Long-Horizon Planning and Execution With Functional Object-Oriented Networks

Following work on joint object-action representations, functional object-oriented networks (FOON) were introduced as a knowledge graph representation for robots. A FOON contains symbolic concepts useful to robot's understanding of tasks and its environment xmlns:xlink="http://www.w3.org/1999/xlink">object-level planning . Prior this work, little has been done show how plans acquired from can be executed by robot, the in are too abstract execution. We thereby introduce idea exploiting object-level task planning Our approach automatically transforms into PDDL leverages off-the-shelf planners, action contexts, robot skills hierarchical pipeline generate executable plans. demonstrate our entire long-horizon CoppeliaSim learned contexts extended never-before-seen scenarios.