Acoustic source localization in distributed sensor networks

Abstract

This paper studies the problem of sound source localization in a distributed wireless sensor network formed by mobile general purpose computing and communication devices with audio I/O capabilities. In contrast to well understood localization methods based on dedicated microphone arrays, in our setting sound localization is performed using a sparse array of arbitrary placed sensors (in a typical scenario, localization is performed by several laptops/PDAs co-located in a room). Therefore any far-field assumptions are no longer valid in this situation. Additionally, localization algorithm's performance is affected by uncertainties in sensor position and errors in A/D synchronization. The proposed source localization algorithm consists of two steps. In the first step, time differences of arrivals (TDOAs) are estimated for the microphone pairs, and in the second step the maximum likelihood (ML) estimation for the source position is performed. We evaluate the Cramer-Rao bound (CRB) on the variance of the location estimation and compare it with simulations and experimental results. We also discuss the effects of distributed array geometry and errors in sensor positions on the performance of the localization algorithm. The performances of the system are likely to be limited by errors in sensor locations and increase when the microphones have a large aperture with respect to the source.

Resource constrained efficient acoustic source localization and tracking using a distributed network of microphones

“…However, we employ an MLE estimator [6], which is asymptotically efficient. As shown in [9], the variance of the MLE is very close to the theoretical CRLB also when a small number of microphones is used.…”

Noise robust footstep location estimation using a wireless acoustic sensor network

“…Other research focusing on estimating sound source positions that is suited for a WASN [1,18], does not focus on footsteps and their specific challenges. However the algorithm described in [18] should have some tolerance against noise.…”

“…So ideally E period describes where the footstep sound energy is located within 1 gait period. 1 Then the most energetic 200 ms (the time a footstep produces sound, determined from the template footstep) within E period is selected as the detection for a single footstep and repeated F times to have detections for all F footsteps.…”

On UAV routing protocols for sparse sensor data exfiltration

“…Many authors have addressed the problem of estimating the source location from the noisy TOA measurements [5], [6], [7]. Others have addressed optimal sensor placements [8], [9], [10], [11].…”