On the use of acoustic particle velocity fields in adjoint-based inversion

Following the recent interest in the use of combined pressure and particle motion sensors in underwater acoustics and signal processing, some general aspects regarding the modeling and multipath phenomenology of acoustic particle velocity fields in shallow water environments have been studied. In this paper we will address a number of issues associated with the incorporation of vector sensor data (pressure and particle velocity) into adjoint-based inversion schemes. Specifically, we will discuss the ability of a semi-automatic adjoint approach to compute the necessary gradient information without the need for an analytic model of the adjoint particle velocity field. Solutions to the forward propagation of acoustic pressure are computed using an implicit finite-difference parabolic equation solver while the particle velocity is calculated locally at each grid point. Some numerical examples of vector sensor inversion results are provided