Andreas M. Ali

An empirical study of collaborative acoustic source localization

Field biologists use animal sounds to discover the presence of individuals and to study their behavior. Collecting bio- acoustic data has traditionally been a difficult and time- consuming process in which researchers use portable microphones to record sounds while taking notes of their own detailed observations. The recent development of new deploy- able acoustic sensor platforms presents opportunities to develop automated tools for bio-acoustic field research. In this work, we implement an AML-based source localization algorithm, and use it to localize marmot alarm-calls. We assess the performance of these techniques based on results from two field experiments: (1) a controlled test of direction-of- arrival (DOA) accuracy using a pre-recorded source signal, and (2) an experiment to detect and localize actual animals in their habitat, with a comparison to ground truth gathered from human observations. Although small arrays yield ambiguities from spatial aliasing of high frequency signals, we show that these ambiguities are readily eliminated by proper bearing crossings of the DOAs from several arrays. These results show that the AML source localization algorithm can be used to localize actual animals in their natural habitat, using a platform that is practical to deploy.

An Empirical Study of Collaborative Acoustic Source Localization

Field biologists use animal sounds to discover the presence of individuals and to study their behavior. Collecting bio-acoustic data has traditionally been a difficult and time-consuming process in which researchers use portable microphones to record sounds while taking notes of their own detailed observations. The recent development of new deployable acoustic sensor platforms presents opportunities to develop automated tools for bio-acoustic field research. In this work, we implement both two-dimensional (2D) and three-dimensional (3D) AML-based source localization algorithms. The 2D algorithm is used to localize marmot alarm-calls of marmots on the meadow ground. The 3D algorithm is used to localize the song of Acorn Woodpecker and Mexican Antthrush birds situated above the ground. We assess the performance of these techniques based on the results from four field experiments: two controlled test of direction-of-arrival (DOA) accuracy using a pre-recorded source signal for 2D and 3D analysis, an experiment to detect and localize actual animals in their habitat, with a comparison to ground truth gathered from human observations, and a controlled test of localization experiment using pre-recorded source to enable careful ground truth measurements. Although small arrays yield ambiguities from spatial aliasing of high frequency signals, we show that these ambiguities are readily eliminated by proper bearing crossings of the DOAs from several arrays. These results show that the AML source localization algorithm can be used to localize actual animals in their natural habitat using a platform that is practical to deploy.

Acoustic sensor networks for woodpecker localization

Sensor network technology can revolutionize the study of animal ecology by providing a means of non-intrusive, simultaneous monitoring of interaction among multiple animals. In this paper, we investigate design, analysis, and testing of acoustic arrays for localizing acorn woodpeckers using their vocalizations. Each acoustic array consists of four microphones arranged in a square. All four audio channels within the same acoustic array are finely synchronized within a few micro seconds. We apply the approximate maximum likelihood (AML) method to synchronized audio channels of each acoustic array for estimating the direction-of-arrival (DOA) of woodpecker vocalizations. The woodpecker location is estimated by applying least square (LS) methods to DOA bearing crossings of multiple acoustic arrays. We have revealed the critical relation between microphone spacing of acoustic arrays and robustness of beamforming of woodpecker vocalizations. Woodpecker localization experiments using robust array element spacing in different types of environments are conducted and compared. Practical issues about calibration of acoustic array orientation are also discussed.

Theoretical and experimental study of DOA estimation using AML algorithm for an isotropic and non-isotropic 3D array

The focus of most direction-of-arrival (DOA) estimation problems has been based mainly on a two-dimensional (2D) scenario where we only need to estimate the azimuth angle. But in various practical situations we have to deal with a three-dimensional scenario. The importance of being able to estimate both azimuth and elevation angles with high accuracy and low complexity is of interest. We present the theoretical and the practical issues of DOA estimation using the Approximate-Maximum-Likelihood (AML) algorithm in a 3D scenario. We show that the performance of the proposed 3D AML algorithm converges to the Cramer-Rao Bound. We use the concept of an isotropic array to reduce the complexity of the proposed algorithm by advocating a decoupled 3D version. We also explore a modified version of the decoupled 3D AML algorithm which can be used for DOA estimation with non-isotropic arrays. Various numerical results are presented. We use two acoustic arrays each consisting of 8 microphones to do some field measurements. The processing of the measured data from the acoustic arrays for different azimuth and elevation angles confirms the effectiveness of the proposed methods.

Particle Filtering Approach to Localization and Tracking of a Moving Acoustic Source in a Reverberant Room

We propose a novel algorithm employing particle filters for acoustic source tracking in a reverberant environment. By incorporating the likelihood function computed through approximate maximum-likelihood (AML) method, the proposed algorithm is applicable to wideband sources and can be implemented for multiple sources tracking. Both computer simulation and experimental results show the effectiveness of the proposed algorithm

Acoustic source localization using the acoustic ENSBox

Field biologists use animal sounds to discover the presence of individuals and to study their behavior. The recent development of new deployable acoustic sensor platforms presents opportunities to develop automated tools for bio-acoustic field research. In this work, we demonstrate a real-time wireless sensor system that implements an AML-based source localization algorithm. We will demonstrate a system that is easy to set up and that can localize a whistle in the demonstration area in real time. This demonstration will use the techniques we described in our corresponding paper [1].

Tracking of random number of targets with random number of sensors using Random Finite Set Theory

Variation in the number of targets and sensors needs to be addressed in any realistic sensor system. Targets may come in or out of a region or may suddenly stop emitting detectable signal. Sensors can be subject to failure for many reasons. We derive a tracking algorithm with a model that includes these variations using Random Finite Set Theory (RFST). RFST is a generalization of standard probability theory into the finite set theory domain. This generalization does come with additional mathematical complexity. However, many of the manipulations in RSFT are similar in behavior and intuition to those of standard probability theory.

Tracking of Acoustic Sources Using Random Set Theory

In a passive acoustical tracking, sources are not expected to cooperate with the system. They may enter and leave the sensing area at various speed and directions. They may even stop vocalizing at any time. The standard solution divides this problem into two steps: estimating the number of sources, followed by tracking their geo-kinematic states. Instead, we propose using the random set theory to jointly estimate both the number of sources and their geo-kinematic states. The random set theory is a mathematical framework that gives a complete picture of the joint estimation problem. With this framework, one can analytically develop a tracking system that is robust against node failures or dynamic multiple source scenario. To cover the common wireless sensing deployment strategy, we investigate two topologies: a uniform grid, and a distributed sub-array. With each topology, we validate the proposed tracking system solution with simulations. For the distributed sub-array topology, we use results from a controlled field experiment and found this approach to be both practical and robust.

Collection and processing of acoustic and seismic array data for source localization

Distributed sensor networks have been proposed for a wide range of applications. In this paper, our goal is to locate a wideband source, generating both acoustic and seismic signals, using both seismic and acoustic sensors. For a far-field acoustic source, only the direction-of-arrival (DOA) in the coordinate system of the sensors is observable. We use the approximate Maximum-Likelihood (AML) method for DOA estimations from severalacoustic arrays. For a seismic source, we use data collected at a single tri-axial accelerometer to perform DOA estimation. Two seismic DOA estimation methods, the eigen-decomposition of the sample covariance matrix method and the surface wave method are used. Field measurements of acoustic and seismic signals generated by vertically striking a heavy metal plate placed on the ground in an open field are collected. Each acoustic array uses four low-cost microphones placed in a square configuration and separated by one meter. The microphone outputs of each array are collected by a synchronized A/D recording system and processed locally based on the AML algorithm for DOA estimation. An array of six tri-axial accelerometers arranged in two rows whose outputs are fed into an ultra low power and high resolution network-aware seismic recording system. Field measured data from the acoustic and seismic arrays show the estimated DOAs and consequent localizations of the source are quite accurate and useful.

Automatic vocal individual recognition of acorn woodpecker (Melanerpesformicivorus) based on hidden Markov models

The acorn woodpecker (Melanerpes formicivorus) is a highly social cooperatively breeding species. A variety of conspecific interactions inside and between family groups is mediated by vocalizations. In the current study, acoustic sensor networks are used to monitor vocal behaviors of acorn woodpecker. In order to identify the callers in our behavioral study, we present a method for the automatic vocal individual recognition using hidden Markov models (HMMs). Field recordings from eight woodpeckers in two family groups were made in northern California. Individual identification accuracy is 92.65% for isolated waka calls, 84.31% for isolated syllable wa, 73.04% for isolated syllable ka, using 7‐state HMMs with MFCC—E—D parametrization. For the continuous recordings, the recognition rate is 66.67%–100.00%. We studied the influence of two major factors in the performance of the HMMs: (i) different structures of HMMs and parametrizations of data, (ii) different qualities of signals, including degraded calls af...