Project Page: Hybrid Multi-Goal Path Planning for Powered Wheelchairs

Source of the static project page, served with GitHub Pages.

Live site: https://egeozgul.github.io/AutonomousWheelchairProject/

What the page shows

• Paper title, authors, abstract, and setup figures
• A synchronized player for each hardware experiment: overhead video, the wheelchair's two forward cameras, and an animated path plot
• A 3D LiDAR point cloud player for experiments that have a LiDAR recording

Repository layout

├── index.html              The page
├── pointcloud_player.js    3D LiDAR player, reads the bags in the browser
├── config.json             Experiments listed on the page: names, files, offsets, environments
├── <experiment>/           One folder per experiment
└── *.jpg, *.png            Figures and logos

An experiment folder can contain:

File	Used for
overhead_web.mp4	Overhead view
cam1_detected.mp4, cam2_detected.mp4	Forward camera views
*.csv	Path plot: t, chair_x, chair_y, chair_yaw, plus <name>_x, <name>_y per pedestrian
lidar_bag/lidar_bag_0_filtered.db3	Point cloud player (ROS 2 SQLite bag with /livox/lidar and /vehicle/odometry)

Adding an experiment

1. Create a folder and upload its files.
2. Add an entry to projects in config.json with its id, name, folder, and files.
3. Commit. GitHub Pages republishes the site within a minute or two.

The point cloud player lists an experiment automatically once its lidar_bag/lidar_bag_0_filtered.db3 exists. To use a different bag name, set "lidarBag" in that experiment's files.

File size: keep files under 25 MB when uploading through the GitHub website, and don't store bags with Git LFS, since GitHub Pages serves only a pointer for LFS files.