The register
Experiments.
Every question the project has put to a room, and what came back. Each card names the capture sessions or simulation runs it rests on, and the archive checks that claim, so a result here is backed by named recordings and never by prose alone.
Showing 18 of 44.
| Identifier | Experiment | Standing | Kind | Evidence behind it | Last activity |
|---|---|---|---|---|---|
| csi-impairment-fit | exp-csi-calibration, finally — fitting the CSI impairment profile from a month of empty roomWhat does our real Wi-Fi card actually add to a clean signal, measured from a month of empty room? | running | simulation | 16 sim runs · 1 campaign session | 2026-08-28 |
| csi-drift-real-floor | CSI layout drift on the surveyed floor — does the synthetic (T*, ε) prior survive the real room?On the real library floor, can moving the furniture even reach the radio beam, and if not, should the drift curve be flat? | running | simulation | 12 sim runs · 1 campaign session | 2026-08-28 |
| csi-crowd-temporal | Time-resolved CSI under real crowd motion (coupled JuPedSim → Sionna)Does the wobble in the Wi-Fi signal really grow with the size of a moving crowd? | running | simulation | 92 sim runs · 7 campaign sessions | 2026-08-07 |
| csi-sim2real-transfer | Sim-to-real count transfer — does synthetic-impaired CSI pretraining beat real-only at a small real budget?Can a counter trained on cheap simulated data and finished on a handful of real measurements beat one trained on the handful alone? | running | simulation | 83 sim runs · 2 campaign sessions | 2026-07-15 |
| crowd-dynamics-replay | Crowd dynamics on real floors — persona-driven motion, seat-seeking, egress, replaysDo the simulated people move through real floor plans the way real crowds do? | running | simulation | 177 sim runs · 9 campaign sessions | 2026-07-14 |
| ap-placement-coverage-counting | The floorplan as a placement oracle — where do APs both cover the space and count the crowd?Do the best places to put Wi-Fi radios for coverage also happen to be the best places for counting the crowd? | running | simulation | 47 sim runs · 5 campaign sessions | 2026-07-13 |
| csi-recalibration-trigger | Event-driven recalibration — can a cross-modal drift statistic beat a fixed clock?Can the disagreement between a Wi-Fi count and a Bluetooth count tell us when to recalibrate, instead of a fixed timer? | running | simulation | 4 campaign sessions | 2026-07-07 |
| csi-topology-vs-area | Topology vs area for CSI occupancy — what drives fingerprint discriminability and sniffer count?Is it the shape of a floor plan or simply its size that decides how many radios you need to count its people? | running | simulation | 6 campaign sessions | 2026-07-07 |
| ble-anchor-placement | BLE anchor placement — does topology-informed placement recover graded count?If Bluetooth anchors are concentrated in the room everyone passes through, can they count people rather than only notice them? | running | simulation | 2 campaign sessions | 2026-07-06 |
| csi-layout-drift | CSI layout drift — a synthetic (T*, ε) prior for recalibration cadenceHow fast does a Wi-Fi people counter go wrong as the furniture is moved, and is the decay regular enough to schedule recalibration by? | running | simulation | 3 campaign sessions | 2026-07-02 |
| csi-band-calibration | CSI band calibration — can the coupled sim reproduce the WiMANS 2.4-vs-5 GHz occupancy gap?Does the simulator show the same 5 GHz advantage for counting people that real hardware shows? | running | simulation | 73 sim runs · 3 campaign sessions | 2026-06-25 |
| csi-ble-fusion-drift | CSI × BLE fusion under motion — does periodic BLE recalibration bound cross-environment drift?If a drifting Wi-Fi count is rescaled now and then by a Bluetooth reading, does the drift stay bounded when the model moves to a new floor? | running | simulation | 6 campaign sessions | 2026-06-25 |
| ble-csi-coregistration | BLE/CSI co-registration — does the RSSI anchor resolve occupancy where temporal CSI collapses?When a seated crowd stops stirring the Wi-Fi signal, can a Bluetooth reading solved from the same scene still tell how many people are there? | running | simulation | 2 campaign sessions | 2026-06-24 |
| csi-impairment-sim-to-real | CSI hardware impairments — does the occupancy signal survive a commodity NIC?Does the occupancy signal survive the flaws of an ordinary Wi-Fi card? | running | simulation | 2 campaign sessions | 2026-06-24 |
| csi-body-em-fidelity | Body-as-scatterer EM fidelity — what the dielectric cylinder hides (ray-traced)Is a person modelled as a water-filled cylinder with one loss number good enough for the radio simulation? | running | simulation | 117 sim runs · 1 campaign session · 1 paper cited | 2026-06-07 |
| csi-static-occlusion | CSI amplitude under static occupancy — the blockage↔variance crossover (ray-traced)When a room fills up, does the Wi-Fi signal mainly get dimmer or mainly get jumpier, and at what headcount does one take over? | running | simulation | 16 sim runs · 1 campaign session · 3 papers cited | 2026-06-06 |
| csi-through-wall-blockage | Locating the blockage-dominant regime — through-wall, sparse-LOS CSIThrough a wall, with little direct line of sight, does the shadow a body casts finally beat the jitter as the thing to count by? | running | simulation | 66 sim runs · 1 campaign session | 2026-06-04 |
| csi-material-fidelity | Wall-material fidelity — does the ITU wall palette shift second-order CSI statistics?Does it matter whether the simulated walls are drywall or concrete? | running | simulation | 8 sim runs · 1 campaign session | 2026-06-04 |
Simulation campaigns 57 briefs · the machinery the simulation experiments ran on
A campaign is one supervised sweep of simulations against a brief. Sessions are sealed and carry their own synthesis and their own adversarial criticism. The run archive holds every run they produced.
| Campaign | Experiment | Sessions | Runs | Latest |
|---|---|---|---|---|
c-ble-graded-count-poweredGraded BLE-anchor occupancy, at power — confirm or refute the hub-placement slope with a CI that can actually exclude the null |
ble-anchor-placement | 1 | 96 | 2026-07-06 |
c-csi-cross-geometry-scaleoutCross-geometry CSI drift at real floor count — 40 floors, nested seeds, honest LOFO baseline |
csi-cross-geometry-generalization | 14 | 94 | 2026-07-09 |
c-ble-trigger-equivalenceTrigger-vs-clock parity (TOST) — does event-driven recalibration reach equivalence with fixed cadence at ¼ the budget? |
csi-recalibration-trigger | 1 | 84 | 2026-07-07 |
c-csi-layout-driftCSI layout drift — a synthetic (T*, ε) prior for recalibration cadence |
csi-layout-drift | 3 | 74 | 2026-07-02 |
c-exps1-fundamental-diagramReproduce the Adrian-2024 / Liddle-2022 fundamental-diagram band on EXP-S1 Geometry A |
EXP-S1 | 3 | 73 | 2026-06-09 |
c-csi-cross-geometry-resplanCross-geometry generalization of CSI occupancy-sensitivity across the ResPlan dataset |
csi-cross-geometry-generalization | 1 | 72 | 2026-06-04 |
c-resplan-crowd-geometry-galleryCross-geometry crowd flow across the ResPlan apartment corpus — a replay gallery |
crowd-dynamics-replay | 3 | 68 | 2026-06-25 |
c-continuity-prior-csi-density-v2Continuity-prior PINN for CSI density — the refutation, this time with rendered evidence |
csi-crowd-temporal | 2 | 48 | 2026-07-15 |
c-csi-through-wall-blockageThrough-wall, sparse-LOS CSI — locating the blockage-dominant regime |
csi-through-wall-blockage | 1 | 48 | 2026-06-04 |
c-feature-type-transferFeature type, not modality — do kinematic features survive the furniture rearrangement that breaks fingerprints? |
csi-cross-geometry-generalization | 2 | 44 | 2026-07-04 |
c-ble-csi-coregistrationBLE/CSI co-registration — does the RSSI anchor channel resolve occupancy where temporal CSI collapses? |
ble-csi-coregistration | 2 | 44 | 2026-06-24 |
c-csi-crowd-temporalTime-resolved CSI under real crowd motion (coupled JuPedSim → Sionna) |
csi-crowd-temporal | 2 | 44 | 2026-06-24 |
c-csi-ble-fusionCSI x BLE fusion under motion — does periodic BLE recalibration bound the cross-environment drift of temporal-CSI counting? |
csi-ble-fusion-drift | 6 | 42 | 2026-06-25 |
c-coverage-estimator-factorialCoverage factorial with a bounded-influence estimator — isolate what actually drives blind-spot footfall |
ap-placement-coverage-counting | 2 | 40 | 2026-07-07 |
c-csi-band-calibrationCSI band calibration — does the coupled sim reproduce the WiMANS-measured 2.4-vs-5 GHz occupancy-discriminability gap? |
csi-band-calibration | 2 | 40 | 2026-06-25 |
c-csi-body-em-fidelityBody-as-scatterer EM fidelity — loss×radius confounding and per-band Fresnel shift (ray-traced) |
csi-body-em-fidelity | 1 | 36 | 2026-06-07 |
c-csi-roomscale-calibrationCSI room-scale calibration — does a WiMANS-scale room let the coupled sim reproduce the measured 2.4-vs-5 GHz occupancy gap? |
csi-band-calibration | 1 | 30 | 2026-06-25 |
c-csi-topology-sniffer-countDoes room-adjacency topology — not floor area — drive CSI/BLE fingerprint discriminability and the required sniffer count? (ResPlan stratified) |
csi-topology-vs-area | 3 | 24 | 2026-06-25 |
c-ble-drift-triggerBLE drift trigger — can an online drift statistic detect CSI layout drift and beat a fixed recalibration schedule? |
csi-recalibration-trigger | 1 | 21 | 2026-07-03 |
c-mall-archcadThe shopping mall as a placement oracle — where do APs both cover the concourse and count the crowd, on a real-ArchCAD-grounded mall floor? |
names csi-mall-grid, not a card |
4 | 20 | 2026-07-20 |
c-flat-day-reactive-vs-scriptedc-flat-day-reactive-vs-scripted |
not linked | 4 | 20 | 2026-06-26 |
c-crowd-motion-mixMulti-group crowd choreography on a real multi-room floor — arrival waves, seat reuse, through-traffic, with full movement replays |
crowd-dynamics-replay | 2 | 19 | 2026-06-04 |
c-resplan-egressEvacuation egress on ResPlan floors — emptying the rooms through the doorways |
crowd-dynamics-replay | 1 | 18 | 2026-07-06 |
c-ble-anchor-placementBLE anchor placement — does topology-informed (high-betweenness hub + large room) placement recover the graded count the dispersed anchors lost? |
ble-anchor-placement | 1 | 18 | 2026-06-25 |
c-csi-impairment-fitexp-csi-calibration — fit the CSI impairment profile from a month of empty room |
csi-impairment-fit | 1 | 16 | 2026-08-28 |
c-csi-crowd-occlusionCSI amplitude under crowd occlusion — locate the blockage↔variance crossover (ray-traced) |
csi-static-occlusion | 1 | 16 | 2026-06-06 |
c-ble-impairment-robustnessBLE-anchor impairment robustness --- does the RSSI calibration anchor survive a commodity NIC where the CSI phase feature does not? |
csi-impairment-sim-to-real | 1 | 14 | 2026-06-24 |
c-csi-impairment-sim-to-realCSI hardware impairments — does the occupancy signal survive a commodity NIC? |
csi-impairment-sim-to-real | 1 | 13 | 2026-06-09 |
c-csi-drift-real-floorCSI layout drift on the surveyed floor — the prediction the geometry already forces |
csi-drift-real-floor | 1 | 12 | 2026-08-28 |
c-csi-temporal-eventsEvent-rich temporal CSI — is crowd-induced CSI dynamics linear between events and forced at entries/exits (HAVOK), on a corpus built to answer it? |
csi-crowd-temporal | 1 | 11 | 2026-08-07 |
c-continuity-prior-csi-densityContinuity-prior PINN for CSI crowd density — does a 1D mass-conservation prior regularize the CSI->density inverse? |
csi-crowd-temporal | 2 | 10 | 2026-06-25 |
c-exp-s1-mechanism-defenseDefend the L4½.5 mechanism claim across the EXP-S1 sandbox |
EXP-S1 | 1 | 9 | 2026-06-09 |
c-csi-fidelity-materialWall-material fidelity — does the ITU wall palette shift second-order CSI statistics (drywall vs concrete sensitivity probe)? |
csi-material-fidelity | 1 | 8 | 2026-06-04 |
c-placement-oracleThe floorplan as a placement oracle — where do APs both cover the space and count the crowd, across 56 floors? |
ap-placement-coverage-counting | 2 | 7 | 2026-07-13 |
c-coverage-meets-crowdsCoverage meets crowds — does the placed-anchor sensing footprint cover where the crowd actually goes? |
ap-placement-coverage-counting | 1 | 7 | 2026-06-25 |
c-public-spacesReal crowds in real buildings — flow, dwell, device-traffic & CSI counting on S3DIS |
not linked | 1 | 6 | 2026-07-20 |
c-csi-ble-condition-switchCondition-switch fusion — the xmodal drift statistic replaces the staleness clock |
csi-recalibration-trigger | 2 | 6 | 2026-07-05 |
c-walk-notebook-flexFlexibility probe — papermill-executed Jupyter notebook as a first-class simulator artefact |
crowd-dynamics-replay | 1 | 6 | 2026-06-06 |
c-flat-day-csic-flat-day-csi |
not linked | 1 | 5 | 2026-06-25 |
c-fiit-find-a-seatSeek-and-occupy — pedestrians find a free seat on a real DB floor (IP-092) |
crowd-dynamics-replay | 2 | 3 | 2026-06-09 |
c-operanet-sim2realSim-to-real — does simulated CSI crowd-counting match real OPERAnet measurements? |
not linked | 1 | — | 2026-07-21 |
c-sim2real-count-transferSim-to-real count transfer: does synthetic-impaired CSI pretraining beat real-only at a small real budget? |
csi-sim2real-transfer | 2 | — | 2026-07-12 |
c-csi-topology-los-balancedTopology vs area for CSI occupancy — a fair fight with LOS/NLOS-balanced placement |
csi-topology-vs-area | 3 | — | 2026-07-07 |
c-cac-driftCalibration-as-control — does a BLE anchor rescue a drifting CSI occupancy filter (sim)? |
not linked | — | — | — |
c-csi-archcad-sensabilityCSI sensability across public-building geometry at scale — the ArchCAD ensemble the residential set couldn't cover |
csi-archcad-public-building-sensability | — | — | — |
c-csi-band-paired-hardwarePaired dual-band CSI on the AX210/Pi5 fleet — does the 5 GHz occupancy edge survive a self-illuminated, matched-rate, swapped-receiver measurement? |
csi-band-paired-hardware | — | — | — |
c-csi-ble-fusion-poweredCSI x BLE fusion, powered and cross-checked — the staleness-switch claim on >1 geometry with a real-data anchor |
csi-ble-fusion-drift | — | — | — |
c-csi-cross-geometry-scaleout-v2Cross-geometry CSI drift at real floor count — re-run with a session that exists and figures that render |
csi-cross-geometry-generalization | — | — | — |
c-fiit-library-dayA library day — campus-rhythm arrivals, friends sharing tables, seat reuse over 8 h (IP-102) |
crowd-dynamics-replay | — | — | — |
c-flat-day-aic-flat-day-ai |
not linked | — | — | — |
c-furniture-csi-observatoryc-furniture-csi-observatory |
names csi-oracle-grid-resplan-2605, not a card |
— | — | — |
c-furniture-walkabilityDoes furniture change where the crowd goes — the movement-side of the furnished floor? |
crowd-dynamics-replay | — | — | — |
c-ip110-showcasec-ip110-showcase |
not linked | — | — | — |
c-multifloor-fsm-observatoryc-multifloor-fsm-observatory |
not linked | — | — | — |
c-operanet-sim2real-fingerprintOPERAnet sim-to-real fingerprint — does ray-traced synthetic CSI reproduce the measured environment confound? |
csi-sim2real-transfer | — | — | — |
c-placement-oracle-poweredPlacement oracle, re-earned — the 57-floor coverage-vs-counting claim with attached runs and rendered figures |
ap-placement-coverage-counting | — | — | — |
c-resplan-crowd-to-csiCoupled crowd → CSI on a ResPlan floor — synthetic CSI–density pairs from watchable motion |
csi-crowd-temporal | — | — | — |