AR Tag Landmark-Based Autonomous Navigation

This ROS package enables autonomous navigation by using AR tags (AprilTags) as landmarks. Robots equipped with cameras can navigate accurately to specified locations while avoiding obstacles by identifying predefined AR tags.

Repository Workflow

Features

• AR Tag Landmarks: Use AprilTags as navigational points.
• Dynamic Landmark Configuration: Easily manage landmarks using YAML files.
• Integration with TF: Automatically manages and publishes transforms to the ROS TF tree.
• CLI Management: Define and update landmarks through a simple command-line interface.

---

Installation

Ensure dependencies are installed:

• AprilTag Module

Clone this repository:

cd ~/catkin_ws/src
git clone https://github.com/ArghyaChatterjee/april-tag-based-autonomous-navigation.git
cd ~/catkin_ws
catkin_make

---

Landmark Configuration

Landmarks are defined in YAML configuration files (example):

landmark:
  mounted: true          # True if AR tag is wall-mounted
  offset: 0.5            # Distance offset from the tag for navigation goal
  id: 1                  # AprilTag ID
  label: "Entrance"       # Landmark label
  set: false             # Automatically managed; do not manually set

---

Modules and Nodes

Modules

• Landmark: Stores landmark data, computes transforms, and manages landmark poses.
• LandmarkCollection: Manages collections of landmarks with additional utilities.

Nodes

• landmark_manager:

  • CLI to define landmarks interactively.

• landmark_finder:

  • Periodically captures images from a camera to detect landmarks, publishing their poses to the TF tree.

• tf_map_to_landmark:

  • Publishes transforms between a fixed frame (usually map) and landmarks to organize the TF tree.

---

Usage

Include necessary launch files in your main ROS launch configuration from the launch directory.

Launch File Arguments

landmark.launch

Argument	Description	Default
apriltag_ns	Namespace for AprilTag nodes and parameters	apriltag
camera_color_topic	ROS topic for camera color images	camera/color/image_raw
camera_info_topic	ROS topic for camera info	camera/color/camera_info
use_cam	Launch RealSense camera package	true

landmark_manager.launch

Argument	Description	Default
landmark_config	Path to landmark YAML config file	landmarks.yaml
fixed_frame	Frame to transform landmark TFs	landmarks
obs_frame	Observation frame for landmarks	camera_color_optical_frame
standalone_tags	Individual AprilTags to search for	See landmark_manager.launch

---

Troubleshooting

Landmark Issues

• "File at <filepath> does not exist yet. No landmarks loaded."

  • Ensure the YAML file exists and is correctly formatted.

---

Future Work

• Special landmark labels triggering custom behaviors (e.g., docking procedures).

---

Acknowledgments

Thanks to Interbotix ROS for their open-source contributions.