Humans Are Not Always the Source of Errors: Characterization and Assessment of Remote Operator (RO) Resilient Performance in m:N sUAS Beyond Visual Line of Sight (BVLOS) Operations

As the use of small Uncrewed Aircraft Systems (sUAS) continues to expand across commercial and public sectors, human-machine teaming paradigms such as m:N (m = number of crew members and N = number of sUAS being managed by ‘m’ crew) Beyond Visual Line of Sight (BVLOS) control where a single Remote Operator (RO) manages multiple sUAS are gaining prominence due to their benefits of scalability and profitability. However, the increased complexity and uncertainty inherent in such complex sociotechnical operations pose a critical cognitive, temporal, physical, and mental demand on the human-in-the-loop in both–during nominal operations and when facing multiple disruptions on multiple sUAS simultaneously. Hence, this dissertation explores how RO resilient performance can be characterized and assessed, within the context of m:N sUAS BVLOS flight operations. It investigates how ROs sustain resilient performance under such complex, dynamic operational conditions, where unexpected disruptions and functional or task interdependencies are common. System resilience is often attributed to technological sophistication; however, it is the human-in-the-loop who acts as the last line of defense by showing safety-producing behaviors–especially when automation fails or behaves erratically in response to increased operational complexities, or unpredictable scenarios unfold. Despite this, existing literature tends to disproportionately focus on humans as sources of error, rather than recognizing their proactive role in adapting to disruptions, leaving a gap in our understanding of one of the major sources of ensuring system and operational resilience–human-in-the-loop–in real-world settings.The research is consists of three main questions. RQ1 identifies systemic and operational factors that influence RO’s resilient performance–both positively and negatively. RQ2 characterizes resilient RO behaviors in response to real-time disruptions and operational demands, resulting in 1) a comprehensive list of RO resilience proficiencies necessary to deal with multiple disruptions on multiple sUAS simultaneously and 2) the proposal of extending the existing model of resilience potentials with the addition of ‘ability to assess’ to make it applicable for m:N sUAS BVLOS use case. Finally, RQ3 presents a novel rubric, combining both–extended resilience potentials model and Bloom’s Taxonomy, to assess RO resilience proficiencies across basic, intermediate, and advanced levels.Data collection was conducted through direct observation, video recording, and semi-structured retrospective think-aloud (RTA) interviews during actual m:N sUAS BVLOS flight test operations conducted as a part of the FAA-BAA project titled ‘697DCK-22-C-00381 Standards for Piloting Multiple UAS BVLOS’. Further, this data was analyzed using qualitative data analysis methods of hybrid coding and thematic analysis. Despite the study being constrained by a limited number of participants and sUAS, due to the resource-intensive and regulatory nature of executing such flight testing, its evidence-based foundation strengthens the findings and makes them more applicable to be used by other researchers and the industry. The list of RO resilience proficiencies and rubric provides a foundation for scenario-based training and assessment, and for crafting RO certification benchmarks. This dissertation aims to deepen the understanding of human resilience in complex sociotechnical, autonomous aviation operations and offer a path toward safer and more resilient m:N sUAS BVLOS operations.Future research should expand on this work through broader crew sampling and operational conditions, simulation-based studies, and the development of quantitative performance metrics to characterize and assess safety-producing behaviors of the human-in-the-loop, aligning with initiatives such as NASA’s System-wide Safety (SWS) project.