====== Lab 07 ======
This lab page has not yet been ported to ROS 2. It will be updated shortly.
During this lab, you should learn how to work with action servers and clients.
===== ROS action server and client =====
The ROS action servers/clients provide the ability to execute and interface preemptable tasks that can take a long time to execute.
It also allows you to get a regular feedback from tasks that are currently being executed.
Examples of such a task can be moving a robot to a target location or following a given path with a robot.
You can check out the official python tutorial for writing [[https://docs.ros.org/en/kilted/Tutorials/Intermediate/Writing-an-Action-Server-Client/Py.html | action server and action client]].
==== Action specification ====
The action specification have three parts, the goal (similar to services' request), the results (similar to services' response), and the feedback messages to enable reporting on the action run.
The action server and action clients can be created using ''rclpy.action'' library in python.
The definition of an example action ''GoToPosition.action'' can look as follows:
#goal definition
geometry_msgs/Point position
---
#result definition
bool success
string message
---
#feedback
float32 position_error
float32 orientation_error
string message
==== Action server ====
The action server part listens for a goal message and once the goal is received, it starts executing procedures to fulfill the given goal.
While working on finishing the goal, the action server can periodically report the current state of goal processing through a feedback message, but also has to check if it was not stopped through preemption request.
Upon finishing or failing to fulfill the goal, the action server sends its result to the client.
The action server code could look like this:
import time
import rclpy
from rclpy.node import Node
from rclpy.action import ActionServer, GoalResponse
from rclpy.executors import ExternalShutdownException
from your_package.action import GoToPosition
class GoToPositionActionServer(Node):
def __init__(self):
super().__init__('goto_position_action_server')
self._action_server = ActionServer(self, GoToPosition, 'goto_position',
execute_callback=self.execute_callback, goal_callback=self.goal_callback)
def execute_callback(self, goal_handle):
self.get_logger().info('Executing goal...')
success = True
rate_hz = 10.0
period = 1.0 / rate_hz
feedback = GoToPosition.Feedback()
while rclpy.ok():
# goal canceled
if goal_handle.is_cancel_requested:
self.get_logger().info('Goal canceled')
goal_handle.canceled()
result = GoToPosition.Result()
result.success = False
result.message = "goal canceled"
return result
.... # do something now, compute position and orientation error
feedback.message = "still working on it"
feedback.position_error = ...
feedback.orientation_error = ...
goal_handle.publish_feedback(feedback)
# check finish/fail conditions
if #finished or #failed:
success = True/False # based on finished or failed
break # break the action loop
time.sleep(period)
# set final state before returning
result = GoToPosition.Result()
result.success = success
if success:
result.message = "finished successfully"
goal_handle.succeed()
else:
result.message = "failed miserably"
goal_handle.abort()
return result
def goal_callback(self, goal_request):
self.get_logger().info('Received goal request')
position_invalid = ...
if position_invalid:
self.get_logger().warn('Rejecting goal: position invalid')
return GoalResponse.REJECT
# Otherwise accept
self.get_logger().info('Goal accepted')
return GoalResponse.ACCEPT
def main(args=None):
try:
with rclpy.init(args=args):
node = GoToPositionActionServer()
rclpy.spin(node)
except (KeyboardInterrupt, ExternalShutdownException):
pass
if __name__ == '__main__':
main()
==== Action client ====
The action client can be used to interface the action server efficiently.
A simple client can look like this:
import rclpy
from rclpy.node import Node
from rclpy.action import ActionClient
from rclpy.executors import ExternalShutdownException
from your_package.action import GoToPosition
class GoToPositionActionClient(Node):
def __init__(self):
super().__init__('goto_position_action_client')
self._action_client = ActionClient(self, GoToPosition, 'goto_position')
def send_goal(self):
# Wait for server
self.get_logger().info('Waiting for action server...')
self._action_client.wait_for_server()
# Create goal
goal_msg = GoToPosition.Goal()
goal_msg.position = ...
self.get_logger().info('Sending goal...')
# Send goal asynchronously
self._send_goal_future = self._action_client.send_goal_async(goal_msg, feedback_callback=self.feedback_callback)
self._send_goal_future.add_done_callback(self.goal_response_callback)
# Called when server accepts/rejects goal
def goal_response_callback(self, future):
goal_handle = future.result()
if not goal_handle.accepted:
self.get_logger().info('Goal rejected')
return
self.get_logger().info('Goal accepted')
# Request result
self._get_result_future = goal_handle.get_result_async()
self._get_result_future.add_done_callback(self.get_result_callback)
# Feedback callback
def feedback_callback(self, feedback_msg):
feedback = feedback_msg.feedback
self.get_logger().info(f'Feedback: {feedback.message}, 'f'pos_err={feedback.position_error},
'f'ori_err={feedback.orientation_error}')
# Result callback
def get_result_callback(self, future):
result = future.result().result
status = future.result().status
self.get_logger().info(f'Result: success={result.success}')
self.get_logger().info(f'Message: {result.message}')
self.get_logger().info(f'Status: {status}')
# Shutdown after receiving result
rclpy.shutdown()
def main(args=None):
try:
rclpy.init(args=args)
node = GoToPositionActionClient()
node.send_goal()
rclpy.spin(node)
except (KeyboardInterrupt, ExternalShutdownException):
pass
if __name__ == '__main__':
main()
===== Homework 5 assignment =====
Read and try to understand the assignment of the homework [[homework05|HW5]].