Garrett Fitzgerald, U. S. Space Force; Rachel Morris, Applied Decision Science; Laura G. Militello, Applied Decision Science; Justin Fletcher, USSF/SSC
Keywords: Human-Machine Teaming, Cognitive Task Analysis, Space Domain Awareness, Trust in Autonomy, Telescope C2
Abstract:
A critical acceptance criterion of autonomous systems in military applications is operator trust in AI. In space sensing tasks for Space Domain Awareness (SDA), such as automated telescope command and control (C2), there exist a limited number of SMEs who can be systematically surveyed for user feedback. Cognitive Task Analysis (CTA) can be employed as a means to probe and elicit relevant SME feedback in the cycle of human-autonomous teaming system development. We establish a tailored framework to explore SME trust in a ground-based telescope autonomous C2 system (MACHINA) using prior concepts in Human-Machine Teaming (HMT) and cognitive engineering. A HMT Knowledge Audit is adopted for the SDA telescope C2 mission area and employed as a knowledge elicitation vehicle for USSF SDA operators. Analysis of interview data from three user groups is mapped to trust antecedents, yielding actionable HMT guidance for system development. This work presents examples of how SME trust in AI can be improved throughout the development of autonomous space sensing applications by combining human factors engineering and HMT concepts.
Date of Conference: September 17-20, 2024
Track: Machine Learning for SDA Applications