Back to Interfacing with ROS.

Your test program might look like

#include <eeros/control/ros/EerosRosTools.hpp>
#include <ros/ros.h>
#include <signal.h>
 
int main(int argc, char **argv) {
  ...
  eeros::control::rosTools::initNode("eerosNode"); // initialize the ROS node in your main function and give it a name.
  log.info() << "ROS node initialized.";
  signal(SIGINT, ...);
  ...

All ROS tools such as rqt will list your EEROS application under the name eerosNode.

If you want to register a signal handler, e.g. for shutting down a system (see Shutting down a System Properly), you have to register it after the ROS node is initialized. The ROS node will properly shut down as soon as the last node handler is destroyed when going out of scope.

Your test program might look like

#include <eeros/control/ros2/EerosRosTools.hpp>
#include <rclcpp/rclcpp.hpp>
#include <signal.h>
 
int main(int argc, char **argv) {
  ...
  RosTools::initRos(argc, argv);
  auto node = RosTools::initNode("eerosNode");
  if (node != nullptr) log.info() << "ROS node initialized: " << node->get_name();
  ...
  ControlSystem cs(node, dt);
  ...

When defining a control system which uses ROS subscriber or publisher blocks, you have to pass this node to them.

All ROS tools such as rqt will list your EEROS application under the name eerosNode. The variable node has to be passed to all blocks in the control system which publish or subscribe to ROS messages.

If you want to register a signal handler, e.g. for shutting down a system (see Shutting down a System Properly), you have to register it after the ROS node is initialized. The ROS node will properly shut down as soon as the last node handler is destroyed when going out of scope.