Use of simulated mental models and active replanning for human-robot interaction
J. Ren, A. Swaminathan, Y. Zhou, S. Coogan, K. Feigh, J. Kolb, H. Miller
AIAA SCITECH 2025 Forum, 2025
Abstract
This paper introduces a communication and planning framework to facilitate efficient state update information between an autonomous robotic system and a human operator under scenarios where continuous robotic monitoring is not available. The framework estimates the operator's mental model and uses the difference between the estimated mental model and the actual world state and robot plans to trigger selective communication updates to the operator. The framework is deployed in a simulation environment where a rapid task planning algorithm adapts robot plans to a dynamic operational environment in real time. The proposed framework aims to improve situational awareness and reduce cognitive load in contrast with baseline methods where communication is triggered by mission milestones, obstacles encountered, or periodically.