Sensor Frame Mounting and Extrinsics

Source: ros2-copilot-skills sensor frame mounting skill

Why This Matters

Sensor extrinsics are the bridge between the physical robot and the software model of the robot. A few millimeters or degrees of error can show up as map distortion, bad obstacle placement, inconsistent fusion, or detections that seem to float in the wrong place.

Distilled Takeaways

• Extrinsics are not metadata. They are geometry that the rest of the stack depends on.
• Mounting accuracy matters most when multiple sensors or world-model steps need to agree spatially.
• Small alignment errors compound quickly across mapping, localization, and perception chains.
• URDF and TF should describe real mounting, not idealized mounting.

Practical Guidance

• Measure sensor placement carefully and revisit it after hardware changes.
• Validate extrinsics visually in RViz against real observed geometry.
• Treat unexplained map or perception offsets as possible extrinsic problems early.
• Document mounts and assumptions so future contributors know what changed.

Corroborating References

• REP 105 coordinate frame conventions
• tf2 tutorials

When to Read the Original Source

Go to the original skill when you want the practical sensor-mounting reminders and the specific extrinsic failure modes that show up most often on mobile robots.