Sim-to-Real Deep Reinforcement Learning with Manipulators for Pick-and-place

TL;DR

This paper introduces a self-supervised vision-based deep reinforcement learning approach for robotic pick-and-place tasks that can be directly transferred from simulation to real-world without fine-tuning, achieving high success rates.

Contribution

The paper presents a novel height-sensitive policy and a self-supervised training method enabling direct sim-to-real transfer for robotic pick-and-place tasks.

Findings

Achieves 90% success rate on real objects without fine-tuning.

Effectively handles crowded and stacked objects in real environments.

Demonstrates successful transfer of simulation-trained models to real robots.

Abstract

When transferring a Deep Reinforcement Learning model from simulation to the real world, the performance could be unsatisfactory since the simulation cannot imitate the real world well in many circumstances. This results in a long period of fine-tuning in the real world. This paper proposes a self-supervised vision-based DRL method that allows robots to pick and place objects effectively and efficiently when directly transferring a training model from simulation to the real world. A height-sensitive action policy is specially designed for the proposed method to deal with crowded and stacked objects in challenging environments. The training model with the proposed approach can be applied directly to a real suction task without any fine-tuning from the real world while maintaining a high suction success rate. It is also validated that our model can be deployed to suction novel objects in a real experiment with a suction success rate of 90\% without any real-world fine-tuning. The experimental video is available at: https://youtu.be/jSTC-EGsoFA.