Local Level Lab

Airborne measurements. Open methods.

Connect the qualified body-source bound to dynamic transport

The transport study failed qualification: four broad globe controls reached the unchanged per-bound time limit. Preserve its 20 successes and four failures; do not restart any identity or extend the frozen limits.

The separate cached implementation now passes all 24 controls with unchanged rational bounds and numerical limits. Continue with its source-integration plan. Keep this earlier failed qualification unchanged.

The next implementation needs a source-state adapter, not another change to termination tolerances or physical calibration bounds:

  1. Track the body-rotation prefix and its complete 27-column reference map at packet midpoints and endpoints. Use the same clipped clock fractions and preceding-packet boundary convention as the native map. An endpoint proof cannot be reused for all earlier prefixes.
  2. Obtain A=R f from the verified training-conditioned acceleration range, retaining the allowed gravity interval. Bound the literal installation matrix's norm defect; do not normalize it.
  3. Construct B=R(P−I)f for each actual prefix. Apply the qualified vertical norm relation and form an affine-error radius about the specified reference row. Include center or row discrepancies explicitly. Keep all columns.
  4. Use (Q−I)(A+B)+B for the first-force source. Retain the calibrated differences of every later force sample. Carry component bounds and useful global norm bounds together through the qualified transport path.
  5. Freeze a finite generated-source comparison against the actual saved training-conditioned baseline. Report unsupported norm branches and failed improvements. Preserve the original limits and every failed identity.
  6. After native center/derivative/containment controls qualify the complete candidate, freeze a separate full prediction comparison. Keep the existing width target; a component improvement alone does not justify aircraft fits.

Use separate new implementation files and study freezes. Preserve the current transport candidate and the earlier source-reference evidence. No observed aircraft fitting or protected ESA confirmation data is needed for this work.