Transferring End-to-End Visuomotor Control from Simulation to Real World for a Multi-Stage Task

TL;DR

This paper demonstrates a simple, scalable end-to-end visuomotor control method that transfers from simulation to real-world multi-stage tasks without real-world training data, achieving robust generalization in complex environments.

Contribution

The authors introduce a domain randomization-based approach for transferring simulation-trained control policies to real robots for multi-stage tasks without real-world data.

Findings

Successfully accomplished multi-stage manipulation in real world

Generalized to dynamic lighting and distractors

Effective for long-horizon tasks

Abstract

End-to-end control for robot manipulation and grasping is emerging as an attractive alternative to traditional pipelined approaches. However, end-to-end methods tend to either be slow to train, exhibit little or no generalisability, or lack the ability to accomplish long-horizon or multi-stage tasks. In this paper, we show how two simple techniques can lead to end-to-end (image to velocity) execution of a multi-stage task, which is analogous to a simple tidying routine, without having seen a single real image. This involves locating, reaching for, and grasping a cube, then locating a basket and dropping the cube inside. To achieve this, robot trajectories are computed in a simulator, to collect a series of control velocities which accomplish the task. Then, a CNN is trained to map observed images to velocities, using domain randomisation to enable generalisation to real world images. Results show that we are able to successfully accomplish the task in the real world with the ability to generalise to novel environments, including those with dynamic lighting conditions, distractor objects, and moving objects, including the basket itself. We believe our approach to be simple, highly scalable, and capable of learning long-horizon tasks that have until now not been shown with the state-of-the-art in end-to-end robot control.