An "ant" is placed in a two-dimensional toroidal grid at the beginning 


a two-dimensional hexagonal lattice with periodic boundary
conditions, a three-dimensional rectangular lattice with periodic boundary conditions, et



"Robot Navigation in Partially Observable Domains 
using Hierarchical Memory-Based Reinforcement Learning"

Proceedings of International Conference on Ubiquitous Robots and Ambient Intelligence

�f (URAmI 2005) held as part of the International Robot Week 2005 organized by the Korea Robotics Society, Robotics Society of Japan and Korea Advanced Institute of Science & Technology (KAIST), Daejeon, Korea in November 2005.

\bibitem[srin]
B.~Srinivasan (2006)
"Analysis of Memory-Based Learning Schemes for Robot
Navigation in Discrete Grid-Worlds with Partial Observability.)
(\verv \http://www.stanford.edu/~bharsrin/docs/papers/flairs06-membased_learning.pdf\.)


In this paper, we attempt to find a solution to the
problem of robot navigation using different memory-based
learning schemes, and compare these schemes in terms of learning
speed and efficiency.


---
Image-based robot navigation from an image memory 

Source  Robotics and Autonomous Systems archive
Volume 55 ,  Issue 4  (April 2007) table of contents

Pages: 345-356   
Year of Publication: 2007 
ISSN:0921-8890  
Authors  Anthony Remazeilles   IRISA, Campus de Beaulieu, 35042 Rennes Cedex, France  
Francois Chaumette   IRISA, Campus de Beaulieu, 35042 Rennes Cedex, France