A multi-range vision strategy for autonomous offroad navigation

Raia Hadsell, Ayse Erkan, Pierre Sermanet, Jan Ben, Koray Kavukcuoglu, Urs Muller, Yann LeCun

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Vision-based navigation and obstacle detection must be sophisticated in order to perform well in complicated and diverse terrain, but that complexity comes at the expense of increased system latency between image capture and actuator signals. Increased latency, or a longer control loop, degrades the reactivity of the robot. We present a navigational framework that uses a self-supervised, learningbased obstacle detector without paying a price in latency and reactivity. A long-range obstacle detector uses online learning to accurately see paths and obstacles at ranges up to 30 meters, while a fast, short-range obstacle detector avoids obstacles at up to 5 meters. The learning-based long-range module is discussed in detail, and field experiments are described which demonstrate the success of the overall system.

Original languageEnglish (US)
Title of host publicationProceedings of the 13th IASTED International Conference on Robotics and Applications, RA 2007 and Proceedings of the IASTED International Conference on Telematics
Pages457-463
Number of pages7
StatePublished - 2007
Event13th IASTED International Conference on Robotics and Applications, RA 2007 and Proceedings of the IASTED International Conference on Telematics - Wurzburg, Germany
Duration: Aug 29 2007Aug 31 2007

Other

Other13th IASTED International Conference on Robotics and Applications, RA 2007 and Proceedings of the IASTED International Conference on Telematics
CountryGermany
CityWurzburg
Period8/29/078/31/07

Fingerprint

Obstacle detectors
Navigation
Actuators
Robots
Experiments

Keywords

  • LAGR
  • Latency
  • Learning
  • Navigation
  • Offroad
  • Vision

ASJC Scopus subject areas

  • Computer Science Applications
  • Computer Vision and Pattern Recognition

Cite this

Hadsell, R., Erkan, A., Sermanet, P., Ben, J., Kavukcuoglu, K., Muller, U., & LeCun, Y. (2007). A multi-range vision strategy for autonomous offroad navigation. In Proceedings of the 13th IASTED International Conference on Robotics and Applications, RA 2007 and Proceedings of the IASTED International Conference on Telematics (pp. 457-463)

A multi-range vision strategy for autonomous offroad navigation. / Hadsell, Raia; Erkan, Ayse; Sermanet, Pierre; Ben, Jan; Kavukcuoglu, Koray; Muller, Urs; LeCun, Yann.

Proceedings of the 13th IASTED International Conference on Robotics and Applications, RA 2007 and Proceedings of the IASTED International Conference on Telematics. 2007. p. 457-463.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Hadsell, R, Erkan, A, Sermanet, P, Ben, J, Kavukcuoglu, K, Muller, U & LeCun, Y 2007, A multi-range vision strategy for autonomous offroad navigation. in Proceedings of the 13th IASTED International Conference on Robotics and Applications, RA 2007 and Proceedings of the IASTED International Conference on Telematics. pp. 457-463, 13th IASTED International Conference on Robotics and Applications, RA 2007 and Proceedings of the IASTED International Conference on Telematics, Wurzburg, Germany, 8/29/07.
Hadsell R, Erkan A, Sermanet P, Ben J, Kavukcuoglu K, Muller U et al. A multi-range vision strategy for autonomous offroad navigation. In Proceedings of the 13th IASTED International Conference on Robotics and Applications, RA 2007 and Proceedings of the IASTED International Conference on Telematics. 2007. p. 457-463
Hadsell, Raia ; Erkan, Ayse ; Sermanet, Pierre ; Ben, Jan ; Kavukcuoglu, Koray ; Muller, Urs ; LeCun, Yann. / A multi-range vision strategy for autonomous offroad navigation. Proceedings of the 13th IASTED International Conference on Robotics and Applications, RA 2007 and Proceedings of the IASTED International Conference on Telematics. 2007. pp. 457-463
@inproceedings{7ec89bcfadfe4242b1f8282a8e84c050,
title = "A multi-range vision strategy for autonomous offroad navigation",
abstract = "Vision-based navigation and obstacle detection must be sophisticated in order to perform well in complicated and diverse terrain, but that complexity comes at the expense of increased system latency between image capture and actuator signals. Increased latency, or a longer control loop, degrades the reactivity of the robot. We present a navigational framework that uses a self-supervised, learningbased obstacle detector without paying a price in latency and reactivity. A long-range obstacle detector uses online learning to accurately see paths and obstacles at ranges up to 30 meters, while a fast, short-range obstacle detector avoids obstacles at up to 5 meters. The learning-based long-range module is discussed in detail, and field experiments are described which demonstrate the success of the overall system.",
keywords = "LAGR, Latency, Learning, Navigation, Offroad, Vision",
author = "Raia Hadsell and Ayse Erkan and Pierre Sermanet and Jan Ben and Koray Kavukcuoglu and Urs Muller and Yann LeCun",
year = "2007",
language = "English (US)",
isbn = "9780889866850",
pages = "457--463",
booktitle = "Proceedings of the 13th IASTED International Conference on Robotics and Applications, RA 2007 and Proceedings of the IASTED International Conference on Telematics",

}

TY - GEN

T1 - A multi-range vision strategy for autonomous offroad navigation

AU - Hadsell, Raia

AU - Erkan, Ayse

AU - Sermanet, Pierre

AU - Ben, Jan

AU - Kavukcuoglu, Koray

AU - Muller, Urs

AU - LeCun, Yann

PY - 2007

Y1 - 2007

N2 - Vision-based navigation and obstacle detection must be sophisticated in order to perform well in complicated and diverse terrain, but that complexity comes at the expense of increased system latency between image capture and actuator signals. Increased latency, or a longer control loop, degrades the reactivity of the robot. We present a navigational framework that uses a self-supervised, learningbased obstacle detector without paying a price in latency and reactivity. A long-range obstacle detector uses online learning to accurately see paths and obstacles at ranges up to 30 meters, while a fast, short-range obstacle detector avoids obstacles at up to 5 meters. The learning-based long-range module is discussed in detail, and field experiments are described which demonstrate the success of the overall system.

AB - Vision-based navigation and obstacle detection must be sophisticated in order to perform well in complicated and diverse terrain, but that complexity comes at the expense of increased system latency between image capture and actuator signals. Increased latency, or a longer control loop, degrades the reactivity of the robot. We present a navigational framework that uses a self-supervised, learningbased obstacle detector without paying a price in latency and reactivity. A long-range obstacle detector uses online learning to accurately see paths and obstacles at ranges up to 30 meters, while a fast, short-range obstacle detector avoids obstacles at up to 5 meters. The learning-based long-range module is discussed in detail, and field experiments are described which demonstrate the success of the overall system.

KW - LAGR

KW - Latency

KW - Learning

KW - Navigation

KW - Offroad

KW - Vision

UR - http://www.scopus.com/inward/record.url?scp=56149097142&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=56149097142&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:56149097142

SN - 9780889866850

SP - 457

EP - 463

BT - Proceedings of the 13th IASTED International Conference on Robotics and Applications, RA 2007 and Proceedings of the IASTED International Conference on Telematics

ER -