ROS2 Lifecycle Node - State Machine

🧠 Concept: Lifecycle Nodes

Standard nodes start immediately and run forever. Lifecycle nodes give you explicit control over a node's startup, operation, and shutdown through a state machine:

Unconfigured ──configure()──▶ Inactive ──activate()──▶ Active
                                   ◀──deactivate()──         
                                   ──cleanup()──▶ Unconfigured

Why this matters: In a robot, you need to ensure sensors are ready before the planner starts, and the planner is stopped before sensors shut down. Lifecycle nodes enforce this ordering.

Key callbacks to implement:

def on_configure(self, state):    # called on configure() — set up publishers, timers
    return TransitionCallbackReturn.SUCCESS

def on_activate(self, state):     # called on activate() — start publishing
    return super().on_activate(state)

def on_deactivate(self, state):   # called on deactivate() — stop publishing
    return super().on_deactivate(state)

Problem Statement

Implement a Managed (Lifecycle) Node. These nodes have a state machine (Unconfigured, Inactive, Active, Finalized) and transitions (Configure, Activate, Deactivate, Cleanup).

Requirements

• Create a class MyLifecycleNode inheriting from LifecycleNode
• Implement on_configure (return SUCCESS)
• Implement on_activate (return SUCCESS, start publisher)
• Implement on_deactivate (return SUCCESS, stop publisher)
• Implement on_cleanup (return SUCCESS)

Input/Output Format

Output:

• Logs indicating state transitions.

⚠️ Common Pitfalls

• Using a regular Node instead of LifecycleNode.
• Forgetting to call super().__init__(...).

📚 Helpful Resources

• Lifecycle Nodes