HumanoidBench: Simulated Humanoid Benchmark for Whole-Body Locomotion and Manipulation

TL;DR

HumanoidBench is a high-dimensional simulated humanoid robot benchmark designed to accelerate research in whole-body locomotion and manipulation, revealing current algorithms' limitations and supporting hierarchical learning approaches.

Contribution

Introduction of HumanoidBench, a comprehensive simulation platform for humanoid robot tasks, enabling systematic evaluation and fostering advancements in algorithms.

Findings

State-of-the-art reinforcement learning algorithms struggle with most tasks.

Hierarchical learning with robust low-level policies outperforms flat approaches.

HumanoidBench facilitates rapid testing and development in humanoid robotics.

Abstract

Humanoid robots hold great promise in assisting humans in diverse environments and tasks, due to their flexibility and adaptability leveraging human-like morphology. However, research in humanoid robots is often bottlenecked by the costly and fragile hardware setups. To accelerate algorithmic research in humanoid robots, we present a high-dimensional, simulated robot learning benchmark, HumanoidBench, featuring a humanoid robot equipped with dexterous hands and a variety of challenging whole-body manipulation and locomotion tasks. Our findings reveal that state-of-the-art reinforcement learning algorithms struggle with most tasks, whereas a hierarchical learning approach achieves superior performance when supported by robust low-level policies, such as walking or reaching. With HumanoidBench, we provide the robotics community with a platform to identify the challenges arising when solving diverse tasks with humanoid robots, facilitating prompt verification of algorithms and ideas. The open-source code is available at https://humanoid-bench.github.io.