Reinforcement Learning on Variable Impedance Controller for High-Precision Robotic Assembly

TLDR

Precise robotic manipulation is essential in industry, and RL offers a way to acquire such skills autonomously. The study aims to integrate operational‑space force/torque feedback into RL to enable high‑precision robotic assembly. The authors combine RL with an operational‑space force controller, use a neural network architecture that generalizes across environments, and test the method on the Siemens Robot Learning Challenge for precise gear‑wheel assembly.

Abstract

Precise robotic manipulation skills are desirable in many industrial settings, reinforcement learning (RL) methods hold the promise of acquiring these skills autonomously. In this paper, we explicitly consider incorporating operational space force/torque information into reinforcement learning; this is motivated by humans heuristically mapping perceived forces to control actions, which results in completing high-precision tasks in a fairly easy manner. Our approach combines RL with force/torque information by incorporating a proper operational space force controller; where we also exploit different ablations on processing this information. Moreover, we propose a neural network architecture that generalizes to reasonable variations of the environment. We evaluate our method on the open-source Siemens Robot Learning Challenge, which requires precise and delicate force-controlled behavior to assemble a tight-fit gear wheel set.