Thursday, March 7, 2013
Mechanical and Civil Engineering Seminar
Novel Means for Localization and Identification of Remote Acoustic Sources
David Dowling, Professor of Mechanical Engineering, Mechanical Engineering, University of Michigan
Sound provides the only viable means for remote sensing in the ocean, and acoustic recordings from listening arrays are the most common starting point for detecting, localizing, and identifying a remote sound source. Unfortunately, these tasks are complicated by the fact that acoustic reflections and scattering from the ocean's boundaries and from inhomgeneities in its interior distort the originally broadcast sound as it propagates away from a remote source. And, the ocean's ever-changing acoustic environment is seldom known with sufficient fidelity to predict the details of this distortion. Thus, robust means for determining the location of a unknown remote source (source localization) and estimating its original broadcast waveform (blind deconvolution) in a poorly-known or unknown environment are enduring underwater remote sensing priorities. This presentation will describe how the basic physics of underwater sound propagation can be combined with novel nonlinear array-signal processing to recover out-of-band lower- or higher-frequency signal information from finite bandwidth signals. This manufactured signal information can be used for source localization to surpass the usual spatial Nyquist and diffraction limits of the receiving array at in-band signal frequencies. In addition, the manufactured below-band signal information can be exploited to overcome the ill-posed character of blind deconvolution, even when the receiving array is sparse in the signal's frequency band and ordinary beamforming is not useful. Results from simulations, laboratory experiments, and ocean acoustic propagation measurements will be shown. Although these novel techniques were developed for sonar applications, they might also be useful in sensor array applications in bio-medical ultrasound, radar, astronomy, and seismic remote sensing.