This is an earlier experiment. Read the amendments below before treating its findings as current.
The question
Craft fly a band about ±12 m tall around their lane. Is that room the dodge planner needs, or is it where the pull back toward the lane (a gentle steering command) holds them?
What was tested
We switched that pull off entirely, sideways and vertical, and measured where craft fly over six layouts.
What was found then
Without the pull, ideal craft (they respond instantly) spread to about ±40 m, against ±12 m with it, and sat inside their lane box only 5–15 % of the time. Slow-responding craft climbed up to about 110 m above their lane. So the band is where the pull balances them, not something dodging needs. A side count in the busy sky showed fewer collisions without the pull; it was treated as a hint, to be tested properly next.
What changed
Nothing.
Corrections retained in the record
Earlier wording: in the still-air probe, the busy-airspace cell had fewer collisions without the pull-back (8–14 per run against 17–30)
Record correction: per-seed counts without the pull-back were 8/11/10/10/5/14, i.e. 5–14 per run
research/reviews/brief-c18.md
Numbers and source references
- 6
- seedsresearch/data/oos/freeband.md
- ±12 m
- spread with the default pull-backresearch/reviews/brief-c15.md
- −38 / +39 m
- ideal craft vertical spread over the run, pull-back offresearch/reviews/brief-c19.md
- −44 / +45 m
- same, second half of the runresearch/reviews/brief-c19.md
- 60 m
- lane altitude on the gridresearch/reviews/brief-c18.md
- 5–15 %
- time in the lane box, pull-back off, second halfresearch/reviews/brief-c19.md
- 110 m
- how far above their lane slow craft reach (+96…+111)research/reviews/brief-c18.md