Home // ADAPTIVE 2015, The Seventh International Conference on Adaptive and Self-Adaptive Systems and Applications // View article

Adaptation for Active Perception, Robustness from Adaptation to Context

Authors:
Paul Robertson
Andreas Hofmann

Keywords: Active Perception, POMDP, Belief State, Planning, Context, Adaptation

Abstract:
Existing machine perception systems are brittle and inflexible, and therefore cannot adapt well to environment uncertainty. Natural perception always occurs in support of an activity that provides context. In our approach, we use context to adapt the perceptual aperatus providing robustness and resilience to noise among other benefits. Evidence supports the view that context driven adaptation occurs in natural perception systems including human vision. This Active Perception approach prioritizes the system’s overall goals, so that perception and situation awareness are well integrated with actions to focus all efforts on these goals in an optimal manner. We use a Partially Observable Markov Decision Process (POMDP) framework, but do not attempt to compute a comprehensive control policy, as this is intractible for practical problems. Instead, we employ Belief State Planning to compute point solutions from an initial state to a goal state set. This approach automatically adapts the perception data flow processes, and generates action sequences for sensing operations that reduce uncertainty in the belief state, and ultimately achieve the goal state set. Our early results described in this paper demonstrate the feasibility of this approach using a restriced set of actions.

Pages: 64 to 73

Copyright: Copyright (c) IARIA, 2015

Publication date: March 22, 2015

Published in: conference

ISSN: 2308-4146

ISBN: 978-1-61208-391-9

Location: Nice, France

Dates: from March 22, 2015 to March 27, 2015