search

izhengfan / se2clam

SE(2)-Constrained Localization and Mapping by Fusing Odometry and Vision (IEEE Transactions on Cybernetics 2019)
https://github.com/izhengfan/se2clam
MIT License
161 stars 55 forks source link

some basic questions about this project #3

Closed Tofupi closed 5 years ago

Tofupi commented 5 years ago

I have read part of source code, It seems no exists initialization process, which is different from ORB_SLAM2. and R t are directly get from odom-2d in triangulate. Local BA part is not read, and it seems use slide window size from what I see in rviz.right? Global BA is alsome and I am curious about it, can you explain it process theory/chart. sorry to bouther you.

Tofupi commented 5 years ago

I am thinking it may be a light vision-odom slam for 2D slam and can be apply to vacuum cleaner, do you agree? or do you have other advice about applying this project to vacuum cleaner?

izhengfan commented 5 years ago

There no actual init process, odom measurement is directly use as the first guess.

Yes, Local BA is like slide window. And there is also nothing new about global BA, it is just pose graph optimization.

I am not familiar with vacuum cleaner, but I think the method is applicable to any ground vehicles moving on a plane.

Tofupi commented 5 years ago

Thanks for your reply. and in your demo exe, I found no Odom and Image Message Synchronization Mechanism. and it assumes that the two messages have the same timestamp. Anyway, Messages can not be absoutely Synchronized with no diff. have you test SLAM accuracy when messages timestamp exists diff and what's the diff range? sorry to bother you again.

izhengfan commented 5 years ago

We have sensors synchronization in the hardware level. Both encoders and industrial camera can be triggered by a signal sent from a MCU.

izhengfan commented 5 years ago

Software synchronization should be ok, but hardware sync is definitely better.

Tofupi commented 5 years ago

ok, thank you very much! I 'll try.