Training and Simulation of Quadrupedal Robot in Adaptive Stair Climbing for Indoor Firefighting: An End-to-End Reinforcement Learning Approach

Baixiao Huang Baiyu Huang, Yu Hou · 2026

Read Training and Simulation of Quadrupedal Robot in Adaptive Stair Climbing for Indoor Firefighting: An End-to-End Reinforcement Learning Approach by Baixiao Huang Baiyu Huang, Yu Hou (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-03087.txt.

Description

Four-legged robots learn to climb stairs for fire searches in a 2026 paper by Baixiao Huang and 2 co-authors. Quadrupeds perform primary searches early in indoor fires, looking for victims under hazardous conditions. Situational awareness and rapid stair climbing across varied staircases remain the main challenges for such robot-assisted searches in complex buildings with smoke and obstacles.

Their two-stage reinforcement learning approach trains navigation and locomotion together, end to end. A Unitree Go2 robot first learns on pyramid-stair terrain in simulation. The policy then transfers to realistic straight, L-shaped, and spiral staircases. A centerline-based navigation formulation unifies movement and direction without hierarchical planning. Using only local height-map perception, the policy generalized across staircases. The paper analyzes success, efficiency, and failure modes.

Catalog record

CREATORBaixiao Huang Baiyu Huang, Yu Hou
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE28 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602-03087

Browse the full library · About the Tower of Babel