Online path planning for industrial robots in varying environments using the curve shortening flow method

Abstract

Handling tasks with robots which are interacting in a shared workspace provide a high risk of collision. A new approach, the so-called "Curve Shortening Flow Method", enables a collision-free path planning method for robots within a varying environment on basis of a workspace model. Thereby a global path planning method based on geometrical curvature flow is combined with the locally and reactively acting potential field method in order to describe the obstacles' influence, whereby the obstacles are modeled as a repulsive field, which displaces the robot's planned path. This path can be formulated as a time-dependant partial differential equation, which can be solved very efficiently using explicit numerical time integration. First results show that this online method is able to generate collision-free paths robustly in real-time. © 2011 Springer-Verlag.