Fast setup and adaptation of industrial assembly tasks with force-based exception strategies

Abstract

In this paper we present a method for fast setup and adaptation of desired movement for industrial assembly tasks. Our method is based on adaption of desired movements acquired by Programming by Demonstration (PbD) and fast setup methods such as the Skill Based System (SBS) that are capable to quickly learn sequential skills to perform assembly. The major novelty of the proposed method is the integration of fast setup methods with force based adaptation skills and exception strategies for fast and efficient task execution. To improve the execution of the skill, the learned movements are online adapted according to forces and torques arising during the execution, effectively eliminating the environmental uncertainties. Results show that this approach can be arbitrarily applied to different robotics platforms. We performed tests on the 7-axis Kuka LWR-4 and on the 6- axis UR-5 Universal robot.