hdl_graph_slam

Requires manual installation of multiple libraries like ndt_omp, fast_gicp, and g2o with suitesparse, which can be time-consuming and prone to compilation issues, as noted in the Requirements section.

Performance heavily depends on tuning parameters like ndt_resolution and registration_method, with the README admitting that mapping quality 'largely depends on the parameter setting,' making it less suitable for out-of-the-box use.

Primarily optimized for 3D LIDAR odometry; lacks built-in support for visual or other sensor modalities without significant modifications, which may limit its use in multi-modal SLAM applications.

The author has released a new SLAM package (glim), suggesting that hdl_graph_slam might not receive active updates, and users may need to migrate for newer features or improvements.