Autonomous Elevator Operation with Unitree Go2 Robot Dog

This academic research project, currently under review for publication in IEEE, focused on enabling a Unitree Go2 quadruped robot to autonomously operate elevators. The system was designed to expand the mobility of legged robots in real-world indoor environments, where vertical navigation is essential.

The robot was equipped with ROS2, the Unitree SDK, and computer vision techniques to achieve perception and control. Using AprilTag detection, the robot could accurately identify elevator panels, perform precise docking maneuvers, and press the correct buttons autonomously.

Highlights:

• Autonomous Elevator Operation: Successfully integrated perception, planning, and actuation to allow the robot to detect panels, dock, and press buttons without human intervention.
  
• Computer Vision Integration: Used AprilTag technology to ensure robust detection of elevator panels in varied lighting and angle conditions.
  
• Seamless Control: Combined ROS2 middleware, the Unitree SDK, and custom navigation algorithms to achieve reliable performance.
  
• Research Contribution: Project submitted to IEEE, contributing to advancements in mobile robotics and service robot autonomy.
  

Robot in Action:

This project demonstrates the potential of quadruped robots in human-centered environments, highlighting how advanced perception and autonomy can extend their usability to multi-floor navigation.