Home // ICAS 2020, The Sixteenth International Conference on Autonomic and Autonomous Systems // View article

Contingent Planning Using Counter-Examples from a Conformant Planner

Authors:
Sébastien Piedade
Alban Grastien
Charles Lesire
Guillaume Infantes

Keywords: contingent planning; autonomous decision-making; uncertainty

Abstract:
Decision-making for autonomous robots in real world applications has to manage uncertainties in order to efficiently accomplish a mission. Some planning methods deal with uncertainty by improving the robustness of the plan embedded in the robot. In this paper, we propose a novel approach to one of these methods, contingent planning. Most of the existing approaches are limited by the computation complexity and the quality of the solutions they return. To deal with these limitations, we propose to limit the number of observations in the plan as observations involve an important cost in computation time and energy. The originality of our approach is that our contingent planner uses an underlying conformant planner, i.e., a planner that is not allowed to make observations, to compute conformant subplans and insert observations between conformant subplans only when a conformant plan cannot be computed. We evaluate this approach by comparing its results with respect to Contingent-FF (Contingent Fast-Forward), a well known contingent planner, on a set of benchmarks. This comparison reveals that, even if our approach has some limitations, as it is not complete, it works quite effectively in terms of solution quality on classic benchmarks of the planning community.

Pages: 16 to 22

Copyright: Copyright (c) IARIA, 2020

Publication date: September 27, 2020

Published in: conference

ISSN: 2308-3913

ISBN: 978-1-61208-787-0

Location: Lisbon, Portugal

Dates: from September 27, 2020 to October 1, 2020