Examination of the Influence of Automated Spacecraft Motion on Operational Decision-Making in Autonomous Rendezvous & Docking Maneuvers

Abstract

As space missions travel beyond lunar orbit, real-time support from mission control will decrease due to larger communication delays. Much of the system-monitoring and decision-making tasks performed by mission control will shift to the crew, resulting in a potentially unmanageable workload that can compromise crew safety and mission success. To mitigate this increased workload, some spacecraft systems can be automated. However, increasing automated systems in complex task performance can result in adverse outcomes where the human operator's workload is not mitigated by the automation, but instead transformed into a high-vigilance monitoring task with the human operating as a fail-safe. Highly automated rendezvous and docking (ARD) systems, which are responsible for bringing a spacecraft into the orbital plane of its docking target and facilitating the connection of the spacecraft to the target, may encounter this workload-shift phenomenon because the astronaut is required to vigilantly monitor the system and take over control when necessary. The cooperative performance between the automated system and human operator must be better characterized to inform methods, processes, and tools that may mitigate human-automation interaction challenges posed in ARD systems. This thesis focuses on the operational decision-making by the astronaut, such as in manual takeover decisions, which arise when the automated system experiences a situation outside of its design domain. Another operational decision arises when the astronaut has to assign control to the automation and focus their attention on other tasks. While more is known about human performance in takeover scenarios in other fields such as automated driving, less is known about the cognitive decision-making processes leading to the takeover or handoff decision by the human in space applications.

This thesis consists of two human-subjects studies and one simulation study that inform the characterization of the cognitive decision-making process of the human in operational decision-making in ARD maneuvers. Automated system and human factors that influence operational decision-making are identified and used to inform system design tradeoffs. Methodologies are presented that work toward characterizing the operational decision-making process of the human operator when monitoring an ARD process. The concept of automated motion legibility, or the intent-expression of the automated system status to the human monitor, is operationalized to understand aspects of automated system design that affect human decision-making in cooperative task performance. The first study shows that a human monitor's manual takeover decision-making process is influenced by spacecraft motion plan factors of initial condition, path curvature, and spacecraft orientation on the path. In the second study, an additional factor of the monitor's perspective when observing the motion was shown to influence manual takeover and automation handoff decision-making. The egocentric perspective provided the best support for takeover and handoff decision-making as measured by the metrics proposed in this thesis. Emergent findings when compared to the literature emphasize the influence of automated system embodiment and decision-making structure on automated motion legibility. The influential decision-making factors of spacecraft motion and monitor viewpoint are then contextualized with relative orbital dynamics in an ARD docking simulation. Results of the ARD dynamics simulation illuminated potential design trade-offs between human decision-making support and spacecraft resource constraints. The contributions of this thesis provided data-based and theory-based approaches to better understanding the cooperative performance between human automation interactive systems in ARD, which are critical to future space crew ability to successfully operate independently of Earth.