Learning in Robot Vision Directed Reaching: A Comparison of Methods

Abstract

Four neural network algorithms were examined for their ability to adaptively associate stereo camera coordinates with joint positions of a three degree of freedom manipulator arm in a 3D reaching task. Given reasonable numbers of training exemplars for an implementation in real hardware, all networks trained to significant errors. Two secondary error correction procedures were then tested. Both further reduced errors, but one method that depended on continuous visual and proprioceptive feedback to train a small set of associative weights that correlated joint and camera velocities was especially effective in eliminating errors. Stereo pan, tilt and vergence information was used to direct ballistic reaching, but relative depth information, was used for the visual feedback of end-effector velocity in the second error correction method.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Nov 01, 1994
Accession Number
ADA483612

Entities

People

  • Hoa G. Nguyen
  • Michael R. Blackburn

Organizations

  • Naval Command, Control and Ocean Surveillance Center

Tags

Communities of Interest

  • Autonomy

DTIC Thesaurus Topics

  • Accuracy
  • Algorithms
  • Camera Controls
  • Cameras
  • Computer Programming
  • Control
  • Control Systems
  • Detectors
  • Feedback
  • Learning
  • Manipulators
  • Neural Networks
  • Ocean Surveillance
  • Simulations
  • Stereo Cameras
  • Three Dimensional
  • Training

Readers

  • Brain and Cognitive Science; Experimental Psychology; Cognitive Neuroscience
  • Geodesy
  • Robotics and Automation.

Technology Areas

  • AI & ML
  • AI & ML - Autonomous Systems
  • AI & ML - Bayesian Inference
  • AI & ML - Neural Networks
  • Autonomy