Particle Filtering Approaches for Multiple Acoustic Source Detection and 2-D Direction of Arrival Estimation Using a Single Acoustic Vector Sensor

This paper considers the problem of tracking multiple acoustic sources using a single acoustic vector sensor (AVS). Firstly, a particle filtering (PF) approach is developed to track the direction of arrivals of fixed and known number of sources. Secondly, a more realistic tracking scenario which assumes that the number of acoustic sources is unknown and time-varying is considered. A random finite set (RFS) framework is employed to characterize the randomness of the state process, i.e., the dynamics of source motion and the number of active sources, as well as the measurement process. As deriving a closed-form solution for the multi-source probability density is difficult, a particle filtering approach is employed to arrive at a computationally tractable approximation of the RFS densities. The proposed RFS-PF algorithm is able to simultaneously detect and track multiple sources. Simulations under different tracking scenarios demonstrate the ability of the proposed approaches in tracking multiple acoustic sources.

I. Introduction

Detection, localization and tracking of 2-D (azimuth and elevation) direction of arrivals (DOA) of multiple acoustic sources in a noisy environment are important topics in signal processing and have many applications such as room speech enhancement, underwater target surveillance, sonar and acoustic radar signal processing. The tasks are traditionally performed by using an array equipped with several pressure sensors together with estimation techniques developed based on the acoustic pressure measurements [1], [2]. However, such techniques usually require either an array with large aperture or multiple hybrid arrays. In recent years, a new technology namely acoustic vector sensor (AVS) has been widely employed for acoustic source detection and localization, and different signal processing algorithms have been developed accordingly [3]–[25].