PNTMAP: GNSS interference alerts as canonical objects

A PNTMAP alert describes an interference episode: a type, a frequency band, a signal strength, a confidence, a severity and a validity interval, with an optional geolocated emitter and an optional affected area.

Two different things are being reported at once. There is something that exists, an interference source at a place over an interval, and something that happened, interference observed over an area. Those are different canonical kinds.

PNTMAP is published by Decent Cybersecurity, the same company that builds SynapseCommand. That relationship is stated here rather than left to be discovered, and it changes nothing about how the adapter behaves.

Satellite navigation denial is now an ordinary condition of operations rather than an exception, and it degrades exactly the inputs a common operating picture depends on most.

An interference report is what makes a position source meaningful. A satellite derived fix inside a jammed area is suspect; an inertially derived fix in the same area is not, and the canonical model records which is which precisely so that this comparison can be made.

Because the alert is translated into the same objects as everything else, the effect of interference on the picture is visible in the picture rather than in a separate advisory channel.

One alert becomes an Entity for the emitter and an Event for the interference. Forcing them into one object would lose whichever half the container was not shaped for.

Affiliation is not inferred. A jamming emitter is very likely hostile, and the adapter still records unknown unless the payload itself states an attribution, because an intelligence judgement made inside a translator is invisible to the audit trail and unattributable.

An alert with no geolocated emitter produces an entity with no position, never a placeholder coordinate.

An unreadable severity raises rather than defaulting, because an alert whose urgency cannot be read must not arrive labelled as routine. Interference type and attribution resolve to unknown instead, because both vocabularies have a member that means exactly that.

The stated geolocation method maps to a position source that reflects how much the fix can be trusted, and the source's own word is kept alongside it. Time of arrival and angle of arrival fixes are recorded as estimated, and none of them is ever labelled as satellite derived, which would be absurd in an alert about satellite navigation being denied.

Every field the adapter does not map is parked verbatim, so a field it has never seen still arrives intact. The one declared transform is the re-rendering of the alert times into the canonical fixed precision form, and the harness prints its reason on every run.

Identity derives from the emitter identifier when the payload has one and from the alert identifier when it does not, and which basis was used is recorded, because an identifier keyed on a per-alert field is not stable across alerts.