Room for me?: Mobile Navigation for Entering a Confined Space Using Deep Reinforcement Learning

Published in Proc. of Int. Conf. on Ubiquitous Robots (UR), IEEE, 2023, 2023

Inspired by challenges in elevator navigation, this work uses deep reinforcement learning to enable robots to safely enter confined spaces.

Highlights

  • Designed reward functions for safe entry without blocking doors
  • Trained in simulation and validated in real-world setups
  • Demonstrated generalization to unseen elevator layouts

Media

🎥 Demo Video
📄 IEEE Xplore