ML Anthology

Using Learning to Recover Side-Effects of Operators in Robotics

Ralph P. Sobek, Jean-Paul Laumond
ICML 1989 pp. 205-208
doi:10.1016/B978-1-55860-036-2.50058-8 /icml/1989/sobek1989icml-using/

Abstract

This paper presents a paradigm for learning operator-level knowledge in a planning system for robotics. The paradigm uses a two-level representation which comprises the operator-level knowledge and domain description rules, nonmonotonic and monotonic, respectively. Our approach comprises three steps: error diagnosis, research of analogous past situations, and experimentation. It provides operator refinement through analysis of operator side-effects in memorized successful plans. An example in the context of a mobile robot system is presented to demonstrate the paradigm.

ICML Semantic Scholar

Cite

Text

Sobek and Laumond. "Using Learning to Recover Side-Effects of Operators in Robotics." International Conference on Machine Learning, 1989. doi:10.1016/B978-1-55860-036-2.50058-8

Markdown

[Sobek and Laumond. "Using Learning to Recover Side-Effects of Operators in Robotics." International Conference on Machine Learning, 1989.](https://mlanthology.org/icml/1989/sobek1989icml-using/) doi:10.1016/B978-1-55860-036-2.50058-8

BibTeX

@inproceedings{sobek1989icml-using,
  title     = {{Using Learning to Recover Side-Effects of Operators in Robotics}},
  author    = {Sobek, Ralph P. and Laumond, Jean-Paul},
  booktitle = {International Conference on Machine Learning},
  year      = {1989},
  pages     = {205-208},
  doi       = {10.1016/B978-1-55860-036-2.50058-8},
  url       = {https://mlanthology.org/icml/1989/sobek1989icml-using/}
}