How arghyachatterjee/april-tag-based-autonomous-navigation works
Project type and purpose
• This is a ROS (Robot Operating System) package for AR-tag (AprilTag) based autonomous navigation.
• Purpose: allow a robot with a camera to detect AprilTag landmarks, publish their poses into the TF tree, and support runtime management of those landmarks for navigation and obstacle avoidance.
File structure
• Top-level ROS build files: CMakeLists.txt, package.xml, setup.py
• Directories:
• launch/ : ROS launch files to bring up each node
• scripts/ : executable Python scripts for three ROS nodes
• landmarks/ : YAML config for defining tag‐based landmarks
• images/ : screenshots for README documentation
• Key scripts:
• landmark_finder
• landmark_manager
• tf_map_to_landmark
• Config:
• landmarks/landmarks.yaml
• launch/*.launch take args to parameterize topics, frames, config paths
README for additional insights
• Architecture is modular:
• Landmark and LandmarkCollection modules handle data structures and transform math
• Three ROS nodes offer CLI management, vision‐based detection, and TF publishing
• Integration points:
• AprilTag detection library from interbotix_perception_modules
• ROS TF tree for coordinate transforms
• Dynamic reconfiguration through YAML + parameters
a. Main components
1. Robot camera (sensor)
2. landmark_finder node
3. tf_map_to_landmark node
4. landmark_manager node
5. landmarks.yaml (configuration store)
6. ROS Master & Parameter Server
7. TF Tree (map → landmarks → robot base_link)
8. AprilTag detection module (external library)
b. Relationships & data flow
• Camera publishes raw images & camera_info → landmark_finder subscribes
• landmark_finder runs AprilTag detector → publishes detected tag poses into /tf
• tf_map_to_landmark reads landmarks.yaml & subscribes to tag frames → computes & publishes map→landmark transforms into /tf
• landmark_manager reads/writes landmarks.yaml via CLI → updates Parameter Server or writes file on disk
• Navigation stack (external) reads TF tree to plan paths to landmarks
c. Architectural patterns & design principles
• ROS node separation: sensing, transform management, and configuration kept in separate processes
• TF‐centric architecture: all coordinate transforms flow through a common TF tree
• Configuration as code: YAML files and ROS parameters for dynamic re‐configuration
d. Key technologies & libraries
• ROS (rospy, tf2_ros, dynamic_reconfigure)
• AprilTag module from interbotix_perception_modules
• YAML for landmark definitions
• ROS launch system for composition