2021/11/09 by Maximilian Diehl, Diehl, Maximilian, Chris Paxton +3
Computer Science · Engineering · #AI-based Problem Solving and Planning #FOS: Computer and information sciences #Logic, Reasoning, and Knowledge #Robot Manipulation and Learning #Robotic Path Planning Algorithms #Robotics (cs.RO)
paper · pdf · doi:10.48550/arxiv.2111.05397
openalex publication_date 2021/11/09 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
We have recently introduced a system that automatically generates robotic\nplanning operators from human demonstrations. One feature of our system is the\noperator count, which keeps track of the application frequency of every\noperator within the demonstrations. In this extended abstract, we show that we\ncan use the count to slim down domains with the goal of decreasing the search\ntime for long-horizon planning goals. The conceptual idea behind our approach\nis that we would like to prioritize operators that have occurred more often in\nthe demonstrations over those that were not observed so frequently. We,\ntherefore, propose to limit the domain only to the most popular operators. If\nthis subset of operators is not sufficient to find a plan, we iteratively\nexpand this subset of operators. We show that this significantly reduces the\nsearch time for long-horizon planning goals.\n