A Perspective on Acoustic Vector Sensors in Passive Surveillance - Real World Measurements, Algorithms and Applications

Abstract

Acoustic signatures of battlefield or underwater sources can be passively exploited towards detecting, localizing and tracking hostile units. Acoustic vector sensors (AVS’s) have come to play an increasingly significant role in this technology with application focus on border control, harbor protection, gunshot localization, and situation awareness. As compared to sound pressure sensors, acoustic vector sensors have advantages of increased mathematical robustness, acoustic bandwidth, reduced system size, low data transmission between nodes, and set up times. These benefits have led to a steep increase in practical usefulness of AVSs with adoption of simple but powerful algorithms. Acoustic measurements in air have always been based upon sound pressure microphones. After the invention in 1994 of the Microflown sensor, capable of measuring directly the acoustic particle velocity in air, assembled AVS’s have become available since 2001 [1], [2], [3]. In 2006, the functionality of this assembled 3D acoustic vector sensor was incorporated on to one single monolithic chip, measuring the three vector components of the particle velocity and the sound pressure in one single point. The chip is extremely light weight and compact, and thus hardly visible [4]. Recently Microflown based AVS’s for underwater use are being developed. Passive monitoring methods to detect, track and classify objects without disclosing their own presence might benefit from these developments for at least two reasons. (1) The acoustic performance of vector based array measurements is superior to sheer sound pressure based beam forming techniques in terms of acoustic bandwidth and required channel count. (2) Product features like minuscule size, low weight and quick set up enable new ground to ground, ground to air and air to ground applications for passive monitoring. Concerning the acoustic properties, beam forming arrays have lower frequency limitations and a line (or plane) symmetry. Data from all measurement points have to be collected and processed first in order to obtain correct results. The Acoustic Vector Sensor (AVS) approach is broad banded, works 3D, and has a better mathematical robustness [14]. The ability of a single AVS to rapidly determine the bearing of a wideband acoustic source is of essence for numerous passive monitoring systems.Depending upon the classification of the acoustic problem, several AVS strategies can be implemented. The location of a transient gun shot can be localized using a pair of AVS’s, triangulating the sound source [16]. First measurements have proven that the range of the method exceeds 1 km in an open field for a 9 mm handgun.The geometric position of a single flying helicopter, considered as a single sound source with a strong tonal component, can also be triangulated with a pair of AVS’s [15], [16]. But the same geometric position can also be determined using one single AVS, using the vector information, the Doppler effect and the r-2 law [17]. There is mathematical evidence that a single AVS can determine the direction and strength of two statistical identical sources [14], [18]. In this paper this mathematical proof is supported by measurements. In order to locate a larger number of broad banded sound sources, a larger number of AVS can be used. Simulations show that a upper limit of (4n-2) independent sources can be found in 3D space where n is the number of AVS [18]. The mathematical background will be presented based on a MUSIC algorithm and first measurement results are presented.