All capabilities
Robot Fleet Witness Physics limit

RollCall

Fleet RF Roll-Call: Per-Robot Identity from Signature

People wish the radio field could recognise each robot by its own reflection — this one is unit 7, that one is unit 12. It cannot.

Readiness Physics limit

Retained for transparency, but blocked from execution because the catalog physics review rejected it.

Evidence Frontier Unvalidated concept

An early research proposal with no capability-specific product validation yet.

Runtime family Not mapped

No browser runtime is assigned to this capability yet.

Why live sensing is unavailable

The mechanism inverts: the same physics that makes robots exceptional RF subjects makes them indistinguishable from one another. A large metal chassis's return is dominated by pose, aspect angle, and the multipath geometry of the moment — within-unit variance across passes exceeds between-unit variance across a homogeneous fleet, so the learned 'fingerprint' identifies the trajectory and the room, not the machine. Two independent kills. First, warehouse baselines are perishable: every pallet moved rewrites the static multipath any signature is conditioned on, so a fingerprint library decays in days and degrades silently into confident misclassification — the worst failure mode an evidence product can have, because it does not know it is wrong. Second, band rule R-A2 bars robot identity from RF alone outright: a witness plane that guesses identities poisons every certificate downstream, and one wrong roll-call entry forfeits the whole ledger's credibility the moment a dispute tests it. Drive-EMI 'fingerprints' die the same way — the ISM floor is the most confounded spectrum in the building and carries no direction and no attribution. The salvageable sliver already ships elsewhere in the band: positional-temporal correlation with the declared plan. In a sparse scene, one robot declared in one zone at one time is attributable by exclusion, stated as correlation with its uncertainty attached, never as recognition — and that sliver is load-bearing in the flow map and misplacement rows, which is where it belongs.

Intended capability

Document why this proposed outcome is blocked by the reviewed physics: Identifies each robot in the fleet by its individual RF return as it passes — a per-unit roll-call from the field alone, no fleet feed required: this return is unit 7, that one is unit 12. Killed before it could ship; retained here as the band's cautionary row.

This describes the intended outcome. Readiness is physics limit, evidence is class C, 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.

Physics limit

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 temptation is real: robots are exceptional RF subjects — large metal radar cross-section, distinctive drive noise, machine-regular Doppler — so surely each unit's return is a fingerprint. It is not. A metal body's return is a function of pose, aspect angle, and the multipath geometry of the moment far more than of unit identity, and a homogeneous fleet is a fleet of physically identical reflectors.

02 Shared processing path

Per-transit CSI signature extraction -> per-unit fingerprint library -> nearest-signature classification. Dead at stage two: the library does not exist as a stable object in a warehouse whose multipath is rewritten by every moved pallet.

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

n/a — killed before protocol design; see reject_reason.

04 Commercial hypothesis

Would have been 'Roll-Call' — per-unit tracking with zero fleet integration — which is exactly why it must be killed loudly rather than quietly.