Feature based neural network acoustic transient signal classification.
Wichert, Terry S.
Collins, Daniel Joseph.
MetadataShow full item record
Utilization of neural network techniques to recognize and classify acoustic signals has long been pursued and shows great promise as a robust application of neural network technology. Traditional techniques have proven effective but in some cases are quite computationally intensive, as the sampling rates necessary to capture the transient result in large input vectors and thus large neural networks. This thesis presents an alternative transient classification scheme which considerably reduces neural network size and thus computation time. Parameterization of the acoustic transient to a set of distinct characteristics (e.g. frequency, power spectral density) which capture the structure of the input signal is the key to this new approach. Testing methods and results are presented on networks for which computation time is a fraction of that necessary with traditional methods, yet classification reliability is maintained. Neural network acoustic classification systems utilizing the above techniques are compared to classic time domain classification networks. Last, a case studs is presented which looks at these techniques applied to the acoustic intercept problem.
Approved for public release; distribution is unlimited.
Showing items related by title, author, creator and subject.
Scott, Alan D. (Monterey, California. Naval Postgraduate School, 1996-06);The US space industry has accumulated a vast amount of expertise in the testing of spacecraft to ensure these vehicles can endure the harsh environments associated with launch and on-orbit operations. Even with this corporate ...
O'Malley, Sean P. (2001-06);The acoustic impedance of a material describes its reflective and absorptive properties. Acoustic impedance may be measured in a wide variety of ways. This thesis describes the construction and testing of an acoustic ...
Schaus, Brian M. (Monterey, California: Naval Postgraduate School, 2015-03);Techniques for classifying maritime domain targets-of-interest within images are explored in this thesis. Geometric and photometric features within each image are extracted from processed images and are used to train a ...