SAREF4AGRI & SAREF4AUTO Join Smart Data Models: 58 Candidate Models for Precision Agriculture and Connected Vehicles

Smart Data Models’ Candidates repository now includes two more ETSI SAREF extensions: saref4agri and saref4auto, together contributing 58 candidate models. Both ship
the full Candidates artifact set — a JSON Schema, an NGSI-LD @context, and all four example serializations — validated against its own schema.

SAREF4AGRI — 21 models

SAREF4AGRI (ETSI TS 103 410-6) extends SAREF into precision agriculture and livestock monitoring, stitching together W3C SSN/SOSA (systems, platforms, deployments), GeoSPARQL (farm and parcel topology), the TIME ontology (deployment periods), OBO TAXRANK (species classification), and FOAF (farmers and farm holdings as organizations). It’s grounded in two worked examples in the standard: a livestock farm with milking and eating-activity sensors, and a smart-irrigation deployment with soil tensiometers and pluviometers reporting through a platform/system/deployment chain.

https://github.com/smart-data-models/Candidates/tree/master/standards/saref4agri

SAREF4AUTO — 37 models

SAREF4AUTO (ETSI TS 103 410-7) is SAREF’s extension for connected and automated vehicles and the roadside infrastructure (C-ITS) that supports them — vehicles, trailers, dangerous-goods cargo, vehicle platoons, pedestrians/cyclists and their clusters, parking spots, roadside units/sensors/actuators/traffic lights, traffic management centres, and a wide catalogue of sensor types (lidar, radar, UWB, stereovision…).

https://github.com/smart-data-models/Candidates/tree/master/standards/saref4auto

Translation notes

• Large named-individual vocabularies, kept open rather than exploded. SAREF4AUTO’s 161-member TrafficEventCause vocabulary (drawn from ETSI’s ITS Common Data Dictionary) became one entity with an open causeCode and a closed category enum of the 20 top-level causes — not ~160 separate models. Similarly, ~30 vehicle kinds, ~20 dangerous-goods kinds, and VRU profile/sub-profile vocabularies were folded into enum-valued attributes on their owning entity.
• Generic value-wrapper classes excluded. Eight SAREF4AUTO OWL “typing” classes with no domain content of their own (BooleanProperty, DecimalProperty, ConfidenceValue…) aren’t modeled as entities — every property using them is instead a plain typed attribute on its owning entity. Net effect: ~46 raw primary classes in scope → 37 genuinely distinct models.
• Road topology deliberately excluded. SAREF4AUTO’s own Road/ RoadSegment/Lane classes were left out of this translation, since they would duplicate rather than extend the official dataModel.Transportation subject’s existing coverage.
• Two naming collisions found, kept verbatim. SAREF4AUTO’s own Vehicle (C-ITS/kinematics-oriented: heading, speed, automation level) and ParkingSpot (thinner, C-ITS-oriented) share a name with existing official Smart Data Models that cover complementary but non-identical ground (fleet/telematics Vehicle; richer occupancy-model ParkingSpot). SAREF4AGRI’s Animal likewise shares a name with the official dataModel.Agrifood‘s own Animal. Per the translation policy, names were kept exactly as the source ontologies define them rather than renamed to dodge the collision — flagged in each standard’s standard-metadata.yaml for the Candidates review process to resolve.

Try it, and tell us what’s missing

These are Candidates — first, carefully-sourced translations, not yet
promoted to official Smart Data Models subjects. The naming collisions
above are the main open item for both standards before promotion.

Browse the models, open an issue, or send a PR:
👉 https://github.com/smart-data-models/Candidates/tree/master/standards/saref4agri
👉 https://github.com/smart-data-models/Candidates/tree/master/standards/saref4auto