FAULT-TOLERANT PARALLEL FFTS ARE IMPLEMENTED BY USING ERROR CORRECTION CODES AND PARSEVAL CHECKS

Abstract

Modern electronic circuits are prone to reliability problems due to the occurrence of soft mistakes.
Consequently, there is a necessity for safeguarding against tiny errors across a wide range of applications. These
advancements also influence the systems responsible for signal processing and communications. The algorithmic
fault tolerance (ABFT) methodologies offer a compelling choice for some applications. The objective of these
methodologies is to detect and rectify faults by leveraging algorithmic traits for this purpose. The ABFT program
demonstrates exceptional performance in both signal processing and information transfer. The utilisation of quick
Fourier transforms, also referred to as FFTs, is essential in a wide range of diverse systems. Several protection
mechanisms for FFTs have been proposed to detect and rectify mistakes. The Parseval check, also referred to as
the sum of squares check, is widely recognised as one of the most famous among these few checks. Contemporary
communication networks are progressively defined by the existence of numerous components that function
simultaneously. This occurrence is increasingly prevalent. Recently, this fact has been utilised to establish a
method for including fault tolerance in parallel filters. In this overview, the primary application for this strategy
is safeguarding FFTs. Subsequently, two recommendations for improved security protocols are proposed and
evaluated. One of these methods utilises parseval checks, while the other employs error correction codes. Based
on the statistics, it appears that the proposed methods have the potential to further reduce the cost of protection.