Is this a SLAM or ACML kind of problem?
So let's say I need to construct an autonomous robot that has to navigate a premade map where obstacles will be randomly placed at the start of every round, but the wall placement and everything else in the map stays the same. Would I use SLAM to generate a new map (for the new obstacles) or should I just implement an ACML type of solution since I already know what the map looks like and since it won't change significantly; the robot just has to avoid the objects (or hit them if necessary)?
Asked by NicolasMontoya on 2018-11-03 00:43:13 UTC
Answers
This sounds to me like an AMCL problem. You could make the map "by hand" or using SLAM. But once you have it, AMCL will be able to figure out where the robot is on the map, and the navigation stack will use that to compute paths, while in real time it uses /scanner data to avoid obstacles not on the map
Asked by pitosalas on 2020-07-06 09:19:31 UTC
Comments
You will also need to find a way to clear spaces marked as obstacles while SLAM
Asked by parzival on 2020-07-06 09:36:58 UTC
Comments