Asking the fleet what it is doing…
monad-knowledge Wi-Fi sensing lab · FIIT STU
Campaign session

01KVZRDA8V0RAN269E4S0JAQFV

finished 2026-06-25 16:04:35.740342+00:00 → 2026-06-25 16:05:47.540892+00:00 · 12 runs · supervisor: none

“Powered link-as-unit re-reduction of the same 12 runs (N=40 link-observations vs 6 floors) REFUTES the wall-count-on-path operationalisation in silico: pooled ρ=+0.14, within-floor partial ρ=+0.14, LOS-vs-NLOS defeater d=-0.13 p≈0.88. Floor area is the stronger predictor (ρ=-0.54); topology does not beat area. layout-topology-fingerprint-discriminability demoted plausible→speculative (literature analogy keeps it from refuted; real-layout is the deciding test). BLE 2.49× more discriminable than CSI per link (corroborates c-ble-csi-coregistration).”

Archive snapshot, as of 23 h ago — the run corpus is rebuilt once a day, so this page is not a live reading. The fleet panel is the live one; it refreshes every 30 s.

Success criteria

CriterionResolved
Re-reduced the existing 12 attached runs with the link-as-unit reduction (csi_topology_linkunit.py): wall-count-on-path computed per link from each run's scene.json (no PostGIS), per-link discriminability = std(mean_amp_db) over the 80 placement frames. 40 link-observations. yes
Area-decorrelated test executed: within-floor partial correlation (floor-mean-demeaned) ρ=+0.14 — the wall-count→discriminability relation is null even after removing the area-confounded floor mean. Pooled area ρ=-0.54 dominates; topology does NOT beat area. yes
Defeater tested at link level (LOS wall-count=0 vs NLOS≥1), now powered: Cohen's d=-0.13, Welch p≈0.88 (n=4 LOS / 36 NLOS) → REFUTED. Honest bound: wall-count spread is limited (most Tx-Rx pairs cross exactly 1 wall), so area-dominance + within-floor null carry the verdict more than the wall-count axis alone; tests one operationalisation, not the full graph-metric set. yes
CSI-vs-BLE per-link discriminability: BLE 2.49× CSI (median ratio), reaffirming c-ble-csi-coregistration's ~2× BLE separability. yes

Synthesis

This session re-reduces the same 12 runs as 01KVXT9B31TF088GH5TE539QZN with the unit of analysis switched from the floor (N=6, underpowered, area∼diameter confounded) to the link (N=40), executing the follow-on design the prior session called for.

Verdict: the wall-count-on-path operationalisation of layout-topology-fingerprint-discriminability is REFUTED in silico; floor area is the operative predictor. Strength demoted plausiblespeculative.

Method

Reduction csi_topology_linkunit.py (monad-free, --selftest passes). Per-link topology metric = wall-count-on-path, the number of wall segments a link's Tx→Rx line-of-sight crosses, computed by 2D segment-intersection from each run's own inputs/scene.json (Tx/Rx coordinates + wall list) — no PostGIS, fully self-contained per run. Per-link discriminability = std (dB) of mean_amp_db across the 80 placement frames (how much a single moving body's position swings that link's amplitude). −240 dB "no path" sentinel clipped to −120 dB first. 3 Rx × 6 floors × 2 seeds = 40 link-observations.

Result

test value reading
wall-count-on-path → D (pooled ρ) +0.14 no relation
wall-count-on-path → D (within-floor partial ρ) +0.14 null even after removing the area-confounded floor mean
floor (bbox) area → D (pooled ρ) −0.54 area is the stronger predictor; topology does not beat area
LOS-vs-NLOS defeater (wall-count 0 vs ≥1) d=−0.13, p≈0.88 (n=4/36) REFUTED
BLE/CSI per-link discriminability ratio 2.49× BLE more separable (reaffirms c-ble-csi-coregistration)

Combined with the prior floor-level finding (required sniffer fraction k* best predicted by area, ρ=−0.74), the in-silico evidence now favours the area + AP-count baseline (adeel2019_ada0) over the topology hypothesis.

Honest bounds

  • Limited wall-count spread: most Tx–Rx pairs cross exactly one wall in this corpus, so the area-dominance + within-floor null carry the verdict more than the wall-count axis on its own.
  • One operationalisation: this tests wall-count-on-path + per-link amplitude std, not the full graph-metric set (the prior weak floor-level mean_degree ρ=+0.37 survives, unpowered).
  • Still in-silico: single-body, uniform-material; the multipath-richness confound (c-csi-fidelity-material) and real-layout transfer are untouched. The literature analogy (jung2025_3f2e betweenness contamination, depatla2018_97a9 shape>area, zhen2022_bb0b adaptive placement) is what keeps the hypothesis from refuted rather than speculative; the deciding evidence is a small real-layout fingerprint study, folded into the IP-106 testbed.

A cheap, valuable negative: it simplifies the sniffer-count deployment story to area + AP count, and removes a speculative arm from the system-design chain. Recorded on layout-topology-fingerprint-discriminability. Reduction monad_knowledge/notebooks/python/csi_topology_linkunit.py; fig under the session prefix + _attachments/topology-linkunit/out/.

Attached runs

Run Gate Purpose Replay
15KJNV42 csi-link-resplan-12419_s0
NZDF4R05 csi-link-resplan-12419_s1
3ZB7DWWE csi-link-resplan-147440_s0
6PX603JG csi-link-resplan-147440_s1
CBTZHMM3 csi-link-resplan-7421_s0
76KBRZXP csi-link-resplan-7421_s1
ZN4YG9X8 csi-link-resplan-16157_s0
QZKRPGK4 csi-link-resplan-16157_s1
HPTD1TDQ csi-link-resplan-12439-multiroom_s0
5Z70QWW2 csi-link-resplan-12439-multiroom_s1
SWMNAVA6 csi-link-resplan-1374_s0
JHG5WSQH csi-link-resplan-1374_s1