Advancing Legible Planning in Robotics through Observer Perspective
A new method enhances legible planning in PDDL domains by considering observer perspectives.
In this research, Michele Persiani and Thomas Hellström introduce a novel approach to legible planning applicable to various PDDL domains. The method builds on existing legibility research in classical planning, eliminating the need for specialized planners. Additionally, the paper explores estimating the observer's perspective using a second-order theory of mind, linking the tasks of both the planner and the observer.