Sim-to-Real Transfer of Robot Learning with Variable Length Inputs

TL;DR

This paper introduces a modular RL framework using deep sets encoding for variable-length inputs, enabling rapid sim-to-real transfer in robotic object sorting tasks with minimal training time.

Contribution

It presents a novel combination of deep sets encoding with modular RL to handle variable-length representations, improving sim-to-real transfer efficiency.

Findings

Effective policies learned within minutes of simulation

Policies directly deployed on robots without additional training

Demonstrated generalization to unseen task variations

Abstract

Current end-to-end deep Reinforcement Learning (RL) approaches require jointly learning perception, decision-making and low-level control from very sparse reward signals and high-dimensional inputs, with little capability of incorporating prior knowledge. This results in prohibitively long training times for use on real-world robotic tasks. Existing algorithms capable of extracting task-level representations from high-dimensional inputs, e.g. object detection, often produce outputs of varying lengths, restricting their use in RL methods due to the need for neural networks to have fixed length inputs. In this work, we propose a framework that combines deep sets encoding, which allows for variable-length abstract representations, with modular RL that utilizes these representations, decoupling high-level decision making from low-level control. We successfully demonstrate our approach on the robot manipulation task of object sorting, showing that this method can learn effective policies within mere minutes of highly simplified simulation. The learned policies can be directly deployed on a robot without further training, and generalize to variations of the task unseen during training.