Distributed Microphone Arrays Research Articles

Microphone Clustering and BP Network based Acoustic Source Localization in Distributed Microphone Arrays

Abstract —A microphone clustering and back propagation (BP) neural network based acoustic source localization method using distributed microphone arrays in an inte lligent meeting room is proposed. In the propos ed method, a novel clustering method is first used to divide all microphones into several clusters where each one corresponds to a specified BP network. Afterwards, the energy-based cluster selecting scheme is applied to the select the clusters which are small and close to the acoustic source. In each chosen cluster, the time difference of arrival of each microphone pair is estimated, and then all estimated time delays act as input of the corresponding BP network for position estimation. Finally, all estimated positions from the chosen clusters are fused for global position estimation. Only subsets rather than all the microphones are responsible for acoustic source localization, which leads to less computational cost; moreover, the local estimation in each chosen cluster can be processed in parallel, which expects to improve the localization speed potentially. Simulation results from comparison with other related localization methods confirm the validity of the proposed method. Index Terms—acoustic source localization, BP neural network, microphone clustering, GCC-PHAT, TDOA.

A Time Delay Estimation Method Based on Wavelet Transform and Speech Envelope for Distributed Microphone Arrays

1 Abstract—A time delay estimation method based on wavelet transform and speech envelope is proposed for distributed microphone arrays. This method first extracts the speech envelopes of the signals processed with multi-level discrete wavelet transform, and then makes use of the speech envelopes to estimate a coarse time delay. Finally it searches for the accurate time delay near the coarse time delay by the crosscorrelation function calculated in time domain. The simulation results illustrate that the proposed method can accurately estimate the time delay between two distributed microphone array signals.

Acoustic Source Localization and Tracking of a Time-Varying Number of Speakers

Particle filter-based acoustic source tracking algorithms track (online and in real-time) the position of a sound source-a person speaking in a room-based on the current data from a distributed microphone array as well as the previously recorded data. This paper develops a multi-target tracking (MTT) methodology to allow for an unknown and time-varying number of speakers in a fully probabilistic manner and in doing so does not resort to independent modules for new target proposal or target number estimation as in previous works. The approach uses the concept of an existence grid to propose possible regions of activity before tracking is carried out with a variable dimension particle filter-which also explicitly supports the concept of a null particle, containing no target states, when no speakers are active. Examples demonstrate typical tracking performance in a number of different scenarios with simultaneously active speech sources.

A distributed array processing using multi-channel signals over a network with an embedded time code by the network time protocol

A conventional microphone array system uses a conductor to wire from a microphone to an input via an amplifier. While, a wireless transmission for an array system makes the configuration flexible, and it is expected to provide novel applications widely, such as measuring of impulse response in wide area. One of the issues is a synchronization of time between channels. This paper proposes multiple signals transmission system over a network with a time code embedding to synchronize those signals. The system consists of clients and a server. Each client at a receiving point runs a network time protocol client daemon, and a packetized audio signal with a time code is sent to a server. The signal from a client is reconstructed at a server based on the time code. Since the time differences between clients affects to performance of the multichannel signal processing, small error in time at a client is preferred. This paper discusses how the error in time between channels affects to performance of the distributed microphone array system.

Reconstructing echolocation behavior using time difference of arrival localization and a distributed microphone array as a virtual Telemike.

Bats use echolocation to probe their 3-D environment. New techniques in distributed microphone array processing allow us to record echolocation behavior, reconstruct the flight paths, and generate an acoustic record of individual bats flying the laboratory or the field. Probabilistic methods of time difference of arrival localization allow us to follow bats navigating in flight room mazes, foraging in the wild and using active sonar in the presence of possible jamming from conspecifics. By reversing the localization operation, we back out the changing propagation times of the sounds emitted by bats flying fast in three dimensions. Through beamforming, we then create an amplified acoustic track that follows the reference frame of the bat, amounting to a virtual Telemike based on consensus data from the microphone array. Spatial tracking and virtual on-board recording allow us to ask new questions about the spectral and temporal strategies employed by echolocating bats and to model the neural and perceptual natures of sonar. [Work supported by ONR and NSF.]

Segmentation, Indexing, and Retrieval for Environmental and Natural Sounds

We propose a method for characterizing sound activity in fixed spaces through segmentation, indexing, and retrieval of continuous audio recordings. Regarding segmentation, we present a dynamic Bayesian network (DBN) that jointly infers onsets and end times of the most prominent sound events in the space, along with an extension of the algorithm for covering large spaces with distributed microphone arrays. Each segmented sound event is indexed with a hidden Markov model (HMM) that models the distribution of example-based queries that a user would employ to retrieve the event (or similar events). In order to increase the efficiency of the retrieval search, we recursively apply a modified spectral clustering algorithm to group similar sound events based on the distance between their corresponding HMMs. We then conduct a formal user study to obtain the relevancy decisions necessary for evaluation of our retrieval algorithm on both automatically and manually segmented sound clips. Furthermore, our segmentation and retrieval algorithms are shown to be effective in both quiet indoor and noisy outdoor recording conditions.

Compact microphone array processing for in-vehicle voice detection applications.

The main goal of voice activity detection algorithms is to determine the presence of human voice signals in a given environment. Voice activity detection is very challenging in vehicle interiors. The main challenge in detecting the presence of voice signals in vehicles is the presence of a large number of interferers and a high background noise level. Further, many types of interferers such as tire or engine noise have signals that are highly nonstationary. In this work, compact microphone arrays mounted in various locations within a vehicle are used to extract signals from locations of interest. Experimental comparisons of the performance of several voice activity detection algorithms are made for various array configurations (including single microphone) and source extraction processing algorithms. Processing algorithms considered include blind source separation algorithms such as fastICA, transfer function based inversion methods, and both fixed and adaptive beamforming techniques. The performance of compact arrays is also compared to the performance of larger distributed microphone arrays. It is shown that the use of compact microphone arrays can significantly improve voice activity detection algorithms. The gain is quantified using receiver operating characteristic type curves plotting probability of detection vs probability of false alarm.

A scanning bi-static SODAR

Field results are given from a bi-static SODAR which uses a single central vertical transmission and three distributed microphone array receivers. Fourier transform delay methods are applied to data sampled from each microphone to retrospectively scan in angle and follow the transmitted pulse. Advantages of sampling a narrow atmospheric column, rather than distributed volumes are discussed.

Robust tracking of moving sound source using multiple model Kalman filter

In this paper a novel method for tracking an active speaker in a noisy and reverberant environment by means of a spatially distributed microphone array is presented. Firstly, a sound source localization algorithm based on time delays of arrival (TDOA) in microphone pairs provides observed position estimates. Then these remarkably noisy estimates are filtered by a multiple model Kalman filter (MMKF) in order to obtain a smoothed trajectory of the speaker’s movement. Compared with the traditional Kalman filter (KF), simulated results prove the MMKF is more robust and effective in noisy environments.

Sound Localization and Separation for Mobile Robot Tele-Operation by Tri-Concentric Microphone Array

The paper describes a tele-operated mobile robot system which can perform multiple sound source localization and separation using a 32-channel tri-concentric microphone array. Tele-operated mobile robots require two main capabilities: 1) audio/visual presentation of the robot’s environment to the operator, and 2) autonomy for mobility. This paper focuses on the auditory system of a tele-operated mobile robot in order to improve both the presentation of sound sources to the operator and also to facilitate autonomous robot actions. The auditory system is based on a 32-channel distributed microphone array that uses highly efficient directional design for localizing and separating multiple moving sound sources. Experimental results demonstrate the feasibility of inter-person distant communication through the tele-operated robot system.

The Fusion of Distributed Microphone Arrays for Sound Localization

This paper presents a general method for the integration of distributed microphone arrays for localization of a sound source. The recently proposed sound localization technique, known as SRP-PHAT, is shown to be a special case of the more general microphone array integration mechanism presented here. The proposed technique utilizes spatial likelihood functions (SLFs) produced by each microphone array and integrates them using a weighted addition of the individual SLFs. This integration strategy accounts for the different levels of access that a microphone array has to different spatial positions, resulting in an intelligent integration strategy that weighs the results of reliable microphone arrays more significantly. Experimental results using 10 2-element microphone arrays show a reduction in the sound localization error from 0.9 m to 0.08 m at a signal-to-noise ratio of 0 dB. The proposed technique also has the advantage of being applicable to multimodal sensor networks.