Deep Reinforcement Learning for Robotic Control in High-Dexterity Assembly Tasks - A Reward Curriculum Approach

IEEE International Conference on Robotic Computing (IRC), pp. 35–42

Abstract

For years, the fully-automated robotic assembly has been a highly sought-after technology in large-scale manufacturing. Yet it still struggles to find widespread implementation in industrial environments. Traditional programming has so far proven to be insufficient to provide the required flexibility and dexterity to solve complex assembly tasks. Although research in robotic control using deep reinforcement learning (DRL) advances quickly, the transfer to real-world applications in industrial settings is lagging behind. In this study, we apply DRL for robotic motion control at the use-case of a multi-body contact automotive assembly task and focus on optimizing the final performance on the real-world setup. We propose a reward-curriculum learning approach in combination with domain randomization to obtain both force-sensitivity and generalizability. We train an agent exclusively in simulation and successfully perform the Sim-to-Real transfer. Finally, we evaluate the controller’s performance and robustness on an industrial setup and reflect its adherence to the high automotive production standards.

Authors 5

1. Lars Leyendecker Aachen Fraunhofer Institute for Production Technology IPT

   Fraunhofer Institute for Production Technology IPT

   Affiliation as printed

   Fraunhofer Institute for Production Technology IPT,52074 Aachen,Germany

   Fraunhofer Institute for Production Technology IPT, 52074 Aachen, Germany

2. Hans Aoyang Zhou Aachen Institute of Information Management in Mechanical Engineering

   RWTH Aachen University

   Affiliation as printed

   RWTH Aachen University,Institute of Information Management in Mechanical Engineering,52068 Aachen,Germany

   Institute of Information Management in Mechanical Engineering, RWTH Aachen University, 52068 Aachen, Germany

3. Владимир Самсонов Aachen Institute of Information Management in Mechanical Engineering

   RWTH Aachen University

   Affiliation as printed

   RWTH Aachen University,Institute of Information Management in Mechanical Engineering,52068 Aachen,Germany

   Institute of Information Management in Mechanical Engineering, RWTH Aachen University, 52068 Aachen, Germany

4. Marius Rittstieg

   BMW (Germany) · BMW Group (Germany)

   Affiliation as printed

   Digitalization at BMW Group,Departments of Data Analytics and Innovation,80807 Munich,Germany

   Departments of Data Analytics and Innovation, Digitalization at BMW Group, 80807 Munich, Germany

5. Daniel Lütticke Aachen Institute of Information Management in Mechanical Engineering

   RWTH Aachen University

   Affiliation as printed

   RWTH Aachen University,Institute of Information Management in Mechanical Engineering,52068 Aachen,Germany

   Institute of Information Management in Mechanical Engineering, RWTH Aachen University, 52068 Aachen, Germany