COCOI: Contact-aware Online Context Inference for Generalizable Non-planar Pushing

TL;DR

COCOI is a novel deep reinforcement learning method that encodes contact-rich interaction contexts online, enabling robots to adapt to diverse dynamics in complex non-planar pushing tasks, demonstrated both in simulation and real-world transfer.

Contribution

This work introduces COCOI, a contact-aware online context inference method that improves adaptability of RL policies in contact-rich manipulation tasks with diverse dynamics.

Findings

COCOI outperforms baseline methods in simulation across various dynamics.

COCOI successfully transfers from simulation to real robot in non-planar pushing.

The method effectively encodes contact-rich interaction contexts online.

Abstract

General contact-rich manipulation problems are long-standing challenges in robotics due to the difficulty of understanding complicated contact physics. Deep reinforcement learning (RL) has shown great potential in solving robot manipulation tasks. However, existing RL policies have limited adaptability to environments with diverse dynamics properties, which is pivotal in solving many contact-rich manipulation tasks. In this work, we propose Contact-aware Online COntext Inference (COCOI), a deep RL method that encodes a context embedding of dynamics properties online using contact-rich interactions. We study this method based on a novel and challenging non-planar pushing task, where the robot uses a monocular camera image and wrist force torque sensor reading to push an object to a goal location while keeping it upright. We run extensive experiments to demonstrate the capability of COCOI in a wide range of settings and dynamics properties in simulation, and also in a sim-to-real transfer scenario on a real robot (Video: https://youtu.be/nrmJYksh1Kc)