Three-Dimensional Camera Space Manipulation

Abstract

This article extends to three-dimensional tasks the method of camera space manipulation. A minimum of two cameras is required to place points on end effectors (or objects in their grasp) of n-degree-of-freedom manipulators relative to other bodies. This is accomplished using a sequential estimation scheme that permits placement of these points in each of the two-dimensional image planes of monitoring cameras. A precise and robust manipulation strategy that is compatible with "real time" results. Simulations are used that show the method to be insensitive to two particular kinds of model error- unmodeled elastic deflection and unmodeled camera distortion. The method is tested experimentally with a three- dimensional point placement task. It is then generalized to rigid body placement tasks and illustrated with experiments involving the positioning of one rigid body on a second. An appendix details the unfolding of one such experimental maneuver at several junctures in the visual data-collecting process.