Ph.D. Dissertation Defense: Siya Kunde
Monday, September 12, 2022
10:00 a.m.
Location: Schorr Center 211
"Recognizing and Building User Proficiency with Small Unmanned Aerial Vehicles (sUAV)"
Unmanned Aerial Vehicles (UAVs) are some of the most widely adopted robotics platforms, and are prevalent across environmental, emergency response, and logistics problem solutions apart from remaining a hobbyist favorite. Currently these systems can be readily purchased and flown almost immediately. The latest drones come equipped with flight assistive modes and can be flown without conforming to a training and adoption schedule. However, user operation proficiency is crucial to (i) adoption of such systems since, users are often the often considered the final option in detecting and correcting abnormal behaviors, (ii) mitigation of accidents resulting from loss of control, leading to lost and crashed vehicles, and, ultimately, (iii) design of high performance adaptive autonomous systems. Consequently, it is useful to examine how user proficiency develops when interacting with UAVs.
We first propose and validate a method which can be used to identify natural user proficiency groups resulting from manual flight skills. We achieve this through a 32-participant exploratory user study, utilizing features derived from the control inputs applied by the user, and the resultant flight trajectory of the sUAV. Second, we identify a training plateau, and different factors that impact flight including the orientation of the UAV and the presence of external factors like wind. We do this by conducting two studies: the \emph{baseline} study with 21 active pilot participants and a longitudinal \emph{training} study with 16 novice participants. Lastly, we provide insights on users' mental model comprised of their perception of sounds, their gaze behaviors, and control abilities, using custom tests administered during both the \emph{baseline and training} studies.
The broader impact of this work is to motivate and advance a to-pronged method comprised of skills based testing and scaffolded learning techniques to overcome the knowledge and skill gap of users, which will be applied to not just formal UAS deployment but also hobbyist and commercial UAS communities.
Supervisory Committee:
Brittany Duncan, Advisor
Bonita Sharif
Carrick Detweiler
Jeffery Stevens