Strategies for managing human workload to improve trust calibration in human-autonomy teaming (HAT)

Abstract

This thesis is situated within Human-Autonomy Teaming (HAT), focusing on human trust in autonomous teammates, particularly in Command and Control (C2) contexts.

Chapter 1 introduces HAT by drawing on Human-Human (HH) team dynamics, outlining motivations for HAT research and key topics such as trust calibration, transparency, and shared mental models.

Chapter 2 examines O’Neill’s critique of the fragmented approaches currently prevalent in HAT. It reviews a selection of conceptual models relevant to HAT, including those developed by De Visser et al. (De Visser et al., 2018, 2020), Hoff and Bashir (Hoff & Bashir, 2015), and Hou et al. (Hou et al., 2021), among others.

Chapter 3 follows up on Chapter 1’s initial proposal of using of Cognitive Fit Theory as a means of addressing the issue of cognitive overload associated with transparency information in HAT. Cognitive Fit Theory offers a mechanism for reducing the cognitive costs of visual information used in decision-making, making it particularly relevant to HAT contexts.

Chapter 4 details some methodologies and tools used in Experiments 1 and 2, focusing on SimpleC2Sim, an open-source web-based testbed developed by the author (GitHub.com). SimpleC2Sim improves accessibility, reduces logistical burdens, and provides an in-task behavioural measure for trust calibration.

Chapter 5 presents Experiment 1, which tests an early version of the Component-based Model of Cognitive Fit using three display types. While the model correctly predicted cognitive fit, it failed to accurately rank misfit displays due to a “distractor effect”, prompting a revision and a followup evaluation in Experiment 2.

Chapter 6 reports Experiment 2, applying the revised model (now incorporating the “distractor effect”) to a new task, but still using the same display types from Experiment 1.

Chapter 7 explores the broader implications of the findings, including partial validation of the Cost-based Team Interaction Cycle, and outlines future research directions based on the "distractor effect".

Date of Thesis submission	2026
Original language	English
Awarding Institution	Western Sydney University
Sponsors	Australian Government Research Training Program Scholarship
Supervisor	Chris Stanton (Supervisor), Kate Stevens (Supervisor) & John Cass (Supervisor)