890 B
Authored second field corridor
The fixed native grid completed all1,681 queries and retained its original failed-survey verdict. Offline comparison selected start X4890/Z11770 and end X4890/Z11870. The171 coarse samples inside that route plus40m margins range28.5625–28.5938m, at most0.1313m from the original lifted start reference. The previous candidate failed at0.31875m.
This is authored test data, not runtime fallback or qualification. The one new candidate must still pass every original2m terrain sample, full-volume obstacle trace, native model/rotor footprint and post-spawn single-use certificate before one finite allocation/control attempt. No terrain tolerance, corridor margin, node cap, landing rule or flight deadline is changed. Full terrain data and the deterministic ranking definition are included; coarse samples can miss intervening terrain and all obstacles.