ROS2 Actions - Fibonacci with Feedback

🧠 Concept: Actions (Long-Running Tasks with Feedback)

Actions are for tasks that take time and need progress updates — navigating to a waypoint, running a calibration, executing a trajectory.

Unlike services (instant response), actions have 3 phases:

1. Goal — client sends what it wants done
2. Feedback — server sends progress updates while working
3. Result — server sends the final outcome

Client sends goal ──▶ Server accepts ──▶ Server sends feedback (N times) ──▶ Server sends result

Key API (server side):

from rclpy.action import ActionServer

self.action_server = ActionServer(self, ActionType, '/action_name', self.execute_callback)

def execute_callback(self, goal_handle):
    # Send feedback during execution
    goal_handle.publish_feedback(feedback_msg)
    # Mark as succeeded and return result
    goal_handle.succeed()
    return result

When to use: Anything that takes more than ~100ms and benefits from progress reporting.

Problem Statement

Actions are for long-running tasks that provide feedback and can be cancelled. Implement an Action Server that computes the Fibonacci sequence.

Requirements

• Create an action server for the Fibonacci action (custom or example)
• Action Name: fibonacci
• Logic: Compute sequence up to order (from goal)
• Publish feedback (partial sequence) at each step
• Return result (full sequence) when done
• Simluated delay: 1 second per step

Input/Output Format

Input:

• Action Goal: order=5

Output:

• Feedback: [0, 1], [0, 1, 1], ...
• Result: [0, 1, 1, 2, 3, 5]

⚠️ Common Pitfalls

• Blocking the callback. You must mark the callback as async or use threads.
• Forgetting to set the goal state to succeed.

📚 Helpful Resources

• ROS2 Actions Tutorial