A fast matched filter in time domain

Abstract

This paper presents a fast matched filter algorithm in time domain. Matched filters compute correlation of signals for optimal detection of signals in noise, for extracting ranging measurements, synchronisation, etc. We consider matched filters in the context of Global Positioning System (GPS) receivers, but the algorithms presented here can be used for other applications as no constraining assumptions are made in derivations. Our method outperforms conventional time domain method several times in arithmetic complexity. It is also competitive with transform domain techniques based on Fast Fourier Transform (FFT). Unlike FFT based methods our approach is accurate and is not using rounding and scaling operations.