Quadruped Navigation
Quadruped navigation research focuses on enabling robots to autonomously traverse diverse terrains, achieving robust and safe locomotion. Current efforts concentrate on integrating multiple sensor modalities (vision, proprioception) using advanced algorithms like vision-language models and reinforcement learning to create adaptable and reliable navigation systems. This work leverages techniques such as cost map generation, motion anisotropy awareness for trajectory planning, and force-responsive control for human-robot interaction, improving both efficiency and safety. The resulting advancements have significant implications for applications ranging from search and rescue to assistive robotics.
Papers
July 23, 2024
March 20, 2024
March 15, 2024
September 8, 2023
November 26, 2022