mirror of
https://github.com/ruvnet/RuView
synced 2026-07-31 18:51:42 +00:00
rUv
650612e5a2
The workspace DSP (vital_signs, multistatic, pose_tracker, tomography) implicitly assumes a forward model that maps scatterer geometry to per-subcarrier phase shifts. Nobody had written it down. This tick makes it explicit. Closed-form first-Fresnel-zone radius + point-scatterer path-delta + per-subcarrier phase prediction over 802.11n/ac 20 MHz channels (52 subcarriers, 312.5 kHz spacing). Pure NumPy demo + JSON output for downstream consumers. Headline numbers: - 5 m link first-Fresnel radius @ midpoint: 40 cm (2.4 GHz), 27 cm (5 GHz) - Inside zone-1: phase spread <0.5 deg across 52 subcarriers (band-flat) - Outside zone-1: phase spread up to 16 deg (band-dispersed) This unifies R5 + R6: R5's experimentally measured band-spread top subcarriers is exactly what the Fresnel forward model predicts for zone-1 occupancy. Closes the loop on three earlier threads: - R7 (mincut adversarial) gets a precise definition of 'physically inconsistent' instead of a learned classifier - R10 (foliage range) needs to retract 100 m sparse estimate to ~70 m to account for Fresnel-zone obstruction - R12 (eigenshift negative result) gets its revision basis: PABS over Fresnel-grounded forward operator Honest scope: point-scatterer only, first Fresnel only, frequency-flat reflectivity, LOS-only (no multipath). The scalar version is the right first-order approximation; volume-integral / multi-zone / multipath extensions catalogued as R6.1+R6.2 follow-ups. Coordination: ticks/tick-8.md, no PROGRESS.md edit.
43 lines
2.5 KiB
Markdown
43 lines
2.5 KiB
Markdown
# Tick 8 — 2026-05-22 05:25 UTC
|
|
|
|
**Thread:** R6 (Fresnel forward model)
|
|
**Verdict:** Working closed-form forward model + numpy demo. Bedrock physics that the entire `wifi-densepose-signal` DSP pipeline implicitly assumes is now explicit.
|
|
|
|
## What shipped
|
|
|
|
- `examples/research-sota/r6_fresnel_zone.py` — pure-numpy Fresnel-zone radius + per-subcarrier phase prediction. Four canonical scenarios over 802.11n/ac 20 MHz channels (52 subcarriers, 312.5 kHz spacing).
|
|
- `examples/research-sota/r6_fresnel_results.json` — machine-readable predictions.
|
|
- `docs/research/sota-2026-05-22/R6-fresnel-forward-model.md` — research note with the model, the demo headline numbers, what it gives each existing workspace module, R12's revision path with a basis, R10 range correction, honest scope.
|
|
|
|
## Headline numbers
|
|
|
|
**First Fresnel envelope (the "channel of maximum sensitivity"):**
|
|
|
|
| Link | 2.4 GHz @ midpoint | 5 GHz @ midpoint |
|
|
|---|---:|---:|
|
|
| 2 m | 25 cm | 17 cm |
|
|
| 5 m | **40 cm** | 27 cm |
|
|
| 10 m | 56 cm | 39 cm |
|
|
|
|
A typical bedroom 5 m WiFi link has a ~40 cm wide ellipsoid where human occupancy dominates the CSI. Outside that, you're picking up only diffracted edge contributions.
|
|
|
|
**Per-subcarrier phase predictions** confirm what R5 measured experimentally: inside zone-1, phase spread across 20 MHz is < 0.5° (band-flat); outside zone-1, spread grows to 15° (band-dispersed). R5's band-spread top-subcarriers are now physically explained, not just measured.
|
|
|
|
## Why this matters for the research loop
|
|
|
|
Three earlier threads were forced to **bootstrap from data** because no forward model existed:
|
|
|
|
- **R7** (mincut adversarial) — could only detect inconsistency, not predict expected. With R6, "physically inconsistent" has a precise definition: residual ≥ noise floor on all links simultaneously.
|
|
- **R10** (foliage range) — used FSPL + ITU foliage but ignored Fresnel-zone obstruction. R6 says the 100 m sparse-foliage range should be retracted to ~70 m (zone obstruction adds ~30% discount).
|
|
- **R12** (eigenshift, negative result) — failed because SVD spectrum loses spatial structure. R6's forward operator is the basis that R12's PABS revision needs.
|
|
|
|
## Coordination
|
|
|
|
Tick-8 via `ticks/tick-8.md`. No PROGRESS.md edit. Branch `research/sota-r6-fresnel-forward`.
|
|
|
|
## Remaining threads
|
|
|
|
R1 (ToA multistatic), R2 (room field model — partly subsumed by R6+R12 path), R3 (cross-room re-ID), R4 (federated learning), R11 (through-bulkhead maritime), R13 (contactless BP), R15 (RF biometric across rooms).
|
|
|
|
~6.6h to cron stop (12:00 UTC).
|