docs: Update ADR-015 with verified dataset specs from research

Corrects MM-Fi antenna config (1 TX / 3 RX not 3x3), adds Wi-Pose
as secondary dataset (exact 3x3 hardware match), updates subcarrier
compatibility table, promotes status to Accepted, adds proof
verification protocol and Rust implementation plan.

https://claude.ai/code/session_01BSBAQJ34SLkiJy4A8SoiL4

docs/adr/ADR-015-public-dataset-training-strategy.md

@@ -2,7 +2,7 @@
 
 ## Status
 
-Proposed
+Accepted
 
 ## Context
 
@@ -25,119 +25,156 @@ the camera is removed. This means any dataset that provides *either* ground-trut
 pose annotations *or* synchronized RGB frames (from which a teacher can generate
 labels) is sufficient for training.
 
+### 56-Subcarrier Hardware Context
+
+The system targets 56 subcarriers, which corresponds specifically to **Atheros 802.11n
+chipsets on a 20 MHz channel** using the Atheros CSI Tool. No publicly available
+dataset with paired pose annotations was collected at exactly 56 subcarriers:
+
+| Hardware | Subcarriers | Datasets |
+|----------|-------------|---------|
+| Atheros CSI Tool (20 MHz) | **56** | None with pose labels |
+| Atheros CSI Tool (40 MHz) | **114** | MM-Fi |
+| Intel 5300 NIC (20 MHz) | **30** | Person-in-WiFi, Widar 3.0, Wi-Pose, XRF55 |
+| Nexmon/Broadcom (80 MHz) | **242-256** | None with pose labels |
+
+MM-Fi uses the same Atheros hardware family at 40 MHz, making 114→56 interpolation
+physically meaningful (same chipset, different channel width).
+
 ## Decision
 
-Use MM-Fi as the primary training dataset, supplemented by XRF55 for additional
+Use MM-Fi as the primary training dataset, supplemented by Wi-Pose (NjtechCVLab)
-diversity, with a teacher-student pipeline for any dataset that lacks dense pose
+for additional diversity. XRF55 is downgraded to optional (Kinect labels need
-annotations but provides RGB video.
+post-processing). Teacher-student pipeline fills in DensePose UV labels where
+only skeleton keypoints are available.
 
 ### Primary Dataset: MM-Fi
 
 **Paper:** "MM-Fi: Multi-Modal Non-Intrusive 4D Human Dataset for Versatile Wireless
-Sensing" (NeurIPS 2023 Datasets Track)
+Sensing" (NeurIPS 2023 Datasets & Benchmarks)
-**Repository:** https://github.com/ybCliff/MM-Fi
+**Repository:** https://github.com/ybhbingo/MMFi_dataset
-**Size:** 40 volunteers × 27 action classes × ~320,000 frames
+**Size:** 40 subjects × 27 action classes × ~320,000 frames, 4 environments
 **Modalities:** WiFi CSI, mmWave radar, LiDAR, RGB-D, IMU
-**CSI format:** 3 Tx × 3 Rx antennas, 114 subcarriers, 100 Hz sampling rate,
+**CSI format:** **1 TX × 3 RX antennas**, 114 subcarriers, 100 Hz sampling rate,
-IEEE 802.11n 5 GHz, raw amplitude + phase
+5 GHz 40 MHz (TP-Link N750 with Atheros CSI Tool), raw amplitude + phase
-**Pose annotations:** 17-keypoint COCO skeleton (from RGB-D ground truth)
+**Data tensor:** [3, 114, 10] per sample (antenna-pairs × subcarriers × time frames)
+**Pose annotations:** 17-keypoint COCO skeleton in 3D + DensePose UV surface coords
 **License:** CC BY-NC 4.0
-**Why primary:** Largest public WiFi CSI + pose dataset; raw amplitude and phase
+**Why primary:** Largest public WiFi CSI + pose dataset; richest annotations (3D
-available (not just processed features); antenna count (3×3) is compatible with the
+keypoints + DensePose UV); same Atheros hardware family as target system; COCO
-existing `CSIProcessor` configuration; COCO keypoints map directly to the
+keypoints map directly to the `KeypointHead` output format; actively maintained
-`KeypointHead` output format.
+with NeurIPS 2023 benchmark status.
 
-### Secondary Dataset: XRF55
+**Antenna correction:** MM-Fi uses 1 TX / 3 RX (3 antenna pairs), not 3×3.
+The existing system targets 3×3 (ESP32 mesh). The 3 RX antennas match; the TX
+difference means MM-Fi-trained weights will work but may benefit from fine-tuning
+on data from a 3-TX setup.
 
-**Paper:** "XRF55: A Radio-Frequency Dataset for Human Indoor Action Recognition"
+### Secondary Dataset: Wi-Pose (NjtechCVLab)
-(ACM MM 2023)
+
-**Repository:** https://github.com/aiotgroup/XRF55
+**Paper:** CSI-Former (MDPI Entropy 2023) and related works
-**Size:** 55 action classes, multiple subjects and environments
+**Repository:** https://github.com/NjtechCVLab/Wi-PoseDataset
-**CSI format:** WiFi CSI + UWB radar, 3 Tx × 3 Rx, 30 subcarriers
+**Size:** 12 volunteers × 12 action classes × 166,600 packets
-**Pose annotations:** Skeleton keypoints from Kinect
+**CSI format:** 3 TX × 3 RX antennas, 30 subcarriers, 5 GHz, .mat format
+**Pose annotations:** 18-keypoint AlphaPose skeleton (COCO-compatible subset)
 **License:** Research use
-**Why secondary:** Different environments and action vocabulary increase
+**Why secondary:** 3×3 antenna array matches target ESP32 mesh hardware exactly;
-generalization; 30 subcarriers requires subcarrier interpolation to match the
+fully public; adds 12 different subjects and environments not in MM-Fi.
-existing 56-subcarrier config.
+**Note:** 30 subcarriers require zero-padding or interpolation to 56; 18→17
+keypoint mapping drops one neck keypoint (index 1), compatible with COCO-17.
 
-### Excluded Datasets and Reasons
+### Excluded / Deprioritized Datasets
 
-| Dataset | Reason for exclusion |
+| Dataset | Reason |
-|---------|---------------------|
+|---------|--------|
-| RF-Pose / RF-Pose3D (MIT) | Uses 60 GHz mmWave, not 2.4/5 GHz WiFi CSI; incompatible signal physics |
+| RF-Pose / RF-Pose3D (MIT) | Custom FMCW radio, not 802.11n CSI; incompatible signal physics |
-| Person-in-WiFi (CMU 2019) | Amplitude only, no phase; not publicly released |
+| Person-in-WiFi (CMU 2019) | Not publicly released (IRB restriction) |
-| Widar 3.0 | Gesture recognition only, no full-body pose |
+| Person-in-WiFi 3D (CVPR 2024) | 30 subcarriers, Intel 5300; semi-public access |
-| NTU-Fi | Activity labels only, no pose keypoints |
+| DensePose From WiFi (CMU) | Dataset not released; only paper + architecture |
-| WiPose | Limited release; superseded by MM-Fi |
+| Widar 3.0 | Gesture labels only, no full-body pose keypoints |
+| XRF55 | Activity labels primarily; Kinect pose requires email request; lower priority |
+| UT-HAR, WiAR, SignFi | Activity/gesture labels only, no pose keypoints |
 
 ## Implementation Plan
 
-### Phase 1: MM-Fi Loader
+### Phase 1: MM-Fi Loader (Rust `wifi-densepose-train` crate)
 
-Implement a `PyTorch Dataset` class that:
+Implement `MmFiDataset` in Rust (`crates/wifi-densepose-train/src/dataset.rs`):
-- Reads MM-Fi's HDF5/numpy CSI files
+- Reads MM-Fi numpy .npy files: amplitude [N, 3, 3, 114] (antenna-pairs laid flat), phase [N, 3, 3, 114]
-- Resamples from 114 subcarriers → 56 subcarriers (linear interpolation along
+- Resamples from 114 → 56 subcarriers (linear interpolation via `subcarrier.rs`)
-  frequency axis) to match the existing `CSIProcessor` config
+- Applies phase sanitization using SOTA algorithms from `wifi-densepose-signal` crate
-- Normalizes amplitude and unwraps phase using the existing `PhaseSanitizer`
+- Returns typed `CsiSample` structs with amplitude, phase, keypoints, visibility
-- Returns `(amplitude, phase, keypoints_17)` tuples
+- Validation split: subjects 33–40 held out
 
-### Phase 2: Teacher-Student Labels
+### Phase 2: Wi-Pose Loader
 
-For samples where only skeleton keypoints are available (not full DensePose UV maps):
+Implement `WiPoseDataset` reading .mat files (via ndarray-based MATLAB reader or
-- Run Detectron2 DensePose on the paired RGB frames to generate `(part_labels,
+pre-converted .npy). Subcarrier interpolation: 30 → 56 (zero-pad high frequencies
-  u_coords, v_coords)` pseudo-labels
+rather than interpolate, since 30-sub Intel data has different spectral occupancy
-- Cache generated labels to avoid recomputation during training epochs
+than 56-sub Atheros data).
-- This matches the training procedure in the original CMU paper
 
-### Phase 3: Training Pipeline
+### Phase 3: Teacher-Student DensePose Labels
 
-- **Loss:** Combined keypoint heatmap loss (MSE) + DensePose part classification
+For MM-Fi samples that provide 3D keypoints but not full DensePose UV maps:
-  (cross-entropy) + UV regression (Smooth L1) + transfer loss against teacher
+- Run Detectron2 DensePose on paired RGB frames to generate `(part_labels, u_coords, v_coords)`
-  RGB backbone features
+- Cache generated labels as .npy alongside original data
-- **Optimizer:** Adam, lr=1e-3, milestones at 48k and 96k steps (paper schedule)
+- This matches the training procedure in the CMU paper exactly
+
+### Phase 4: Training Pipeline (Rust)
+
+- **Model:** `WiFiDensePoseModel` (tch-rs, `crates/wifi-densepose-train/src/model.rs`)
+- **Loss:** Keypoint heatmap (MSE) + DensePose part (cross-entropy) + UV (Smooth L1) + transfer (MSE)
+- **Metrics:** PCK@0.2 + OKS with Hungarian min-cost assignment (`crates/wifi-densepose-train/src/metrics.rs`)
+- **Optimizer:** Adam, lr=1e-3, step decay at epochs 40 and 80
 - **Hardware:** Single GPU (RTX 3090 or A100); MM-Fi fits in ~50 GB disk
 - **Checkpointing:** Save every epoch; keep best-by-validation-PCK
 
-### Phase 4: Evaluation
+### Phase 5: Proof Verification
 
-- **Keypoints:** PCK@0.2 (Percentage of Correct Keypoints within 20% of torso size)
+`verify-training` binary provides the "trust kill switch" for training:
-- **DensePose:** GPS (Geodesic Point Similarity) and GPSM with segmentation mask
+- Fixed seed (MODEL_SEED=0, PROOF_SEED=42)
-- **Held-out split:** MM-Fi subjects 33-40 (20%) for validation; no test-set leakage
+- 50 training steps on deterministic SyntheticDataset
+- Verifies: loss decreases + SHA-256 of final weights matches stored hash
+- EXIT 0 = PASS, EXIT 1 = FAIL, EXIT 2 = SKIP (no stored hash)
 
 ## Subcarrier Mismatch: MM-Fi (114) vs System (56)
 
-MM-Fi captures 114 subcarriers at 5 GHz with 40 MHz bandwidth. The existing system
+MM-Fi captures 114 subcarriers at 5 GHz with 40 MHz bandwidth (Atheros CSI Tool).
-is configured for 56 subcarriers. Resolution options in order of preference:
+The system is configured for 56 subcarriers (Atheros, 20 MHz). Resolution options:
 
-1. **Interpolate MM-Fi → 56** (recommended for initial training): linear interpolation
+1. **Interpolate MM-Fi → 56** (chosen for Phase 1): linear interpolation preserves
-   preserves spectral envelope, fast, no architecture change needed
+   spectral envelope, fast, no architecture change needed
-2. **Reconfigure system → 114**: change `CSIProcessor` config; requires re-running
+2. **Train at native 114**: change `CSIProcessor` config; requires re-running
-   `verify.py --generate-hash` to update proof hash
+   `verify.py --generate-hash` to update proof hash; future option
-3. **Train at native 114, serve at 56**: separate train/inference configs; adds
+3. **Collect native 56-sub data**: ESP32 mesh at 20 MHz; best for production
-   complexity
 
-Option 1 is chosen for Phase 1 to unblock training immediately.
+Option 1 unblocks training immediately. The Rust `subcarrier.rs` module handles
+interpolation as a first-class operation with tests proving correctness.
 
 ## Consequences
 
 **Positive:**
-- Unblocks end-to-end training without hardware collection
+- Unblocks end-to-end training on real public data immediately
-- MM-Fi's 3×3 antenna setup matches this system's target hardware (ESP32 mesh, ADR-012)
+- MM-Fi's Atheros hardware family matches target system (same CSI Tool)
-- 40 subjects with 27 action classes provides reasonable diversity for a first model
+- 40 subjects × 27 actions provides reasonable diversity for first model
+- Wi-Pose's 3×3 antenna setup is an exact hardware match for ESP32 mesh
 - CC BY-NC license is compatible with research and internal use
+- Rust implementation integrates natively with `wifi-densepose-signal` pipeline
 
 **Negative:**
 - CC BY-NC prohibits commercial deployment of weights trained solely on MM-Fi;
   custom data collection required before commercial release
-- 114→56 subcarrier interpolation loses some frequency resolution; acceptable for
+- MM-Fi is 1 TX / 3 RX; system targets 3 TX / 3 RX; fine-tuning needed
-  initial training, revisit in Phase 2
+- 114→56 subcarrier interpolation loses frequency resolution; acceptable for v1
-- MM-Fi was captured in controlled lab environments; expect accuracy drop in
+- MM-Fi captured in controlled lab environments; real-world accuracy will be lower
-  complex real-world deployments until fine-tuned on domain-specific data
+  until fine-tuned on domain-specific data
 
 ## References
 
-- He et al., "MM-Fi: Multi-Modal Non-Intrusive 4D Human Dataset" (NeurIPS 2023)
+- Yang et al., "MM-Fi: Multi-Modal Non-Intrusive 4D Human Dataset" (NeurIPS 2023) — arXiv:2305.10345
-- Yang et al., "DensePose From WiFi" (arXiv 2301.00250, CMU 2023)
+- Geng et al., "DensePose From WiFi" (CMU, arXiv:2301.00250, 2023)
+- Yan et al., "Person-in-WiFi 3D" (CVPR 2024)
+- NjtechCVLab, "Wi-Pose Dataset" — github.com/NjtechCVLab/Wi-PoseDataset
 - ADR-012: ESP32 CSI Sensor Mesh (hardware target)
 - ADR-013: Feature-Level Sensing on Commodity Gear
 - ADR-014: SOTA Signal Processing Algorithms