TANGO: Humanoid Navigation in Cluttered Environments with a Whole-Body Vision-Language-Action Model

TL;DR

TANGO is a whole-body vision-language navigation framework for humanoid robots in cluttered environments. It predicts 29-DoF joint-space actions for collision-free movement. TANGO demonstrates state-of-the-art performance in simulation and real-world deployment.

Why It Matters

TANGO enables humanoid robots to navigate complex environments with ease, which has significant implications for search and rescue, healthcare, and service robotics. This technology can improve robot autonomy and safety in various applications. Effective navigation in cluttered environments is a crucial step towards achieving human-robot collaboration.

Key Findings

• whole-body vision-language navigation
• 29-DoF joint-space action prediction
• state-of-the-art performance in simulation and real-world deployment