| ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange |
Hi, I'm working on a project in which I have to use SLAM and navigation for the robot. I want to know if it is possible to use SLAM for my virtual robot by moving in a Gazebo world , or is it mandatory to have the real robot (hardware) running around? Thank you
You can definitely use Gazebo or any other simulator (e.g. Stage). The SLAM algorithm will be embedded into a node that requires streams of sensor data.
From the algorithm point of view there are no differences if these are coming from real sensors or "fake" ones from a simulator.
Here a tutorial with Gazebo and Gmapping SLAM algorithm http://wiki.ros.org/turtlebot_gazebo/...
While this comment is in general true (ie: due to the decoupling afforded by ROS a SLAM node should not care about whether it receives input from "real" or "virtual" sensors), there is one thing I've observed when using SLAM in simulated environments: real world environments are messy and highly unstructured. SLAM algorithms typically like that (lots of features, high saliency). Simulated environments (ie: models) might be too clean for a particular algorithm to work.
This will not affect all algorithms in the same way, but it may be something to keep in mind.
Yes, agree 100%. If the user wants to test laser-based SLAM algorithm, an alternative to "clean models" consists in using as ground truth a map obtained from a different SLAM run, such as these ones http://www2.informatik.uni-freiburg.d... or from the RADISH database https://dspace.mit.edu/handle/1721.1/...
Hi, thank you all that is definitely helpful. And about gmapping on melodic, Since I'm a beginner I don't really know what it means to "build gmapping from source" as indicated in the following link https://answers.ros.org/question/2961... any help?? thanks
Asked: 2019-05-09 07:22:22 -0600
Seen: 972 times
Last updated: May 09 '19
ROS Answers is licensed under Creative Commons Attribution 3.0 Content on this site is licensed under a Creative Commons Attribution Share Alike 3.0 license.
