6 Future Developments
The current system is based upon matching the positions and
orientations of two planes - the gripper and target.
Since the gripper and target planes
must be visible in both images at all times, neither can rotate
through more than about 120 degrees.
We intend to develop the system to track the gripper using a
three-dimensional rigid model [6],
drawing information
from both images simultaneously. We also aim to identify and track more than
one of the surfaces of the object to be grasped, both for 3-D
tracking and also for analysis of its 3D shape.
We plan to equip such a system with a grasp planner that uses
the relative positions, sizes and orientations of the visible surfaces on the
object, to direct the robot to grasp unmodelled objects in a suitable
way.
Next
Contents