0
$\begingroup$

Rosanswers logo

Is rclcpp::Clock(RCL_ROS_TIME).now() always the same as node_->get_clock()->now()?


Originally posted by galou on ROS Answers with karma: 265 on 2021-12-02

Post score: 1

$\endgroup$

2 Answers 2

0
$\begingroup$

Rosanswers logo

Well, it looks that the answer is NO. I had some issues with this and wasn't satisfied with the answer so I wrote a small node to test these clocks. The test is carried out with Galactic. The complete code can be found at https://github.com/galou/test_clock and the source is below

#include <chrono>
#include <string>

#include <rclcpp/rclcpp.hpp>

using namespace std::chrono_literals;

class TestClockNode : public rclcpp::Node
{
  public:

    TestClockNode(
        const std::string & node_name,
        const rclcpp::NodeOptions & options = rclcpp::NodeOptions()) :
      rclcpp::Node{node_name, options}
    {
      timer_ = create_wall_timer(1s, [this] () {timerCallback();});
    }

  private:

    rclcpp::TimerBase::SharedPtr timer_;
    void timerCallback()
    {
      RCLCPP_INFO_STREAM(get_logger(), "now(): " << now().seconds());
      RCLCPP_INFO_STREAM(get_logger(), "  rclcpp::Clock{}.now(): " << rclcpp::Clock{}.now().seconds());
      RCLCPP_INFO_STREAM(get_logger(), "  rclcpp::Clock{RCL_ROS_TIME}.now(): " << rclcpp::Clock{RCL_ROS_TIME}.now().seconds());
      RCLCPP_INFO_STREAM(get_logger(), "  rclcpp::Clock{RCL_SYSTEM_TIME}.now(): " << rclcpp::Clock{RCL_SYSTEM_TIME}.now().seconds());
      RCLCPP_INFO_STREAM(get_logger(), "  rclcpp::Clock{RCL_STEADY_TIME}.now(): " << rclcpp::Clock{RCL_STEADY_TIME}.now().seconds());
    }

};

int
main(int argc, char * argv[])
{
  rclcpp::init(argc, argv);

  auto node = std::make_shared<TestClockNode>("test_clock");
  if (node == nullptr)
  {
    return 1;
  }

  auto executor = rclcpp::executors::SingleThreadedExecutor{};
  executor.add_node(node);
  executor.spin();

  rclcpp::shutdown();

  return 0;
}

You need a running simulated time provider (i.e. a simulator) to test and you need to launch with

ros2 run test_clock test_clock --ros-args --param use_sim_time:=true --

Example of output:

[INFO] [1645622804.092181382] [test_clock]: now(): 2216.74
[INFO] [1645622804.092376229] [test_clock]:   rclcpp::Clock{}.now(): 1.64562e+09
[INFO] [1645622804.092445844] [test_clock]:   rclcpp::Clock{RCL_ROS_TIME}.now(): 1.64562e+09
[INFO] [1645622804.092512516] [test_clock]:   rclcpp::Clock{RCL_SYSTEM_TIME}.now(): 1.64562e+09
[INFO] [1645622804.092567862] [test_clock]:   rclcpp::Clock{RCL_STEADY_TIME}.now(): 199532

Originally posted by galou with karma: 265 on 2022-02-23

This answer was ACCEPTED on the original site

Post score: 1


Original comments

Comment by aprotyas on 2022-02-27:
I see the results, but looking at the source I'm really not sure why there would be a difference between the two. rclcpp::Node defaults to a clock of type RCL_ROS_TIME, and the node_->now() function simply delegates to the clock's now() function instead.

$\endgroup$
0
$\begingroup$

Rosanswers logo

When use_sim_time is enabled, RCL_ROS_TIME type clock will use time source from topic /clock.

But we have to attach the rcrclcpp::Clock(RCL_ROS_TIME) to one ros node(and furthor one executor) to be able subscibe to topic /clock and update simulated time stored in RCL_ROS_TIME clock instance. According to this understanding, we can also conclude that, as long as the attached node and executor is not spin()ing, the simulated clock won't be updated.

That's the behavior of node_->get_clock().

A standalone rcrclcpp::Clock(RCL_ROS_TIME) have no way to subscribe to topic /clock, then as a fallback strategy, it will behave like a RCL_SYSTEM_TIME type clock.


Originally posted by tsingakbar with karma: 26 on 2022-09-23

This answer was NOT ACCEPTED on the original site

Post score: 1

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.