Construction of source description database for machine learning analysis

Abstract

A large set of acoustic measurements were conducted in conjunction with Sinclair Community College Center for UAS Training and Certification at the Flying Pavilion in Dayton, Ohio. Ten aircraft were characterized at multiple altitudes above a circular array placed on the ground. A subset of these data, with two biologic signals, were presented to a sound jury who provided pairwise similarity judgments. These human perception data were organized into a similarity matrix that then passed through a multi-dimensional scaling analysis. The output of the scaling analyses were examined for each subject and the average across the data collection. This dimensional analysis provides a basis for machine learning analysis to determine the salient features.