Scilit:
- The NLMS Algorithm in Block Floating-Point FormatIEEE Signal Processing Letters, 2004
- Asymptotic behavior of digital filters with block floating point arithmetic2002
- Realization of block floating-point digital filters and application to block implementationsIEEE Transactions on Signal Processing, 1999
- Roundoff errors in block-floating-point systemsIEEE Transactions on Signal Processing, 1996
- Absolute error bounds for block floating-point direct-form digital filtersIEEE Transactions on Signal Processing, 1995
- Implementation of high-order direct- form digital filter structuresIEEE Transactions on Circuits and Systems, 1986
- Convergence properties of simplified gradient adaptive lattice algorithmsIEEE Transactions on Acoustics, Speech, and Signal Processing, 1985
- Block floating-point distributed filtersIEEE Transactions on Circuits and Systems, 1984