FLIGHTOPS

← back to the bench

TEST COURSE

Same robot, same map. One of them is lost.

Re-enacted from two real Nav2 runs on 18 October 2026. In run B, AMCL listens on /lidar/scan while the lidar publishes /scan, so the robot never gets a map pose. FlightOps puts both runs side by side and writes down the difference.

SCENE WORKBENCH

Two runs. One cursor. The difference in writing.

A native desktop app that draws what the recordings show. Here, a healthy Nav2 run sits beside one where AMCL listens on the wrong lidar topic and the robot never gets a map pose.

Scene Workbench in Compare mode. Run A places the robot on the map. Run B is flagged NO MAP POSE. The difference panel reads map to odom from 3.13 seconds in A, never in B.
18 Oct 2026 · two real Nav2 runs
1

The fault, flagged. Run B never gets a map pose, and the app says so instead of drawing a guess.

2

Difference at cursor. map → odom from 3.13 s in A, never in B. Every differing signal, listed.

3

One timeline. Every topic, every gap, one cursor. Play, scrub or jump to a message.

4

Read-only by design. Simulation only, hardware off. No live robot commands, ever.

REPLAYCOMPARELIVEPreflight › Change › Review › Focused check › Commit › Plan › Gate › Paired runCTRL KLIGHT · DARK

NEXT DRAWER