Avoiding automation surprise: Identifying requirements to support pilot intervention in automated Uncrewed Aerial Vehicle (UAV) flight

Abstract

The breadth and depth of Uncrewed Aerial Vehicle (UAV) operations are expanding at a considerable rate. With expansion comes challenges for the design of automation to support decision making. This research takes the perceptual cycle model (PCM) and the derived trust version of the Schema World Action Research Method (T-SWARM), to identify the issues and challenges of pilot intervention in UAVs operating during highly automated states. Nine UAV pilots with current experience operating medium to large UAVs were interviewed, using T-SWARM, about incidents in which they initiated an intervention in system operation (i.e. to avoid weather or collision) and an event where the system initiated the intervention (i.e. due to system failure). The coded responses highlighted the challenges with what information is displayed, how it is displayed and how it influences decision-making in the UAV context. In addition, the responses also identified aspects that influence trust in the system, including personal disposition, affect interventions with the automation. Against each of the key factors identified recommendations are made to increase safety and operational efficiency of UAV operations. This research adds to the growing body of literature that supports the application of T-SWARM for eliciting knowledge in the aviation domain and specifically within the UAV domain.