2014 IEEE/RSJ International Conference on Intelligent Robots and Systems 2014
DOI: 10.1109/iros.2014.6942756
|View full text |Cite
|
Sign up to set email alerts
|

Abstract: We present a method to specify tasks and synthesise cost-optimal policies for Markov decision processes using co-safe linear temporal logic. Our approach incorporates a dynamic task handling procedure which allows for the addition of new tasks during execution and provides the ability to replan an optimal policy on-the-fly. This new policy minimises the cost to satisfy the conjunction of the current tasks and the new one, taking into account how much of the current tasks has already been executed. We illustrat… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
1
1

Citation Types

0
40
0

Year Published

2015
2015
2023
2023

Publication Types

Select...
3
3
2

Relationship

1
7

Authors

Journals

citations
Cited by 64 publications
(40 citation statements)
references
References 14 publications
(21 reference statements)
0
40
0
Order By: Relevance
“…LTL is being increasingly used to formulate task specifications for robotics and automation [16] and there have been some efforts to solve MDPs with LTL constraints. Lacerda et al [13] solve a planning problem with LTL specification in a cost optimal form. Ding et al [8] focused on long term surveillance problems.…”
Section: State Of the Artmentioning
confidence: 99%
“…LTL is being increasingly used to formulate task specifications for robotics and automation [16] and there have been some efforts to solve MDPs with LTL constraints. Lacerda et al [13] solve a planning problem with LTL specification in a cost optimal form. Ding et al [8] focused on long term surveillance problems.…”
Section: State Of the Artmentioning
confidence: 99%
“…We apply an approach based on [20], which allows planning for tasks specified in linear temporal logic (LTL), using Markov decision process (MDP) models of the environment. We present this approach mostly as a case study of how our frequency-based predictions can be employed; our work is not limited to this particular planning technique.…”
Section: Spatio-temporal Topological Representationsmentioning
confidence: 99%
“…In [20], it is shown how, given a co-safe LTL formula ϕ and a cost function defined over state-action pairs of the MDP (in our case, such function would be the expected time to navigate between two nodes in the environment), one can create policies that minimize the accumulated cost to generate a trace of the system that satisfies ϕ. Broadly speaking, LTL allows for the specification of goals that are not simply reaching a given target node in the environment, but can be temporally extended goals that require, for example, a set of nodes to be visited in a given order, or to visit a given node while avoiding a set of forbidden nodes.…”
Section: Spatio-temporal Topological Representationsmentioning
confidence: 99%
See 1 more Smart Citation
“…During the last few years, several methods have been developed for deterministic, e.g., [1][2] [3], non-deterministic, e.g., [4][5] [6] and stochastic systems, e.g., [7] [8]. These methods specify properties using temporal logics such as linear temporal logic (LTL) and computation tree logic (CTL).…”
Section: Introductionmentioning
confidence: 99%