STPA Hazard Analysis of Human Supervisory Control of Multiple Unmanned Aerials Systems

Abstract

Unmanned Aircraft Systems (UAS) operations are shifting from multiple operators controlling a single-UAS to a single operator supervising multiple-UAS engaged in complex mission sets. To enable this paradigm change, there is wide consensus in the literature that limitations in human cognitive capacity require shifting low-level control responsibilities to automation so that human operators can focus on supervisory control. However, hazard analyses to identify related safety concerns have largely used traditional hazard analysis techniques that cannot handle the level of complexity of these systems and none can provide recommendations for the early stages of system development. To begin to address this shortfall, this thesis applies System-Theoretic Process Analysis (STPA) on a model of a multi-UAS system with human-supervisory control. This hazard analysis approach handles complex software and human-machine control interactions together. This thesis details both how the hazard analysis was executed and the implications of the analysis results. Numerous traceable causal scenarios are systematically identified and used to generate design recommendations. These recommendations, if applied, will help ensure multi-UAS systems with human supervisory control are designed with safety in mind.