HumanoidBench: Simulated Humanoid Benchmark for Whole-Body Locomotion and Manipulation
This provides the robotics community with a platform to accelerate algorithmic research for humanoid robots, though it is incremental as it builds on existing simulation and benchmarking approaches.
The authors tackled the problem of costly and fragile hardware setups bottlenecking humanoid robot research by creating HumanoidBench, a simulated benchmark for whole-body locomotion and manipulation tasks. They found that state-of-the-art reinforcement learning algorithms struggle with most tasks, while a hierarchical learning approach with robust low-level policies achieves superior performance.
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.