Human-Autonomy Teaming for Supervised Scientific Exploration: A Pilot Study in the Wild

Recent robotic competitions such as the DARPA SubT Challenge have spurred novel robotic autonomy capabilities. Among these capabilities, the NeBula autonomy solution was deployed during the 2022 NASA BRAILLE field campaign to verify its ability to perform in extreme terrain. Here we describe a pilot user study investigating different interfaces and interaction modalities for real-time supervisory control of robot teams in a scientific exploration scenario. We deployed a VR-based interface and an on-screen, browser-based interface during a field campaign at the Lava Beds National Monument (CA, USA). During this field test, the terrestrial Valentine cave (a lava tube) served as a proxy for planetary subsurface environments. We tasked the human-robot team to explore, map, and locate objects of interest using varying levels of autonomy support. Here, we present first insights based on raw NASA TLX scores for four tested conditions and motivate the need for further experiments with larger participant numbers in order to obtain more statistically significant results. We further note the challenges and limitations of large-scale, real-world deployments and raise questions that could lead to valuable insights for the preparation of future planetary missions.