How arghyachatterjee/gps-waypoint-based-autonomous-navigation-in-ros works
Project type and purpose
• This is a full-stack robotics application built on ROS (Robot Operating System).
• Purpose: Use predefined or collected GPS waypoints to autonomously navigate a rover, fusing GPS, IMU, odometry, and lidar data for mapping and obstacle avoidance.
File structure
• Top‐level:
• CMakeLists.txt, package.xml – ROS packaging and build
• assets – static images
• launch/ & launch/navigation/ – ROS launch files to spin up sensors, nodes, controllers, navigation stacks, mapviz, safety, etc.
• params/ – YAML parameter files for EKF, filters, navsat transforms
• rviz_config/ – RViz configurations for different modes
• scripts/ – Python nodes for joystick control and waypoint interaction
• src/ – C++ nodes for heading calibration, waypoint following, data collection, plotting
• waypoint_files/ – sample waypoint outputs
• Build & deploy: catkin_make, shell scripts (run.sh, run_continuous.sh).
README for additional insights
• Describes hardware setup (Jetson TX2, IMU, ZED-F9P GPS, RPLidar).
• Dependencies: robot_localization (EKF), navsat_transform, GMapping, move_base, mapviz, rviz_satellite, rosboard.
• Custom nodes: gps_waypoint, gps_waypoint_continuous1/2, collect_gps_waypoints, calibrate_heading, plot_gps_waypoints, gps_waypoint_mapping.
Main components & their interactions
a. Hardware & Drivers (Sensors)
• GPS sensor → /navsat/fix
• IMU → /imu/data
• Wheel odometry → /odom
• Lidar → /scan
b. Localization Stack
• robot_localization EKF node (fuses IMU + odom + GPS → /odometry/filtered)
• navsat_transform node (converts /navsat/fix to /navsat/odom, projects lat/lon into map frame)
c. Mapping & Obstacle Detection
• GMapping SLAM node subscribes to /scan + /odometry/filtered → publishes /map
d. Path Planning & Control
• move_base
• global_planner uses /map, /odom, and goals
• local_planner uses /scan, dynamic reconfigure
e. Waypoint Management
• gps_waypoint* nodes read waypoint files, transform to map frame, publish goals to move_base
• collect_gps_waypoints node logs current GPS/odometry to a file
• calibrate_heading node sets initial yaw offset
f. Visualization & Monitoring
• RViz (rviz_config/*.rviz)
• Mapviz (mapviz launch)
• rosboard (http visualization of topics)
Architectural patterns & design principles
• ROS node‐topic architecture (loosely coupled, message-based)
• Separation of concerns:
• Sensing/Driver nodes
• Filtering/fusion (EKF)
• Coordinate transforms (navsat_transform)
• Mapping (GMapping)
• Planning (move_base + custom waypoint publishers)
• Data collection & visualization
• Launch files group and configure nodes per mission profile (mapping vs. waypoint navigation vs. joystick control).
Essential elements
• All custom ROS nodes, grouped by functionality
• External packages (robot_localization, navsat_transform, GMapping, move_base, Mapviz, RViz, rosboard)
• Sensor sources and produced topics
• Launch file groupings (e.g., gps_waypoint_nav.launch, joy_launch_control.launch)
• Parameter/YAML files attached to their nodes