Page 7 - Summer 2008

SIGNAL PROCESSING IN ACOUSTICS: SCIENCE OR SCIENCE FICTION?
James V. Candy
Lawrence Livermore National Laboratory and University of California, Santa Barbara Livermore, California 94551
Signal processing in acoustics is
based on one fundamental con-
cept—extracting critical informa-
tion from noisy, uncertain measurement
1-5
data.
Acoustical processing problems
can lead to some complex and intricate
paradigms to perform this extraction especially from noisy,
sometimes inadequate, measurements. Whether the data
are created using a seismic geophone sensor from a moni-
toring network or an array of hydrophone transducers
located on the hull of an ocean-going vessel, the basic pro-
cessing problem remains the same—extract the useful
information. Techniques in signal processing (e.g., filtering,
Fourier transforms, time-frequency and wavelet trans-
forms) are effective; however, as the underlying acoustical
process generating the measurements becomes more com-
plex, the resulting processor may require more and more
information about the process phenomenology to extract
the desired information. The challenge is to formulate a
meaningful strategy that is aimed at performing the pro-
cessing required, even in the face of these high uncertain-
ties. This strategy can be as simple as a transformation of
the measured data to another domain for analysis or as
complex as embedding a full-scale propagation model into
5-7
For example, think of trying to extract a set of resonances (damped sinusoids) from accelerometer time series. It is nearly impossible to calculate zero-crossings from the time series but it is a simple matter to transform the data to the spectral domain using the Fourier transform and then applying the property that sinusoids are impul- sive-like in Fourier space facilitating their extraction through peak detection. Finding a sinusoidal source propa- gating in the ocean is another matter that is quite complex due to the attenuation and dispersion characteristics of this harsh environment. Here, a complex propagation model must be developed and applied to “unravel” the highly dis- torted data to reveal the source—a simple Fourier trans- form will not work. The aims of both approaches are the same—to extract the desired information and reject the extraneous, and therefore, develop a signal processing scheme to achieve this goal. In this article, we briefly dis- cuss this underlying signal processing philosophy from a “bottoms-up” perspective enabling the problem to dictate the solution rather than vise-versa. Once accomplished, we ask ourselves the final and telling question, “Did it work or are we kidding ourselves?” Are the results science or are
the processor.
they science fiction?
More specifically, signal processing (Note that through-
out this article we will use the term “signal processing” to
“...to extract the desired information and reject the extraneous...”
encompass all of the techniques used to extract the useful information from data. Such techniques as image processing, tomography, array processing, spectral processing, model-based processing, etc. are implied by this terminology) forms
the basic nucleus of many acoustical applications. It is a spe- cialty area that many acousticians apply in their daily techni- cal regimen with great success such as the simplicity in Fourier analysis of resonance data or in the complexity of analyzing the time-frequency response of dolphin sounds. Acoustical applications abound with unique signal process- ing approaches offering solutions to the underlying problem. For instance, the localization of a target in the hostile under- water ocean acoustic environment not only challenges the acoustician, but also taxes the core of signal processing basics thereby requiring that more sophistication and a priori knowledge be incorporated into the processor. This particu- lar application has led to many advances both in underwater signal processing as well as in the development of a wide vari- ety of so-called model-based or physics-based processors. A prime example of this technology is the advent of the model- based matched-field processor7-9 that has led not only to a solution of the target localization problem, but also to many applications in other acoustical areas such as nondestructive evaluation and biomedical imaging. So the conclusion is the same, signal processing is a necessary ingredient as a working tool that must be mastered by the acoustician to extract the useful information from uncertain measurements.
Acoustics
Let us look at acoustical signal processing from a slight- ly different perspective. Acoustical data can be used to extract useful information about signal sources, the sur- rounding environment and background noise much the same as any other modality (e.g., electromagnetics: radio frequen- cy (RF), infrared (IR), optics, etc.). The information is clear- ly different but can also be used effectively. The uniqueness afforded is determined by how the acoustic signals propagate within the particular environment. The information available or carried by the acoustic signal (wave) depends heavily on the source characteristics and the environment supporting the propagation in which the wave interacts or causes these signals to bounce, bend and spread in a multitude of direc- tions distorting both their shape and arrival times at sensor locations. Localization (incoherent) of the source is per- formed by estimating the arrival times (time delays) and using geometric relations (triangularization). The source characteristics also determine the underlying acoustical
6 Acoustics Today, July 2008