DexPBT: Scaling up Dexterous Manipulation for Hand-Arm Systems with Population Based Training

Aleksei Petrenko
University of Southern California
Arthur Allshire
University of Toronto
Gavriel State
Ankur Handa
Viktor Makoviychuk
Paper Website

Paper ID 37

Session 5. Simulation and Sim2Real

Poster Session Wednesday, July 12

Poster 5

Abstract: In this work, we propose algorithms and methods that enable learning dexterous object manipulation using simulated one- or two-armed robots equipped with multi-fingered hand end-effectors. Using a parallel GPU-accelerated physics simulator (Isaac Gym), we implement challenging tasks for these robots, including regrasping, grasp-and-throw, and object reorientation. To solve these problems we introduce a decentralized Population-Based Training (PBT) algorithm that allows us to massively amplify the exploration capabilities of deep reinforcement learning. We find that this method significantly outperforms regular end-to-end learning and is able to discover robust control policies in challenging tasks. Video demonstrations of learned behaviors and the code can be found at