CaltechAUTHORS
  A Caltech Library Service

Dimensionality Reduction Using Automatic Supervision for Vision-Based Terrain Learning

Angelova, Anelia and Matthies, Larry and Helmick, Daniel and Perona, Pietro (2008) Dimensionality Reduction Using Automatic Supervision for Vision-Based Terrain Learning. In: Robotics: Science and Systems III. MIT Press , Cambridge, MA, pp. 225-232. ISBN 9780262255868. https://resolver.caltech.edu/CaltechAUTHORS:20150903-125600893

[img] PDF - Accepted Version
See Usage Policy.

1346Kb

Use this Persistent URL to link to this item: https://resolver.caltech.edu/CaltechAUTHORS:20150903-125600893

Abstract

This paper considers the problem of learning to recognize different terrains from color imagery in a fully automatic fashion, using the robot’s mechanical sensors as supervision. We present a probabilistic framework in which the visual information and the mechanical supervision interact to learn the available terrain types. Within this framework, a novel supervised dimensionality reduction method is proposed, in which the automatic supervision provided by the robot helps select better lower dimensional representations, more suitable for the discrimination task at hand. Incorporating supervision into the dimensionality reduction process is important, as some terrains might be visually similar but induce very different robot mobility. Therefore, choosing a lower dimensional visual representation adequately is expected to improve the vision-based terrain learning and the final classification performance. This is the first work that proposes automatically supervised dimensionality reduction in a probabilistic framework using the supervision coming from the robot’s sensors. The proposed method stands in between methods for reasoning under uncertainty using probabilistic models and methods for learning the underlying structure of the data. The proposed approach has been tested on field test data collected by an autonomous robot while driving on soil, gravel and asphalt. Although the supervision might be ambiguous or noisy, our experiments show that it helps build a more appropriate lower dimensional visual representation and achieves improved terrain recognition performance compared to unsupervised learning methods.


Item Type:Book Section
Related URLs:
URLURL TypeDescription
http://www.roboticsproceedings.org/rss03/p29.htmlOrganizationConference Paper
http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=6280116&PublisherIEEE Xplore Chapter
ORCID:
AuthorORCID
Perona, Pietro0000-0002-7583-5809
Additional Information:© 2008 Massachusetts Institute of Technology. This research was carried out by the Jet Propulsion Laboratory, California Institute of Technology, under a contract with NASA, with funding from the Mars Technology Program. We thank Navid Serrano and the anonymous reviewers for their very useful comments on the paper.
Funders:
Funding AgencyGrant Number
NASAUNSPECIFIED
Record Number:CaltechAUTHORS:20150903-125600893
Persistent URL:https://resolver.caltech.edu/CaltechAUTHORS:20150903-125600893
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:60052
Collection:CaltechAUTHORS
Deposited By: Caroline Murphy
Deposited On:09 Mar 2020 14:56
Last Modified:03 Oct 2019 08:53

Repository Staff Only: item control page