Toward a faster mapping onto configuration space for obstacle avoidance in robotics

Abstract

As robotic manipulators become used more widely in industry, there may be a desire to include them in more complicated environments. In such workspaces, obstacle avoidance will be an element of the hierarchical control system. Most of the algorithms proposed for obstacle avoidance require that the robot be represented as a point in a configuration space or C-Space. Obstacles must also be represented in the configuration space. In this work, a faster approach to the mapping of obstacles into configuration space is presented. In a plane, the scheme is based on constraining the manipulator to contact the boundary of the obstacle in the workspace. The solutions to the equations of constraint define closed curves which represent the boundary of the configuration space obstacle. The Jacobian of the mapping for a two degree-of-freedom revolute manipulator is evaluated for use in estimating the new mapping of a displaced obstacle. Finally, the proposed planar mapping is extended for obstacles in the workspace of a three degree-of-freedom spatial revolute manipulator.