Affordable Access

Designing visually servoed tracking to augment camera teleoperators

Authors
Publication Date
Keywords
  • Engineering
  • Robots
  • Servomechanisms
Disciplines
  • Design

Abstract

Robots have now far more impact in humans life then ten years ago. Vacuum cleaning robots are already well known. Making today’s robots to work unassisted requires appropriate visual servoing architecture. In the past, a lot of efforts were directed towards designing controllers that relies exclusively on image data. Still most robots are servoed kinematically using joint data. Visual servoing architecture has applications not only in robotics. Video cameras are often mounted on platforms that can move like rovers, booms, gantries and aircrafts. People can operate such platforms to capture desired views of a scene or a target. To avoid collisions, with the environment and occlusions, such platforms demands much skill. Visual-servoing some degrees-of-freedom may reduce the operator burden and improve tracking. We call this concept human-in-the-loop visual servoing. Human-in-the-loop systems involve an operator who manipulates a device for desired tasks based on feedback from the device and environment. For example, devices like rovers gantries and aircrafts possess a video camera. The task is to control maneuver the vehicle and position the camera to obtain desired fields of view. To overcome joint limits, avoid collisions and ensure occlusion-free views, these devices are typically equipped with redundant degrees-of-freedom. Tracking moving subjects with such systems is a challenging task and requires a well skilled operator. In this approach, we use computer vision techniques to visually servo the camera. The net effect is that the operator just focuses on safely manipulating the boom and dolly while computer-control automatically servos the camera.

There are no comments yet on this publication. Be the first to share your thoughts.