Fault tolerant Fermat number transform domain adaptive filters based on modulus replication RNS architectures

Authors:
C. Radhakrishnan;W. K. Jenkins;Z. Raza;R. M. Nickel
Affiliations:
Dept. of Elec. Engr., The Pennsylvania State University, University Park, PA;Dept. of Elec. Engr., The Pennsylvania State University, University Park, PA;Dept. of Elec. and Comp. Engr., George Washington University, Washington, DC;Dept. of Elec. Engr., Bucknell University, Lewisburg, PA
Venue:
Asilomar'09 Proceedings of the 43rd Asilomar conference on Signals, systems and computers
Year:
2009

Citing 5
Cited 0

On Modulus Replication for Residue Arithmetic Computations of Complex Inner Products

IEEE Transactions on Computers
Modeling the Effect of Technology Trends on the Soft Error Rate of Combinational Logic

DSN '02 Proceedings of the 2002 International Conference on Dependable Systems and Networks
Inventory based speech enhancement for speaker dedicated speech communication systems

ICASSP '09 Proceedings of the 2009 IEEE International Conference on Acoustics, Speech and Signal Processing
Fault tolerance in transform-domain adaptive filters operating with real-valued signals

IEEE Transactions on Circuits and Systems Part I: Regular Papers
Speech enhancement with inventory style speech resynthesis

IEEE Transactions on Audio, Speech, and Language Processing

Quantified Score

Hi-index	0.00

Visualization

Abstract

In previous work it was shown that a combination of Fermat Number Transform (FNT) domain block processing and polynomial ring theory provides computationally efficient fault tolerance for adaptive filters implemented in highly scaled nanotechnology circuits. Since the FNT requires finite field integer arithmetic it is compatible with Modulus Replication RNS (MRRNS) arithmetic that has powerful error detection/correction capabilities as well as the potential for low power realizations. This paper proposes a hybrid combination of Fermat Number Transform block processing and MRRNS coding that can be used to realize adaptive filters that are immune to a wide range of transient and hard errors.