Mapped to a shared offline runtime family and usable with a recording or compatible ESP32 CSI stream.
Intended capability
Pilot this intended outcome through the shared Presence family, then validate it against site-specific ground truth: Holds the presence of a motionless human inside an autonomous operating zone — the maintenance tech kneeling in an aisle, the picker who stepped over the chain — via aimed mmWave cones on the zones' human-exclusion volumes, and records presence-during-machine-motion intervals as evidence events. This is the nightmare case by the fatality record: US robot-fatality studies covering 1992-2017 put most robot-related deaths near a stationary machine's envelope, with the victim's presence in no machine log — and CSI fades on exactly that person. The row is an evidence layer, never a safety function: the certified stack — fences, E-stops, safety PLCs — remains authoritative, and every event this row emits says so.
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.
No rendered revision is available yet.
Bundled recording
A related Fleet scene
This is one recorded vertical scenario. It is not separate validation of every capability in the catalog.
01 The physics
The LD2410C cone (bench-proven 2026-08-10) holds a stationary human to ~6 m, ±60°, via breathing micro-motion at 24 GHz — mm-scale chest displacement resolvable at a 12.5 mm wavelength — with ms latency and no dependence on the 2.4 GHz field. The sensor reports one composite target, no azimuth, no count; a metal chassis transiting the cone dominates the return and collapses 'still human plus moving robot' into 'moving target', so robot-transit windows are declared no-assessment intervals on CSI evidence alone — the estate feed may only add no-assessment windows, never attest robot absence or open assessment. CSI supplies the surrounding motion context; the cone supplies the hold CSI cannot.
02 Shared processing path
LD2410 static-target energy + per-gate range readout on the aimed cone (the hold) -> robot-transit masking gate (CSI machine-motion evidence in the cone's zone marks no-assessment windows; estate data may add windows, never open assessment) -> presence-interval assembly with dwell hysteresis -> join against CSI machine-motion verdicts in the same zone -> presence-during-motion event with range-gate evidence, filed to the ledger and certificate stream.
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
Bench and site protocol extending the 2026-08-10 LD2410C validation: stationary and breath-slowing subjects at 1-6 m across cone angles; robot-transit masking characterization with an AMR crossing at set ranges, including the human-tuned firmware's response to robot targets; false-hold rate against empty cells and against parked powered machinery — spinning lidar, fans, servo hold dither — over a multi-day soak.
04 Commercial hypothesis
'Witness Hold' — for the EHS director who has read the fatality statistics: an independent record that a human was inside the envelope while the machine moved, made by sensors that owe nothing to the machine.