Publication
Learning Visual Representations for Interactive Systems
Justus H. Piater; Sébastien Jodogne; Renaud Detry; Dirk Kraft; Norbert Krüger; Oliver Krömer; Jan Peters
In: Cédric Pradalier; Roland Siegwart; Gerhard Hirzinger (Hrsg.). Robotics Research - The 14th International Symposium ISRR. International Symposium of Robotics Research (ISRR-2009), August 31 - September 3, Lucerne, Switzerland, Pages 399-416, Springer Tracts in Advanced Robotics (STAR), Vol. 70, Springer, 2009.
Abstract
We describe two quite different methods for associating action parameters to visual percepts. Our RLVC algorithm performs reinforcement learning directly on the visual input space. To make this very large space manageable, RLVC interleaves the reinforcement learner with a supervised classification algorithm that seeks to split perceptual states so as to reduce perceptual aliasing. This results in an adaptive discretization of the perceptual space based on the presence or absence of visual features. Its extension RLJC also handles continuous action spaces. In contrast to the minimalistic visual representations produced by RLVC and RLJC, our second method learns structural object models for robust object detection and pose estimation by probabilistic inference. To these models, the method associates grasp experiences autonomously learned by trial and error. These experiences form a nonparametric representation of grasp success likelihoods over gripper poses, which we call a . Thus, object detection in a novel scene simultaneously produces suitable grasping options.