Azimuth-Elevation Direction Finding With a Pair of Acoustic Vector Sensors in the Presence of a Reflecting Boundary

Abstract

This paper is aiming at addressing an azimuth-elevation direction-finding algorithm using a pair of identically oriented acoustic vector sensors located near a reflecting boundary. Two fourth-order cumulant matrices related in terms of a translationally invariant structure are formed for recovering the acoustic vector sensor steering vectors in the particle-velocity coarray domain. Afterward, source azimuth-elevation directions are extracted in closed-form from the estimates of the coarray steering vectors. The identifiability study shows that the algorithm proposed herein can uniquely resolve up to 13 sources, thereby being applicable to underdetermined scenarios, where the number of sources exceeds that of the sensors. This capability constitutes the primary advantage over the many seminal works Hawkes and Nehorai (2000); Wu et al. (2016); Ahmadi-Shokouh and Keshavarz (20007); Tao et al. (2007); Xu and Liu (2007) that pioneered AVS array processing involving reflecting boundaries. Finally, numerical examples are provided to verify the theoretical analysis and demonstrate the efficacy of the proposed algorithm.