Measuring Autonomy Pilot Interactions (MAPI)

Autonomy that can observe and react to pilot workload conditions is critical for improved mission performance in the Distributed Battle Management (DBM) vision of future air operations. This vision, advanced through DARPA's initiatives and transitioning into programs like AFRL’s Skyborg and the joint Collaborative Combat Aircraft (CCA), emphasizes the integration of autonomous uncrewed systems (AUS) in naval fleets by the late 2020s. Central to this integration is training pilots to optimize their interactions with AUS, addressing issues of under trust, over trust, under monitoring, and over monitoring. The research will focus on creating a workload estimation system that enables adaptive AUS to respond to pilots' cognitive and physical workload in real-time, mitigating performance degradation caused by either underload or overload.The proposed technical effort aims to develop a reliable and unobtrusive system for performing real-time operator workload estimation that can be used as input to scalable autonomy. This effort will build upon Dr. Adams’ adaptive human-robot teaming architecture, which leverages multi-dimensional workload estimation algorithms validated across various human-UAS teams. These algorithms use data from multiple sensors to accurately estimate cognitive, speech, auditory, visual, gross motor, fine motor, and tactile components of workload, supporting real-time, context-aware UAS adaptations. The integration of this system into NAWCAD’s Human Computer Integration Laboratory (HCIL) will involve identifying and installing a sensor suite and writing software to integrate Adams’ human-robot teaming architecture such that it can provide real-time autonomy adaptations to the HCIL’s virtual AUS.The Phase I objectives include formalizing Adams’ workload metrics in the context of Naval Aviation, finalizing the context-relevant data collection methods, adapting Adams’ human-robot teaming architecture for crewed-uncrewed-teaming missions, modifying HCIL to support this new technology, and demonstrating the feasibility of the system using a combination of real and surrogate data. This research will enhance the effectiveness of human-AUS teams by enabling adaptive, real-time responses to dynamic operator workload conditions, increasing mission performance by minimizing suboptimal states of overload and underload.