Active Vision Techniques for Uncalibrated Hand-Eye Coordination and Motor Control

We present a promising approach for combined visual model acquisition and visual servo control. The approach differs from previous work in that a full coupled Jacobian is estimated on-line without any prior models, introduction of special calibration movements, or the use of absolute world coordinate systems. A trust region controller allow stable and convergent control when the underlying visual-motor model is highly non-linear. On the high level actions and tasks are coded in terms of desired general perceptions rather than motor sequences. We argue that our vision space approach is particularly suited for easy teaching/programming of a robot. For instance a task can be taught by supplying an image sequence illustrating it. The resulting robot behavior is robust to changes in the environment, dynamically adjusting the motor control rules in response to environmental variation. We provide an extensive experimental evaluation of the positioning precision and convergence of the visual s..