In autonomous robotics, telling a robot to “go to point X” is a standard task. But what if you need it to “go to point A, then B, then C, and patrol all day?” This is the challenge of waypoint navigation, and it’s a critical one for tasks like warehouse patrols, agricultural surveys, or search and rescue.

A fantastic, focused ROS package for solving this is follow_waypoints, forked by Chris Sunny from an original project by Daniel Snider. It’s a simple, powerful tool that acts as a mission manager for the ROS navigation stack.

The Problem: move_base Only Takes One Goal at a Time

The core navigation “brain” in ROS is move_base. It’s brilliant at planning a path from A to B while avoiding obstacles. However, it’s designed to handle one goal at a time. If you want to give it a sequence of goals, you’d have to write a custom script to wait for the first goal to complete, then send the second, and so on.

The Solution: A Smart Waypoint Queue

This is exactly what the follow_waypoints package does. It’s a Python-based ROS node that acts as a smart buffer on top of move_base.

It allows a user to “load up” a list of waypoints and then, with a single command, tell the robot to “go!” The node then takes over, feeding each waypoint to move_base one by one, only sending the next goal after the previous one has been successfully reached.

How It Works in Practice

The workflow is simple and designed for easy use with standard ROS tools:

  1. Create Waypoints: The easiest way to create a path is right in RViz (the ROS visualization tool). By using the “2D Pose Estimate” tool, a user can simply click on the map to define a series of waypoints. The follow_waypoints node listens for these clicks and adds them to its queue.
  2. Visualize the Path: As waypoints are added, the node publishes the entire list as a PoseArray, which can be viewed in RViz as a series of arrows showing the robot’s intended path.
  3. Start the Mission: When the user is ready, they publish a simple message to the /path_ready topic. This is the “go” signal.
  4. Execute: The node enters its FOLLOW_PATH state, takes the first waypoint from its queue, and sends it as a goal to move_base using actionlib.
  5. Repeat: When move_base reports success, the node waits for a specified wait_duration (if any), then sends the next waypoint. This continues until the queue is empty.
  6. Clear: A message to /path_reset can clear the queue at any time.

This package is a perfect example of the ROS philosophy: a small, modular node that does one job extremely well. It’s an essential tool for anyone, from students in the University Rover Challenge (a listed use case) to researchers who need to automate complex navigation tasks.

If you’re working with ROS navigation, this project is a must-see for turning a simple point-to-point robot into a fully autonomous patrol machine. Check it out on GitHub here!

<
Previous Post
Automatic Number Plate Recognition in Python: A Look at the “Vehicle Logger” Project
>
Next Post
Indoor_fingerprinting