Improving the Efficiency of Online POMDPs by using Belief Similarity Measures

J. Ballesteros, L. Merino, M. A. Trujillo, A. Viguria, and A. Ollero. Improving the Efficiency of Online POMDPs by using Belief Similarity Measures. In Proceedings of the IEEE International Conference on Robotics and Automation, ICRA, pp. 1792–1798, 2013.

Download

[PDF] [html] 

Abstract

In this paper, we introduce an approach calledFSBS (Forward Search in Belief Space) for online planning inPOMDPs. The approach is based on the RTBSS (Real-TimeBelief Space Search) algorithm of [1]. The main departurefrom the algorithm is the introduction of similarity measures inthe belief space. By considering statistical divergence measures,the similarity between belief points in the forward search treecan be computed. Therefore, it is possible to determine if acertain belief point (or one very similar) has been alreadyvisited. This way, it is possible to reduce the complexity of thesearch by not expanding similar nodes already visited in thesame depth. This reduction of complexity makes possible thereal-time implementation of more complex problems in robots.The paper describes the algorithm, and analyzes differentdivergence measures. Benchmark problems are used to showhow the approach can obtain a ten-fold reduction in thecomputation time for similar obtained rewards when comparedto the original RTBSS. The paper also presents experimentswith a quadrotor in a search application.

BibTeX Entry

@INPROCEEDINGS{ballesteros_icra13,
  author = {J. Ballesteros and L. Merino and M. A. Trujillo and A. Viguria and A. Ollero},
  title = {Improving the Efficiency of Online {POMDP}s by using Belief Similarity Measures},
  booktitle = ICRA,
  year = {2013},
  pages = {1792--1798},
  doi = {10.1109/ICRA.2013.6630813},
  abstract={In this paper, we introduce an approach called
FSBS (Forward Search in Belief Space) for online planning in
POMDPs. The approach is based on the RTBSS (Real-Time
Belief Space Search) algorithm of [1]. The main departure
from the algorithm is the introduction of similarity measures in
the belief space. By considering statistical divergence measures,
the similarity between belief points in the forward search tree
can be computed. Therefore, it is possible to determine if a
certain belief point (or one very similar) has been already
visited. This way, it is possible to reduce the complexity of the
search by not expanding similar nodes already visited in the
same depth. This reduction of complexity makes possible the
real-time implementation of more complex problems in robots.
The paper describes the algorithm, and analyzes different
divergence measures. Benchmark problems are used to show
how the approach can obtain a ten-fold reduction in the
computation time for similar obtained rewards when compared
to the original RTBSS. The paper also presents experiments
with a quadrotor in a search application.},
}

Generated by bib2html.pl (written by Patrick Riley ) on Thu Aug 16, 2018 19:28:37