Active Bayesian perception and reinforcement learning

Nathan F. Lepora, Uriel Martinez-Hernandez, Giovanni Pezzulo, Tony J. Prescott

Research output: Chapter in Book/Report/Conference proceedingConference Contribution (Conference Proceeding)

2 Citations (Scopus)

Abstract

In a series of papers, we have formalized an active Bayesian perception approach for robotics based on recent progress in understanding animal perception. However, an issue for applied robot perception is how to tune this method to a task, using: (i) a belief threshold that adjusts the speed-accuracy tradeoff; and (ii) an active control strategy for relocating the sensor e.g. to a preset fixation point. Here we propose that these two variables should be learnt by reinforcement from a reward signal evaluating the decision outcome. We test this claim with a biomimetic fingertip that senses surface curvature under uncertainty about contact location. Appropriate formulation of the problem allows use of multi-armed bandit methods to optimize the threshold and fixation point of the active perception. In consequence, the system learns to balance speed versus accuracy and sets the fixation point to optimize both quantities. Although we consider one example in robot touch, we expect that the underlying principles have general applicability.

Original languageEnglish
Title of host publicationIEEE International Conference on Intelligent Robots and Systems
Pages4735-4740
Number of pages6
DOIs
Publication statusPublished - 2013
Event2013 26th IEEE/RSJ International Conference on Intelligent Robots and Systems: New Horizon, IROS 2013 - Tokyo, Japan
Duration: 3 Nov 20138 Nov 2013

Conference

Conference2013 26th IEEE/RSJ International Conference on Intelligent Robots and Systems: New Horizon, IROS 2013
CountryJapan
CityTokyo
Period3/11/138/11/13

Fingerprint Dive into the research topics of 'Active Bayesian perception and reinforcement learning'. Together they form a unique fingerprint.

Cite this