A concise Nav2 tuning checklist

When Nav2 tuning feels mysterious, the problem is usually observability. Before changing parameters, verify the robot model, transforms, odometry, and map quality.

Quick preflight

ros2 run tf2_tools view_frames
ros2 topic hz /odom
ros2 topic echo /diagnostics --once
ros2 launch nav2_bringup navigation_launch.py use_sim_time:=true

Parameter groups

Tune one group at a time. Start with the footprint and costmap inflation, then the global planner, then the local controller.

local_costmap:
  local_costmap:
    ros__parameters:
      footprint: "[[0.32, 0.24], [0.32, -0.24], [-0.32, -0.24], [-0.32, 0.24]]"
      inflation_layer:
        inflation_radius: 0.65
        cost_scaling_factor: 3.0

Record each run with a bag file and keep the changed parameters in source control.