Guidance of a Refinement-based Acting Engine with a Hierarchical Temporal Planner

Jérémy Turi (1) , Arthur Bit-Monnot (1, 2)

1 LAAS-RIS - Équipe Robotique et InteractionS
2 INSA Toulouse - Institut National des Sciences Appliquées - Toulouse

Endowing robotic platforms with advanced deliberation capabilities to support autonomous and intelligent behavior has been a key objective of the scientific community for several decades. Where many acting systems propose to first perform offline planning and then execute the plan, we propose here the OMPAS system, based on hierarchical operational models, which proposes to leverage planning as a guidance to an otherwise reactive system. For these hierarchical operational models, this paper proposes both a Lisp-based language to define them, and its automatic analysis to generate predictive models of the system's behavior that can be leveraged by a hierarchical planner. The generation of predictive models allows to have a unified language for execution and planning, while having a rich structure to program operational models.

Domaines: Intelligence artificielle [cs.AI]

Jérémy Turi, Arthur Bit-Monnot. Guidance of a Refinement-based Acting Engine with a Hierarchical Temporal Planner. ICAPS Workshop on Integrated Planning, Acting, and Execution (IntEx), Jun 2022, Singapore (virtual), Singapore. ⟨hal-03690039⟩