This repository provides an example for using the SCRIMMAGE simulator with the ROS 2D Navigation Stack. The interface between SCRIMMAGE and ROS is almost identical to the interface between stage_ros and ROS.

You will need to build the SCRIMMAGE simulator

1. Setup a ROS catkin workspace, if you don't have one already:

    $ mkdir -p ~/catkin_ws/src && cd ~/catkin_ws/
    $ catkin_make
   

2. Clone this repo into into ~/catkin_ws/src:

    $ cd ~/catkin_ws/src
    $ git clone https://github.com/SyllogismRXS/scrimmage_ros.git
   

3. Build it:

    $ cd ~/catkin_ws
    $ catkin_make
   

The example SCRIMMAGE mission / roslaunch simulates two ground robots running the ROS 2D Navigation stack in a simple maze world. It's probably a good idea to open a new terminal at this point, so that your terminal has access to any new SCRIMMAGE environment variables.

$ source ~/catkin_ws/devel/setup.bash               # source the catkin_ws
$ roslaunch scrimmage_ros multiple-robots.launch    # launch scrimmage/ros

The SCRIMMAGE simulator controls the time, generates sensor data, and integrates the motion models. To kill the simulation, type "CTRL + c" in the window that you ran the roslaunch command from.

The robots will be initialized using the positions in the scrimmage mission file (e.g., ros-ex1.xml), but you can specify an initial pose estimate for each robot in the multiple-robots.launch file. You will want to modify the initial pose estimates to match the actual poses in the scrimmage mission file. This may be automated in the future, but it's manual configuration for now.

You need to specify the same map file in both multiple-robots.launch (name="map_name" default="simple-maze-1") and the ros-ex1.xml scrimmage mission file (e.g., entity_interaction order="0" map="simple-maze-1").

You don't need to use an URDF file for a simple ground robot with a few sensors. Specify the sensor transform from the robot's center using the xyz and rpy XML attribute tags of a sensor tag.