Robot Air Hockey: A Manipulation Testbed for Robot Learning with Reinforcement Learning

TL;DR

This paper introduces a dynamic robot air hockey testbed for reinforcement learning that supports diverse tasks, sim-to-real transfer, and evaluation of learning algorithms in interactive, fast-paced environments.

Contribution

It presents a novel, versatile RL testbed based on robot air hockey with multiple task types and transfer domains, enabling comprehensive assessment of RL methods.

Findings

Testbed supports varied RL algorithms including behavior cloning and offline RL.

Demonstrates effective sim-to-real transfer across multiple domains.

Enables evaluation of RL in fast-paced, interactive tasks.

Abstract

Reinforcement Learning is a promising tool for learning complex policies even in fast-moving and object-interactive domains where human teleoperation or hard-coded policies might fail. To effectively reflect this challenging category of tasks, we introduce a dynamic, interactive RL testbed based on robot air hockey. By augmenting air hockey with a large family of tasks ranging from easy tasks like reaching, to challenging ones like pushing a block by hitting it with a puck, as well as goal-based and human-interactive tasks, our testbed allows a varied assessment of RL capabilities. The robot air hockey testbed also supports sim-to-real transfer with three domains: two simulators of increasing fidelity and a real robot system. Using a dataset of demonstration data gathered through two teleoperation systems: a virtualized control environment, and human shadowing, we assess the testbed with behavior cloning, offline RL, and RL from scratch.