From human-human trust to human-autonomous system trust

• Main Author: Pushparaj, Kiranraj
• Other Authors: Vu N. Duong
• Format: Thesis-Doctor of Philosophy
• Language: English
• Published: Nanyang Technological University 2022
• Subjects: Engineering::Aeronautical engineering::Aviation
• Online Access: https://hdl.handle.net/10356/161893
• Institution: Nanyang Technological University

A sustained increase in air traffic has resulted in a greater workload for Air Traffic Controllers (ATCOs). Researchers have tried to alleviate this issue by attempting to introduce advanced intelligent decision aids for ATCOs, that are able to assist them in operation. However, for an efficient integration of these autonomous tools into Air Traffic Management (ATM) operations, ATCOs must use these decision aids appropriately. Even though the predominant human factor that governs how ATCOs use their tools is trust, it remains a field that has not been able to keep up with the pace of research in the more technical aspects of ATM. Matching the pace of development is absolutely necessary in order to gain a deeper insight into the nature of ATCO-Autonomous System interaction. The research that has been conducted so far tends to treat autonomous systems as a subset of automation, where they are considered to be a more evolved version of automation tools. However, the added element of autonomy warrants further specialised examination. Autonomy has progressed to the stage where it is starting to be viewed as a legitimate teammate, rather than just technology. This shift in dynamic calls for ATCO-Autonomous System Interaction to be analysed in its own right. The characteristic autonomy that is present in these interactions suggests that there may be similarities with Human-Human Interaction, which also involves an autonomous trustee. This brought about the question of whether finding from Human-Human Trust can be applied to Human-Autonomous System Trust. As such, the strength of the relationship between Human-Human Trust and Human-Autonomous System trust was examined empirically in the context of both Propensity to Trust, as well as Momentary Trust using less subjective and more quantitative methods such as an oblique questionnaire. Furthermore, when exploring the extent of the parallels between Momentary Human-Human Trust and Momentary Human-Autonomous System Trust, a neuroergonomic approach was adopted to improve the quality of data obtained. The results indicated a strong relationship between Human-Human Trust and Human-Autonomous System Trust. Subsequently, the similarities between them were assimilated into the development of a robust model of ATCO-Autonomous System Trust that simultaneously juxtaposed two opposing representations of trust and distrust. Past research presented the views that trust and distrust could be on opposing ends of the same continuum, or mutually exclusive of one another. The use of a novel and innovative quantum-inspired model reconciled both of these perspectives, with objective neuroimaging data that was obtained from experiments with ATCOs. This model was further calibrated with supplemental behavioural data that was also collected. The development of this novel model demonstrated the feasibility of utilising Human-Human Trust as a starting point to speed up the growth of Human-Autonomous System Trust. While they are by no means virtually indistinguishable, the empirically determined similarities, such as similar interaction patterns and neural activity dictate that insights from Human-Human Trust are certainly applicable to Human-Autonomous System Trust. As such, this can certainly serve as a realistic and practical approach to accelerate the growth of ATCO-Autonomous System Trust research and catch up to the development of intelligent decision aids for ATCO use. A more comprehensive understanding of Human-Autonomous System Trust will likely lead to a smoother integration of the various decision support tools that are being developed for ATCOs by ensuring appropriate use through calibrated trust levels.