Plan-Space Hierarchical Planning with the Action Notation Modeling Language
Résumé
Planning in robotics must be considered jointly with Acting. Planning is an open loop activity which produces a plan, based on action models, the current state of the world and the desired goal state. Acting, on the other hand, is a closed loop on the environment activity (to execute command and perceive the state of the world). These two deliberative activities must be integrated and need to handle time, concur-rency, synchronization, deadlines and resources. The timeline representation for temporal plan space planning and acting is very expressive; it is also quite flexible for integrating planning and acting. The ANML language is a recent proposal motivated by combining the expressiveness of the timeline representation with the decomposition of HTN methods. This paper reports on FAPE (Flexible Acting and Planning Environment), to our knowledge the first system integrating an ANML planner and actor. Our current focus is not efficient temporal planning per se, but the tight integration of acting and planning, which is addressed by: (i) extending HTN methods with refinements, given by PRS procedures, of planned action primitives into low-level commands, (ii) interleaving the planning process with acting, the former implements plan repair, extension and replanning, while the latter follows PRS skills refinements, and (iii) executing commands with a dispatching mechanism that synchronizes observed time points of action effects and events with planned time. FAPE has been integrated to a PR2 robot and experimented in a home-like environment. The paper presents how planning is performed and integrated with acting, and describes briefly the robotics experiments and reports on initial performances.
Origine | Fichiers produits par l'(les) auteur(s) |
---|
Loading...