Abstract (100 words or less): |  | We will present an iterative algorithm that numerically calculates the boundary value curve at discretized time instants for a constrained variational problem. The idea is to use proper integration and approximation schemes so that the cost function can be represented by a set of discretized positions (and Lagrange multipliers), while the characteristics of the original optimization system are still preserved.
Our algorithm can be applied on motion planning for robot manipulator performing a compliant motion task, e.g., the robot's end-effector is required to move on an environmental surface. We will propose a two-step calculation scheme aiming to find a joint trajectory fulfilling a synthetically geometrical optimality covering both itself and its resulting end-effector path.
An example of motion planning for a 4-DOF robot will be given at the end of this talk, some 3D-animation and experimentation demos will be shown as well. |