Peter Stone's Selected Publications

Classified by TopicClassified by Publication TypeSorted by DateSorted by First Author Last NameClassified by Funding Source


Toward Agile Maneuvers in Highly Constrained Spaces: Learning from Hallucination

Toward Agile Maneuvers in Highly Constrained Spaces: Learning from Hallucination.
Xuesu Xiao, Bo Liu, Garrett Warnell, and Peter Stone.
IEEE Robotics and Automation Letters (RA-L), January 2021.
5-minute video demonstration
Project webpage

Download

[PDF]3.2MB  

Abstract

While classical approaches to autonomous robot navigation currently\ enable operation in certain environments, they break down in tightly constrained\ spaces, e.g., where the robot needs to engage in agile maneuvers to squeeze\ between obstacles. Recent machine learning techniques have the potential to\ address this shortcoming, but existing approaches require vast amounts of navigation\ experience for training, during which the robot must operate in close proximity to\ obstacles and risk collision. In this paper, we propose to side-step this requirement by\ introducing a new machine learning paradigm for autonomous navigation called\ learning from hallucination (LfH), which can use training data collected in completely\ safe environments to compute navigation controllers that result in fast, smooth, and\ safe navigation in highly constrained environments. Our experimental results show\ that the proposed LfH system outperforms three autonomous navigation baselines on\ a real robot and generalizes well to unseen environments, including those based on\ both classical and machine learning techniques.

BibTeX Entry

@article{ral21-xiao,
  title={Toward Agile Maneuvers in Highly Constrained Spaces: Learning from Hallucination},
  author={Xuesu Xiao and Bo Liu and Garrett Warnell and Peter Stone},
  journal={IEEE Robotics and Automation Letters (RA-L)},  
  abstract={While classical approaches to autonomous robot navigation currently\\ enable operation in certain environments, they break down in tightly constrained\\ spaces, e.g., where the robot needs to engage in agile maneuvers to squeeze\\ between obstacles. Recent machine learning techniques have the potential to\\ address this shortcoming, but existing approaches require vast amounts of navigation\\ experience for training, during which the robot must operate in close proximity to\\ obstacles and risk collision. In this paper, we propose to side-step this requirement by\\ introducing a new machine learning paradigm for autonomous navigation called\\ learning from hallucination (LfH), which can use training data collected in completely\\ safe environments to compute navigation controllers that result in fast, smooth, and\\ safe navigation in highly constrained environments. Our experimental results show\\ that the proposed LfH system outperforms three autonomous navigation baselines on\\ a real robot and generalizes well to unseen environments, including those based on\\ both classical and machine learning techniques.},
  year={2021},
  month={January},
  wwwnote={<a href="https://www.youtube.com/watch?v=AE-KgxJS-iE">5-minute video demonstration</a><br><a href="https://www.cs.utexas.edu/~xiao/Research/LfH/LfH.html">Project webpage</a>}
}

Generated by bib2html.pl (written by Patrick Riley ) on Mon Mar 25, 2024 00:05:09