Path planning for a manipulator with constraints on orientation

Abstract

This paper discusses the work undertaken at the U.K. National Advanced Robotics Research Centre in the area of path planning. Experiments have been undertaken with the Best First Planner (BFP) and Random Path Planner (RPP) algorithms developed at Stanford University by Barraquand and Latombe. These look particularly suitable for solving a wide range of motion planning problems and for providing the basic mechanism for path planning in an advanced robotic architecture. From this basic mechanism higher level functional primitives have been developed to enable paths to be planned with a pre-defined orientation of the end effector at the goal location or maintaining an orientation throughout the whole path. These functions are achieved through placing additional constraints on the search through configuration space and modifying the workspace potential.