Learning to Centralize Dual-Arm Assembly

Abstract

Even though industrial manipulators are widely used in modern manufacturing processes, deployment in unstructured environments remains an open problem. To deal with variety, complexity and uncertainty of real world manipulation tasks a general framework is essential. In this work we want to focus on assembly with humanoid robots by providing a framework for dual-arm peg-in-hole manipulation. As we aim to contribute towards an approach which is not limited to dualarm peg-in-hole, but dual-arm manipulation in general, we keep modeling effort at a minimum. While reinforcement learning has shown great results for single-arm robotic manipulation in recent years, research focusing on dual-arm manipulation is still rare. Solving such tasks often involves complex modeling of interaction between two manipulators and their coupling at a control level. In this paper, we explore the applicability of model-free reinforcement learning to dual-arm manipulation based on a modular approach with two decentralized singlearm controllers and a single centralized policy. We reduce modeling effort to a minimum by using sparse rewards only. We demonstrate the effectiveness of the framework on dualarm peg-in-hole and analyze sample efficiency and success rates for different action spaces. Moreover, we compare results on different clearances and showcase disturbance recovery and robustness, when dealing with position uncertainties. Finally we zero-shot transfer policies trained in simulation to the realworld and evaluate their performance.

Dual-Arm Adversarial Robot Learning

“…With the recent progress in robot learning, many robotic tasks became possible to learn either via imitation or trial-and-error. These tasks range from classical robotics problems such as reaching [1,2] and collision avoidance [3,4] to more complex tasks involving locomotion [5,6,7] and manipulation [8,9,10,11]. However, most methods that solve these tasks are either tested in simulation, assume to have access to a perfect estimate of the environment' state, or strongly restrict the interaction between the robot and its surrounding based on predefined heuristics or physical constraints.…”