Efficient POMDP Forward Search by Predicting the Posterior Belief Distribution

Online, forward-search techniques have demonstrated promising results for solving problems in partially observable environments. These techniques depend on the ability to efficiently search and evaluate the set of beliefs reachable from the current belief. However, enumerating or sampling action-obs...

Full description

Bibliographic Details
Main Authors: Roy, Nicholas, He, Ruijie
Other Authors: Nicholas Roy
Published: 2009
Online Access:http://hdl.handle.net/1721.1/46820
_version_ 1826214993932058624
author Roy, Nicholas
He, Ruijie
author2 Nicholas Roy
author_facet Nicholas Roy
Roy, Nicholas
He, Ruijie
author_sort Roy, Nicholas
collection MIT
description Online, forward-search techniques have demonstrated promising results for solving problems in partially observable environments. These techniques depend on the ability to efficiently search and evaluate the set of beliefs reachable from the current belief. However, enumerating or sampling action-observation sequences to compute the reachable beliefs is computationally demanding; coupled with the need to satisfy real-time constraints, existing online solvers can only search to a limited depth. In this paper, we propose that policies can be generated directly from the distribution of the agent's posterior belief. When the underlying state distribution is Gaussian, and the observation function is an exponential family distribution, we can calculate this distribution of beliefs without enumerating the possible observations. This property not only enables us to plan in problems with large observation spaces, but also allows us to search deeper by considering policies composed of multi-step action sequences. We present the Posterior Belief Distribution (PBD) algorithm, an efficient forward-search POMDP planner for continuous domains, demonstrating that better policies are generated when we can perform deeper forward search.
first_indexed 2024-09-23T16:15:07Z
id mit-1721.1/46820
institution Massachusetts Institute of Technology
last_indexed 2024-09-23T16:15:07Z
publishDate 2009
record_format dspace
spelling mit-1721.1/468202019-04-10T11:56:26Z Efficient POMDP Forward Search by Predicting the Posterior Belief Distribution Roy, Nicholas He, Ruijie Nicholas Roy Robotics, Vision & Sensor Networks Online, forward-search techniques have demonstrated promising results for solving problems in partially observable environments. These techniques depend on the ability to efficiently search and evaluate the set of beliefs reachable from the current belief. However, enumerating or sampling action-observation sequences to compute the reachable beliefs is computationally demanding; coupled with the need to satisfy real-time constraints, existing online solvers can only search to a limited depth. In this paper, we propose that policies can be generated directly from the distribution of the agent's posterior belief. When the underlying state distribution is Gaussian, and the observation function is an exponential family distribution, we can calculate this distribution of beliefs without enumerating the possible observations. This property not only enables us to plan in problems with large observation spaces, but also allows us to search deeper by considering policies composed of multi-step action sequences. We present the Posterior Belief Distribution (PBD) algorithm, an efficient forward-search POMDP planner for continuous domains, demonstrating that better policies are generated when we can perform deeper forward search. 2009-09-28T21:00:15Z 2009-09-28T21:00:15Z 2009-09-23 http://hdl.handle.net/1721.1/46820 MIT-CSAIL-TR-2009-044 Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Unported http://creativecommons.org/licenses/by-nc-nd/3.0/ 12 p. application/pdf application/postscript
spellingShingle Roy, Nicholas
He, Ruijie
Efficient POMDP Forward Search by Predicting the Posterior Belief Distribution
title Efficient POMDP Forward Search by Predicting the Posterior Belief Distribution
title_full Efficient POMDP Forward Search by Predicting the Posterior Belief Distribution
title_fullStr Efficient POMDP Forward Search by Predicting the Posterior Belief Distribution
title_full_unstemmed Efficient POMDP Forward Search by Predicting the Posterior Belief Distribution
title_short Efficient POMDP Forward Search by Predicting the Posterior Belief Distribution
title_sort efficient pomdp forward search by predicting the posterior belief distribution
url http://hdl.handle.net/1721.1/46820
work_keys_str_mv AT roynicholas efficientpomdpforwardsearchbypredictingtheposteriorbeliefdistribution
AT heruijie efficientpomdpforwardsearchbypredictingtheposteriorbeliefdistribution