CaltechAUTHORS
Published August 31, 2018 | Version Published
Journal Article Open

Temporal Logic Control of POMDPs via Label-based Stochastic Simulation Relations

Creators

Abstract

The synthesis of controllers guaranteeing linear temporal logic specifications on partially observable Markov decision processes (POMDP) via their belief models causes computational issues due to the continuous spaces. In this work, we construct a finite-state abstraction on which a control policy is synthesized and refined back to the original belief model. We introduce a new notion of label-based approximate stochastic simulation to quantify the deviation between belief models. We develop a robust synthesis methodology that yields a lower bound on the satisfaction probability, by compensating for deviations a priori, and that utilizes a less conservative control refinement.

Additional Information

© 2018, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. Available online 31 August 2018. This research was carried out at JPL and Caltech under a contract with the NASA and funded through the President's and Director's Fund Program.

Attached Files

Published - 1-s2.0-S2405896318311625-main.pdf

Files

1-s2.0-S2405896318311625-main.pdf

Files (493.2 kB)

Name Size Download all
md5:d77533357b91bab7c927ae34bfbd6b02
493.2 kB Preview Download

Additional details

Identifiers

Eprint ID
89577
Resolver ID
CaltechAUTHORS:20180912-130453647

Funding

NASA/JPL/Caltech
JPL President and Director's Fund

Dates

Created
2018-09-12
Created from EPrint's datestamp field
Updated
2022-05-17
Created from EPrint's last_modified field