Thorsten Belker and Michael Beetz

Learning To Exceute Navigation Plans

KI-2001: Advances in Artificial Intelligence

---

Abstract

Most state-of-the-art navigation systems for autonomous service robots decompose navigation into global navigation planning and local reactive navigation. While the methods for navigation planning and local navigation are well understood, the plan execution problem, the problem of how to generate and parameterize local navigation tasks from a given navigation plan, is largely unsolved.

This article describes how a robot can autonomously learn to execute navigation plans. We formalize the problem as a Markov Decision Problem (MDP), discuss how it can be simplified to make its solution feasible, and describe how the robot can acquire the necessary action models. We show, both in simulation and on a RWI B21 mobile robot, that the learned models are able to produce competent navigation behavior.

---

Bibtex

@InProceedings{Belker01KI,
  author =       {T. Belker and M. Beetz},
  title =        {Learning To Execute Navigation Plans}, 
  booktitle =    {In: KI-2001: Advances in Artificial Intelligence}, 
  OPTcrossref =  {},
  OPTkey =       {},
  OPTpages =     {},
  year =         {2001},
  OPTeditor =    {F. Baader and G. Brewka and T. Eiter},
  volume =       {2174},
  OPTnumber =    {},
  OPTseries =    {},
  OPTaddress =   {},
  OPTmonth =     {},
  OPTorganization = {},
  OPTpublisher = {Springer},
  OPTnote =      {},
  OPTannote =    {}
}