Deep Reinforcement Learning for Robotic Manipulation with Asynchronous Off-Policy Updates

TL;DR

This paper presents an off-policy deep reinforcement learning algorithm that efficiently trains neural network policies for complex 3D robotic manipulation tasks, including real-world door opening, by parallelizing updates across multiple robots.

Contribution

It introduces an asynchronous off-policy deep Q-learning method that scales to real robots and complex tasks without prior demonstrations or manual representations.

Findings

Successfully learned 3D manipulation skills in simulation.

Achieved real-world door opening without demonstrations.

Reduced training time through parallel asynchronous updates.

Abstract

Reinforcement learning holds the promise of enabling autonomous robots to learn large repertoires of behavioral skills with minimal human intervention. However, robotic applications of reinforcement learning often compromise the autonomy of the learning process in favor of achieving training times that are practical for real physical systems. This typically involves introducing hand-engineered policy representations and human-supplied demonstrations. Deep reinforcement learning alleviates this limitation by training general-purpose neural network policies, but applications of direct deep reinforcement learning algorithms have so far been restricted to simulated settings and relatively simple tasks, due to their apparent high sample complexity. In this paper, we demonstrate that a recent deep reinforcement learning algorithm based on off-policy training of deep Q-functions can scale to complex 3D manipulation tasks and can learn deep neural network policies efficiently enough to train on real physical robots. We demonstrate that the training times can be further reduced by parallelizing the algorithm across multiple robots which pool their policy updates asynchronously. Our experimental evaluation shows that our method can learn a variety of 3D manipulation skills in simulation and a complex door opening skill on real robots without any prior demonstrations or manually designed representations.