// the find
SYSU-RoboticsLab/FAEL
FAEL: Fast Autonomous Exploration for Large-Scale Environments with a Mobile Robot
FAEL is a ROS1 (Noetic) package implementing the autonomous exploration algorithm from the authors' 2023 IEEE RA-L paper: a mobile ground robot plans paths to explore an unknown large-scale environment as fast as possible. It's aimed at robotics researchers who want to reproduce a published exploration benchmark or build on it, not at anyone wanting a production exploration stack.
Backed by a peer-reviewed paper with a real algorithmic contribution (a heuristic for the NP-hard path optimization), not just a demo. Ships a complete, runnable Gazebo simulation setup with four distinct world files so you can actually reproduce the claimed results rather than taking them on faith. Reuses established building blocks (cmu_planner for local planning/terrain analysis, ufomap for 3D mapping, RotorS/jackal for simulation) instead of reinventing basic robotics infrastructure. The README's known-issues section honestly documents a real failure mode (LiDAR point cloud degrading on integrated GPUs) instead of burying it.
Hard-pinned to ROS1 Noetic and Ubuntu 20.04, both now past or near end-of-life, with no ROS2 port — anyone adopting this today is building on a dead stack. The repo vendors entire third-party ROS packages (jackal, rotors_fly, mav_comm, nmea_msgs) as committed source under 3rdparty/ instead of using rosdep or a .rosinstall/vcstool manifest, which bloats the tree and makes license provenance and upstream updates a headache. The core contribution — the heuristic solver for the path optimization — isn't documented or commented in the code itself; you're expected to go read the paper to understand what the planner is actually doing. No CI, no tests, and no activity since mid-2024, so this reads as a one-off research artifact release rather than something intended to take bug reports or contributions.