Awesome effort! I am not a SLAM guys at all, but is there such a thing as "Multi-agent SLAM" where multiple sources are building the map simultaneously (or asynchronously)?
Comments
Log in with your Bluesky account to leave a comment
There is multi-session SLAM, the most recent paper that comes to mind is https://arxiv.org/pdf/2404.15263
Hopefully the related work section can point you to other works :)
Thanks for the reference! We addressed a multi-agent version of dense neural SLAM, but I guess Vincent's question was more general. There is a large body of classical SLAM methods addressing this problem: There are several surveys or see this overview by Andrew Davison: https://youtu.be/OH4NnALUxYQ?si=BlMH2ENUU79Q7Zuw&t=2391
Comments
Hopefully the related work section can point you to other works :)
https://vladimiryugay.github.io/magic_slam/index.html
And there's also a repo here
https://github.com/MIT-SPARK/Kimera-Multi