Automating Transfer of Robot Task Plans using Functorial Data Migrations

2024-06-22

arXiv.org

AI Summary

This work addresses the challenge of automatic cross-domain transfer of robot task plans. We propose a category-theoretic approach to plan migration that constructs functorial mappings between types and predicates in source and target domains, enabling structure-preserving generalization without re-planning or reliance on domain-specific plan representations. Our method is the first to introduce functor-based data migration into robotic task planning, overcoming limitations of conventional approaches—such as dependence on fixed formalisms (e.g., PDDL) and manual, domain-specific mapping. Leveraging ontology-based modeling and Blocksworld formalization, we successfully migrate and execute original Blocksworld task plans in the AI2-THOR kitchen environment. Experimental results demonstrate the feasibility, generality, and generalization capability of our framework on a realistic simulation platform.

Technology Category

• Intelligent Robots: Motion and Path Planning
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Planning, Routing, and Scheduling: Learning for Planning and Scheduling

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Vertical and domain-specific search

Abstract

This paper introduces a novel approach to ontology-based robot plan transfer using functorial data migrations from category theory. Functors provide structured maps between domain types and predicates which can be used to transfer plans from a source domain to a target domain without the need for replanning. Unlike methods that create models for transferring specific plans, our approach can be applied to any plan within a given domain. We demonstrate this approach by transferring a task plan from the canonical Blocksworld domain to one compatible with the AI2-THOR Kitchen environment. In addition, we discuss practical applications that may enhance the adaptability of robotic task planning in general.

Problem

• Transferring robot task plans between domains using functorial data migrations
• Enabling plan transfer without replanning via structured ontology mappings
• Demonstrating framework by transferring Blocksworld plans to AI2-THOR Kitchen

Innovation

• Uses functorial data migrations for plan transfer
• Applies universally within source domain
• Transfers plans without replanning