The CONVINCE Perspective on Task and Motion Planning in Dynamic Environments
摘要
Combined task and motion planning (TAMP) has been extensively studied within robotics. This paper explores TAMP in dynamic environments in the context of the European Union Horizon project CONVINCE (context-aware verifiable and adaptive dynamic deliberation). CONVINCE aims to develop cognitive deliberation capabilities for safe and autonomous robot operation over extended periods. The technical contributions of CONVINCE, including solutions to TAMP, will be validated on three real-world use cases: a robot vacuum cleaner, an assembly robot, and a robotic museum guide. In this paper, we briefly explain our proposed TAMP strategies for these use cases.