Multi-Robot system: development and tuning issues. [closed]

asked 2012-01-26 19:17:00 -0500

Nikkio gravatar image

Hi community!! I'm trying to develop a multi-robot system where the single bots are able to cooperate. At the moment all the environment, a set of empty rooms and two pioneer3AT, is simulated on gazebo that is run on a pc with enough HW resources (sorry i don't remember the features of it, i will write these on Monday if needed). For each robot i launch a .launch file which it spawn the robot model on gazebo, start a move_base node, gmapping (each robot builds map on real-time), and another node that makes some really light operations. When I start all the launch files (gazebo and the 2 bots) on the pc mentioned above everything works ok, awesome. When i try to launch gazebo on the "big" pc and just one robot's launch file from another pc (i7 core, 4gb RAM, Gbit ethernet) the simulation becomes really slow..without strange errors (Ok sometimes move_base lost the map update)!! Any suggestions about how to proceed for get remarkable improvement??

edit retag flag offensive reopen merge delete

Closed for the following reason question is not relevant or outdated by tfoote
close date 2013-09-09 15:07:58


I'm a bit confused about your use-case and setup. Are you running multiple instances of gazebo on different machines (each with their own ros master)? Or are you running multiple robots within a single gazebo instance?
JeffRousseau gravatar image JeffRousseau  ( 2012-01-30 07:14:24 -0500 )edit
I'm running a single istance of Gazebo (and the ros master too) on a workstation. From other pc on the same net I launch several nodes, which are going to connect with that ros master on the workstation. If you want i could publish my roslaunch files and yaml configurations
Nikkio gravatar image Nikkio  ( 2012-01-31 20:14:01 -0500 )edit