Author (s)
Kurpakus Keegan
Affiliation
Masters of Science Thesis, The University of Mississippi, ProQuest, 2020
Publication date
2020
Abstract
Unmanned Aerial System (UAS) usage has continually increased in recent years for both recreational and military applications. One particular military application being researched is utilizing a UAS as a host platform for Hostile Fire Detection Systems (HFDS), with particular interest being focused on multi-rotor drone platforms. The type of HFDS considered in this work is based upon acoustic sensors. An acoustic based HFDS utilizes an array of microphones to measure acoustic data and then applies signal processing algorithms to determine if a transient signal is present and if present then estimates the direction from which the sound arrived. The main issue with employing an acoustic based HFDS on a multi-rotor drone is the high level of background noise due to motors, propellers, and flow noise. In this thesis a study of the acoustic near field, particularly relevant to microphones located on the drone, was performed to understand the noise produced by the UAS. More specifically, the causes and characteristics of the sources of noise were identified. The noise characteristics were then used to model the noise sources for multiple motor assemblies based upon position of the microphone and revolutions per minute (RPM) of the motors. Lastly, signal processing techniques were implemented to identify if transient signals are present and if present estimate the direction from which the sound arrives.
Full paper
https://www.proquest.com/openview/91a7574eb4c5bc4f5300e24b80eda999/1?pq-origsite=gscholar&cbl=44156
Keywords
unmanned aerial systems (UAS), hostile fire detection system (HFDS), acoustic signal processing, microphone array, drone noise modeling, transient sound detection, direction of arrival estimation (DoA)