A Probabilistic Programming by Demonstration Framework Handling Constraints in Joint Space and Task Space

We present a probabilistic architecture for solving generically the problem of extracting the task constraints through a Programming by Demonstration (PbD) framework and for generalizing the acquired knowledge to various situations. We propose an approach based on Gaussian Mixture Regression (GMR) to find automatically a controller for the robot reproducing the essential characteristics of the skill by handling simultaneously constraints in joint space and in task space. Experiments with two 5-DOFs Katana robots are then presented with two manipulation tasks consisting of handling and displacing a set of objects.