frank/ruvnet--RuView
1
0
Fork 0
mirror of https://github.com/ruvnet/RuView synced 2026-07-13 15:53:18 +00:00
Code Issues Packages Projects Releases Wiki Activity

docs: precise calibration validation scope (capture+API+auth proven; clean enroll→train→infer not yet on-target)

Browse Source 
Aligns ADR-151 §7 + the appliance integration doc with the PR #989 scope
clarification: nothing has run a clean baseline → enroll → train → infer on
live CSI; the live breathing read used the stateless head, not a trained bank.
Adds --source-format adr018v6 to the backlog.

Co-Authored-By: claude-flow <ruv@ruv.net>
This commit is contained in:
ruv
2026-06-09 13:35:45 -04:00
parent 4ea51df8a3
commit 5e4260e2f7
2 changed files with 15 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/adr/ADR-151-room-calibration-specialist-training.md
+1 -1
View File
@@ -247,7 +247,7 @@ wifi-densepose room-status --room living-room
| **P5** | heartbeat + restlessness + anomaly specialists; `runtime.rs` mixture + veto + confidence gating | End-to-end RoomState on hardware; anomaly veto verified | ✅ Done (`runtime.rs`, CLI `room-watch`; breathing read live on COM8 ESP32) |
| **P6** | Baseline-drift `STALE` invalidation; SONA online adaptation; optional ADR-105 federation; HF teacher–student distillation | Drift marks bank STALE; AetherArena entry | ◐ Partial (STALE done; SONA/federation/HF-backbone = follow-ups) |
**Current status (2026-06-09):** Stages 1–5 implemented with *statistical* specialists (threshold/prototype/autocorrelation) — runnable and hardware-validated today. 48 unit tests. Open refinements: phase-based (vs amplitude) breathing carrier for stronger SNR, RVF/HNSW storage, and the ADR-150 frozen HF backbone the specialists would distill from.
**Current status (2026-06-09):** Stages 1–5 implemented with *statistical* specialists (threshold/prototype/autocorrelation). 54 unit tests (incl. multistatic), all passing under qemu-aarch64. **Validation scope is precise:** baseline capture + HTTP API + auth are proven on real CSI (Pi-5 nexmon, 6,813 frames; and an ESP32-S3). The one live runtime signal (breathing ~16–31 BPM via `room-watch`) used the *stateless* breathing head, **not** a trained bank; a clean empty-room `enroll → train-room → room-watch` loop has **not** yet run on-target — the headline remaining gap. Open refinements: `--source-format adr018v6` (drive from the Pi's own nexmon), phase-based breathing carrier, RVF/HNSW storage, and the ADR-150 frozen HF backbone the specialists would distill from.
Validation per CLAUDE.md: `cargo test --workspace --no-default-features` green; hardware verification on the ESP32-S3 (currently COM8) before any release; witness bundle regenerated if the proof surface changes.
docs/integration/calibration-appliance-integration.md
+14 -3
View File
@@ -212,9 +212,20 @@ exchange bank/model deltas, never raw CSI. Hardening already in place:
## 7. Status & validation
- **Implemented:** all 5 stages + multistatic fusion; CLI + Stage-1 HTTP API. **54 tests** (35 calibration + 19 CLI).
- **Hardware-validated** (ESP32-S3, COM8, `edge_tier=0`): baseline capture (120 frames → 52-subcarrier baseline); live parser → feature-extraction → mixture runtime detecting **breathing (~16–31 BPM)**; multistatic ingest grouping/fusing by `node_id`.
- **Known follow-ups (carry into the appliance backlog):** phase-based (vs mean-amplitude) breathing carrier for higher SNR; RVF/HNSW persistence (currently JSON); enroll/train HTTP endpoints; the ADR-150 Hailo backbone; true 2-node multistatic on hardware; ADR-105 federation across appliances.
- **Implemented:** all 5 stages + multistatic fusion; CLI + Stage-1 HTTP API (auth + path-traversal hardened). **54 tests** (35 calibration + 19 CLI), all passing under qemu-aarch64.
**Precise validation matrix (don't overstate this — no clean full calibration has run on-target yet):**
| Stage | Pi-5 (real nexmon→`0xC5110001`, 6,813 frames) | ESP32-S3 (COM8, `edge_tier=0`) | qemu / unit |
|---|---|---|---|
| baseline capture + HTTP API + **auth gate** | ✅ | ✅ (120-frame) | — |
| **clean** empty-room baseline | ❌ `motion_flagged` (artifact) | ❌ (occupied) | — |
| enroll → train-room | ❌ | ❌ (needs operator poses) | logic ✅ |
| runtime infer | ❌ on-target | ◐ single-node breathing ~16–31 BPM via the **stateless** head (not a trained bank) + node-id fusion | veto/STALE/fusion ✅ |
So capture + API + auth are proven on real CSI (both boxes); a **clean-baseline `enroll → train-room → room-watch`** loop is **not yet proven on-target** — that's the headline gap to close on the appliance.
- **Known follow-ups (appliance backlog):** `--source-format adr018v6` to drive calibration from the Pi's own nexmon (no ESP32/transcoder); a clean empty-room enroll→train→infer on-target; phase-based (vs mean-amplitude) breathing carrier; RVF/HNSW persistence (currently JSON); enroll/train HTTP endpoints; ADR-150 Hailo backbone; true 2-node multistatic; ADR-105 federation.
**Reference:** ADR-151 (`docs/adr/ADR-151-room-calibration-specialist-training.md`), ADR-135 (baseline),
ADR-029 (multistatic), ADR-150 (RF Foundation Encoder), ADR-105 (federation), ADR-147 (OccWorld/Hailo).