Application of MUSIC, ESPRIT and SAGE Algorithms for Narrowband Signal Detection and Localization

Abstract

This study presents the experimental verification of multiple signal classification (MUSIC), estimation of signal parameters via rotational invariance techniques (ESPRIT) and space-alternating generalized expectation-maximization (SAGE) algorithms in estimating the direction-of-arrival (DoA) of a narrowband source. MUSIC and ESPRIT are subspace-based methods while SAGE is a maximum-likelihood technique. A BPSK signal was transmitted off-the-air and was received under varied system parameters. The three algorithms were then applied on the recorded data. The performance of the three algorithms was compared based on their capability to estimate the signal's general DoA. The results showed that SAGE has an average deviation in signal localization of 3.5deg, followed by MUSIC 3.8deg and ESPRIT 9.7deg. Further tests show that SAGE can closely estimate the signal DoA with lower data samples than MUSIC and ESPRIT. For subspace-based methods, a 4-element antenna array gives better estimates than a 2-element array