Complete SLAM Implementation
I have searched many places for implementing SLAM but did not find a complete solution . I am using a quadrotor with Asus Xtion Pro RGBD and I also have a Hokuyo LIDAR. Could someone please give the concise complete blueprint or roadmap of implementing the SLAM process in ROS taking one particular package as example. I know the theory behind it all and so can fiddle with the code, and make packages on my own. But being fairly new to the ROS framework, I'm facing a hard time getting the ideas together for implementation. I hope that someone who has passed through the mill will be able to guide me.
Asked by Raman on 2015-05-28 08:49:15 UTC
Answers
There are many pre-existing SLAM packages available in ROS. Each has different performances. Most packages have easy integration tutorials on ROS wiki.
The most common 2D SLAM packages are - gmapping, hector_slam, MRPT slam
For 3D SLAM, I haven't found a lot of support but here are a couple that might be useful. Some of them are not supported for ros-indigo so you might have to modify a little bit - RGBD slam, ICP
Asked by ajain on 2015-05-28 10:34:39 UTC
Comments
Thanks for your answer. Hector_Slam has not implemented SLAM really. They just did some standalone mapping. Also, RGBD Slam is giving lots of errors in Indigo, even after some modifying. If you want, I can share the results with you. These incomplete packages are really troubling a beginner like me.
Asked by Raman on 2015-05-28 18:25:47 UTC
@Raman I suggest that you take a little more time to research the packages mentioned by @ajain , you appear to be jumping to incorrect conclusions. Another SLAM algorithm in ROS is slam_karto
Asked by tfoote on 2015-05-28 22:04:38 UTC
Comments