Spatial diversity keeps corruption localized — the SBI precondition (#133)#147
Merged
Conversation
…133) The SBI sub-block-salvage layer only pays when corruption is localized within a frame; a deep fade smears it frame-wide and SBI recovers nothing. Spatial diversity cuts the SNR *variance* (deep fades are variance), so it moves frames out of the frame-wide (lost) bin into the localized (SBI-salvageable) one — improving SBI's precondition, not just delivery. - tools/precoder/spatial_sbi_sim.py (+ pytest): Monte-Carlo correlated-Rayleigh branches, combined at matched mean SNR to isolate the diversity/variance effect from the array/mean gain, classified clean / localized / lost. --sweep-rho and --self-test. - docs/fused-fec.md: "Spatial diversity feeds the same precondition" in the localization section + module-map entry. Result (matched mean, isolating variance — consistent with the measured static-≈-nil / mobile-pays finding): static/correlated (ρ=0.9) barely moves the frame-wide-loss bin (0.31 vs single-chain 0.33); mobile/decorrelated (ρ=0.2) collapses it (0.12) and lifts the SBI-salvageable fraction of corrupt frames 0.59 → 0.86. Corruption localization is a mobility effect. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
josephnef
added
enhancement
This was referenced Jul 2, 2026
Closed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
The last substantive cross-layer piece (#133): spatial diversity keeps
corruption localized, which is the precondition the fused-FEC sub-block-integrity
(SBI) salvage layer needs.
The mechanism
SBI salvages an FCS-failed frame by keeping its surviving CRC-guarded sub-blocks —
but only when corruption is localized. A deep fade drops SNR far enough that
the whole frame smears (frame-wide corruption) and SBI recovers nothing. Spatial
diversity cuts the SNR variance — and deep fades are variance — so it moves
frames out of the frame-wide (lost) bin into the localized (SBI-salvageable) one.
It improves SBI's precondition, not just raw delivery.
Model
Monte-Carlo correlated-Rayleigh branches (equicorrelation ρ), combined at
matched mean SNR to isolate the diversity/variance effect from MRC's array/mean
gain (that array gain is a separate delivery/energy effect). Each frame is
classified clean / localized / lost by its combined SNR.
Result (mean SNR 10 dB)
Static/correlated combining barely moves the frame-wide bin (consistent with the
measured "static MRC ≈ nil"); decorrelated/mobile combining collapses it and
lifts the salvageable fraction. Corruption localization is a mobility effect —
largest where a long-range link lives.
Contents
tools/precoder/spatial_sbi_sim.py(+ pytest):--sweep-rho,--self-test.docs/fused-fec.md: "Spatial diversity feeds the same precondition".Testing
--self-testgreen; 6 pytest pass. Pure-Python, consistent with the sibling sims.This completes the roadmap's substantive work (#128–#136); only the speculative
#137 beam-steer study remains.
🤖 Generated with Claude Code