Visual navigation: image profiles for odometry and control

ABSTRACT

This paper investigates the extent to which image profiles, pixel-intensity sums across subsets of a video stream, offer a basis for autonomous robotics. Building on previous work that uses image profiles for odometry, we introduce an improved algorithm for odometric estimation based on visual input alone. In addition, we extend prior results by showing how image profiles can support control and exploration tasks. We validate these new approaches through results implemented atop a low-cost vision-only vehicle consisting of a laptop and an iRobot Create.

REFERENCES

• 1: Milford, M. J. and Wyeth, G. F. Single Camera Vision-only SLAM on a Suburban Road Network, in Proceedings, 2008 Int. Conf. on Robotics and Automation, Pasadena, CA, USA, May 19--23, pp. 3684--3689, 2008.
• 2: Available from www.cs.hmc.edu/twiki/bin/view/Robotics

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
      I.2.9 Robotics
          Subjects: Sensors

Additional Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
        I.2.10 Vision and Scene Understanding
            Subjects: Perceptual reasoning; Representations, data structures, and transforms; Motion

General Terms:
Design, Experimentation, Performance

Keywords:
interest operators, vision-based robot control, visual odometry