Vicky - Navigation Software

Independent Research

"Implementing and Enhancing Shared Control Navigation in ROS-based Smart Wheelchairs"

by. ZHANG Yunxin (Vicky)

Through our user research, we have observed that electric wheelchair users face significant challenges in manually navigating through narrow paths, doorways, and other constrained environments. To address these issues, we propose designing a wheelbase with independent steering wheels to allow for a smaller turning radius and enhanced maneuverability. My research will focus on implementing and improving a shared control model for our ROS-based smart wheelchairs, balancing user control and autonomous navigation for a safer and more efficient experience.

Lidar Demo

Touch Screen Demo

Project Journey

1 / Learning ROS2 and Nav2 Fundamentals

Started by understanding the core concepts of ROS2 and Nav2, the navigation stack for autonomous robots
Explored key components such as cost maps, types of navigation servers, and behavior trees

2 / Simulation and Visualization with TurtleBot3

Explored Nav2 stack in Gazebo simulation and visualized robot behavior in RViz

3 / Developing a GUI Menu for Waypoint Navigation

Designed a QT-based GUI to allow users to select predefined waypoints
Developed a ROS node to send selected waypoint coordinates as navigation goals to the robot

4 / LiDAR Setup

Set up Livox Mid-360 LiDAR to capture point cloud data
Used FAST-LIO for accurate and efficient point cloud generation

5 / Converting Point Clouds into a 2D Occupancy Grid Map

Used Octomap Server for realtime conversion from 3D point cloud data to 2D occupancy grid map
Ensured compatibility with Nav2 for path planning and obstacle avoidance