All capabilities
Camera Shield & Cage Watch Shared-family mapped

Null Map

CSI Null-Zone Map: Which Aisle Is Being Shadowed

Tells you which aisle is being blocked, by watching which signals suddenly go dead.

Contested Space · Contested capability

What this capability does not do.

Does not act.

Latent emits a reading. It does not switch cameras, unlock cages, cut power or open a ticket. Your VMS, PoE switch or access controller does that, on your side of the boundary. A Latent blueprint contains no executable code, no remote URLs and no device secrets — by schema, not by policy.

Does not identify.

No face, no plate, no MAC, no device identity, no payload. A movement signature is a shape moving through a room, not a person. Nothing stored here resolves to an individual.

Does not survive everything.

A re-racked row, a moved crash cart or a propped door permanently rewrites the fade map. The observable is amplitude nulls, never phase — ESP32 CSI phase is CFO/SFO-scrambled per frame with no inter-node reference.

Across Camera Shield Jam the mesh and the wake never fires. Two people moving together read as one disturbance. Re-rack a row and the fade map is wrong until you recalibrate.

Observable
Departure of per-link CSI statistics from a learned empty-space baseline, at the mesh frame rate.
Retention
The entry event, its zone and its severity. Neither baseline frames nor raw CSI are exported.
Node minimum
2 nodes
Export policy
Conditional — Exportable into a signed release, subject to the deployment’s own preflight.
Chain of custody

Node identity

Each node mints its own key over USB, in your hand. Physical possession is the root authority: a device that cannot prove physical presence is refused enrollment outright. A cloud pairing PIN is a weaker ownership claim and is never treated as liveness.

Reading provenance

Every reading is attributable to a device id, profile id, hardware id and firmware version recorded at install and verified by mutual HMAC-SHA256 proof over an LFW1 nonce exchange. A node whose proof did not verify is marked untrusted, not merely offline.

Install attestation

Completing an installation is a high-risk action. It requires your explicit browser approval, an approval snapshot that still matches, an idempotency key, and heartbeats less than three minutes old. There is no silent commissioning.

How enrollment works

Properties you can map to your own controls

  • No optical sensor is present in this capability’s node set.
  • Raw CSI does not leave the node. Only derived event rows are stored or exported.
  • No stored identifier resolves to a person. Where access events are correlated, that correlation happens in your systems, against your logs.
  • The runtime executes on the node and in your browser. This capability has no cloud inference path.

These are properties, not certifications. Latent does not assert compliance with any framework on your behalf.

A capability description, not an incident record.
Readiness Shared-family mapped

Mapped to a shared offline runtime family and usable with a recording or compatible ESP32 CSI stream.

Evidence Strong theory Supported hypothesis

Published physics or adjacent results support the hypothesis; capability-specific product and site validation are still required.

Runtime family Intrusion event

Entry or unexpected-presence event relative to a learned empty-space baseline.

When no ESP32 is connected

No compatible ESP32 stream is connected. Use the bundled Camera Shield recording; it demonstrates the shared Intrusion event runtime, not independent proof of this capability.

Intended capability

Pilot this intended outcome through the shared Intrusion event family, then validate it against site-specific ground truth: Watches which stationary links lose which subcarriers to deep amplitude fade and names the zone being shadowed, plus the zones the mesh currently cannot see at all. The observable is amplitude nulls, never phase: ESP32 CSI phase is scrambled frame to frame by carrier and sampling frequency offset with no inter-node reference, so phase is not used here. A re-racked row, a moved crash cart or a propped door permanently rewrites the fade map, and the entry is only trustworthy against a maintained baseline — see cage-calibration-drift-audit, which exists to police exactly that. Sibling declared: ghost-in-the-cage-through-wall tracks a mover inside one cage; this entry maps which zones are visible and names the one being blocked.

This describes the intended outcome. Readiness is shared-family mapped, evidence is class B, and a catalog mapping or recording is not proof of this outcome at a real site.

Solution blueprint

See the environment before installing it.

This exact kit is one of 191 first-class designs. It includes geometry, objects, nodes, wording, scenarios, installation, limitations, and catalog-bound readiness.

Shared-family mapped

No rendered revision is available yet.

Bundled recording

A related Camera Shield scene

This is one recorded vertical scenario. It is not separate validation of every capability in the catalog.

01 The physics

An HT20 Wi-Fi channel exposes 56 used subcarriers, each with its own amplitude, and a stationary link's per-subcarrier fade structure is a stable fingerprint of the room's geometry. A body parked in a Fresnel zone of that link drives specific subcarriers into deep fade in a reproducible pattern, so attributing which links lost which subcarriers constrains the disturbance to a link intersection — zone granularity, without a directional antenna and without a lens.

02 Shared processing path

Per-subcarrier amplitude series on stationary links -> per-link fade-structure baseline (amplitude only; phase discarded) -> adaptive PCA + MVE deviation scoring per subcarrier group -> link-intersection attribution across the mesh to a named zone -> coverage report naming zones with no currently trustworthy link.

This is a capability design path. Components may be shared with other catalog entries; it is not presented as a unique algorithm.

03 Validation plan

Through-wall and cage-intrusion detection are already evidenced at B on ESP32-wall and OPERAnet. Own-hardware: six nodes around a real cage, a person holding position at 12 surveyed spots for 60 s each versus an empty baseline, report zone-attribution accuracy and per-zone detection AUC; then deliberately re-rack one row and re-measure to quantify how far the fade map degrades before recalibration.

04 Commercial hypothesis

'Null Map' — camera-free zone attribution where a lens is not permitted, sold to the compliance-bound facility security director with the maintenance cost stated plainly: rearrange the room and the map must be retaught before it can be trusted.

A quiet reading is not an all-clear. If nodes are stale, degraded or jammed, this runtime says so explicitly. It never infers safety from missing data.