// the take
A hand-curated list of SLAM (Simultaneous Localization and Mapping) research papers, organized by sub-topic (VIO, lidar odometry, calibration, bundle adjustment, etc.) with links to arXiv and IEEE. It's for robotics/computer vision researchers who want a reading list of recent SLAM papers rather than any runnable code — there is none here.
The topic tagging per entry ([VIO], [Calibration], [Fusion], etc.) makes it easy to scan for a specific sub-area instead of reading titles cold. It links out to primary sources (arXiv PDFs, IEEE Xplore) rather than summarizing secondhand, so you're reading the actual papers. Coverage spans the full 2021 SLAM landscape from classical filter-based VIO to early deep-learning approaches like DROID-SLAM and NeuralRecon.
Dead since October 2021 — the README itself says updates are paused indefinitely, so anything from the last four-plus years of SLAM research (a field that's moved fast, especially on learned/NeRF-based approaches) is missing. It's a link list, not an evaluation: no commentary on which papers panned out, which methods got superseded, or which code releases are worth using. Several entries duplicate the same paper under different numbers (e.g. #73/#87/#91), and some links are already rotting (Google Sites-hosted PDFs with signed URLs that expire). There's nothing to adopt or run — the value is entirely in link-following, and for that purpose a live-maintained source (arXiv listings, Papers With Code SLAM tag) will be more current.
We read the README, the directory tree and a key source file in October 2026. Nobody pays to be listed. How we judge →