Nav2 changes and improvements from SAR

[ErikCald]

All the changes and improvements made for SAR.

This pull request introduces several important updates across navigation, localization, and perception components to improve system robustness, path planning quality, and integration with Aruco marker workflows. The major changes include improved EKF configuration for odometry and pose fusion, updates to Aruco detection topics and message types, enhancements to navigation parameter tuning (especially smoothing and costmap behavior), and the addition of a visualization tool for search patterns.

Localization and Sensor Fusion Improvements:

• Updated EKF (Extended Kalman Filter) configurations to use /chassis_controller/wheel_odom instead of /drive/odom for odometry, and enabled additional velocity and angular velocity components for improved state estimation in both global and local configurations (gps_ekf_global.yaml, gps_ekf_local.yaml, indoor_aruco_boards_ekf_global.yaml, indoor_aruco_boards_ekf_local.yaml). Also, added fusion with ZED pose and switched EKFs from 2D to 3D mode (two_d_mode: false) for better accuracy. [1] [2] [3] [4] [5] [6]
• Lowered odometry republishing frequency from 10 Hz to 1 Hz to reduce computational load (republish_odometry.py).

Aruco Marker Detection Integration:

• Changed Aruco detection topic from /aruco_detections to /marker_detected and message type from Int32MultiArray to Int32 for compatibility with the latest detection node output (nav_to_gps_aruco.py). [1] [2] [3] [4] [5]
• Updated default Aruco dictionary to DICT_4X4_50 for board detection, likely to match the physical marker set (aruco_board_params.yaml).

Navigation Parameter Tuning:

• Increased global costmap publish frequency, added transform tolerance, and tuned planner parameters for tighter turns and smoother, more robust path generation. Switched to a constrained smoother plugin for path smoothing and adjusted its parameters for better obstacle avoidance and smoother paths (nav2.yaml). [1] [2] [3]
• Adjusted traversability filter weights to increase the influence of slope and roughness in elevation mapping, potentially making the robot more cautious in rough terrain (postprocessing_traversability.yaml).

New Features and Utilities:

• Added a new ROS2 node, search_pattern_viz_node, for visualizing lawnmower search patterns in RViz, aiding debugging and mission planning (search_pattern_viz.py, setup.py). [1] [2]
• Added a markdown file with launch and command-line instructions for navigation and perception stack startup (nav_launch.md).

Other Changes:

• Commented out the Spin behavior in the dynamic replanning behavior tree, possibly to disable unnecessary spinning during navigation (bt_swerve_dynamic_replanning.xml).
• Minor launch file import cleanup (rover.launch.py).

These changes collectively improve the system's localization accuracy, navigation smoothness, robustness to terrain, and compatibility with the latest Aruco detection workflows.

[Copilot]

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.

[ConnorNeed]

I don't see an explination but also I don't really care

[ConnorNeed]

Is this a debug thing or is there a production use?