Case Study: Defence Data Interoperability
IES to HQDM: an open 4D ontology crosswalk for defence data
The first public crosswalk between the UK Information Exchange Standard (IES), the 4D ontology behind UK defence and security data, and HQDM, the model underpinning the National Digital Twin. Open, validated, and released so that suppliers building across the two can start from something concrete.
Focus
Interoperability
and autonomy assurance
Standards bridged
2
IES and HQDM, both 4D
Verification
17 / 0
correspondences resolve; SHACL conforms
The Challenge
The UK Defence Investment Plan (June 2026) commits over £5bn to autonomous systems and £7.5bn to a Digital Backbone and Digital Targeting Web. All of it depends on heterogeneous systems, and coalition partners, sharing data a machine can reason over. On the UK side that shared vocabulary is the Information Exchange Standard (IES), an open 4D ontology in the BORO tradition, stewarded through the cross-government IES Working Group. Alongside it, the built and physical environment runs on HQDM and the National Digital Twin Foundation Data Model.
An autonomous system reasoning about a mission in a real place has to connect what IES says about the operational picture to what HQDM says about the terrain and infrastructure. Both are 4D and share a common heritage, yet no machine-readable crosswalk between them had ever been published. For a smaller supplier, the cost of working out that alignment by hand is the real gate between a promising prototype and a trusted, fielded capability.
What we built
| Deliverable | What it is |
|---|---|
| Crosswalk (SSSOM) | 17 backbone correspondences with predicate, confidence and justification. |
| Crosswalk (RDF) | SKOS mapping triples with PROV-O provenance and reified correspondences. |
| Divergences record | The curated pairs that look like they map and do not: the original scholarship. |
| SHACL shapes | Validate every correspondence has a subject, object, SKOS predicate, confidence and provenance. |
| SAPIENT safety case | A worked example grounding one autonomous sensor node in IES-typed world states. |
| Reference pipeline | Candidate generation and validation you can run against the live ontologies. |
A finding, not just a mapping
A clean crosswalk is not one with no disagreements; it is one where the disagreements are named. The headline example: a name matcher aligns ies:Event to hqdm:event and gets it exactly backwards. In IES an Event is a happening with participants, so its true HQDM counterpart is hqdm:activity; hqdm:event is an instantaneous boundary point with no participants. A tool that maps them on the shared label corrupts the shared picture. The repository records this and five further divergences with evidence from both ontologies, which is where an implementer would otherwise silently get it wrong. All six are set out below.
ies:Event vs hqdm:eventA name matcher aligns these first and gets it backwards. ies:Event is a durative happening with participants, so its true counterpart is hqdm:activity; hqdm:event is a zero-duration boundary point. Recorded as relatedMatch at confidence 0.25 purely to carry the warning.
ies:BoundingState vs hqdm:eventBoth mark where a 4D extent begins and ends, so they are functionally equivalent, but in IES a boundary is a state and in HQDM it is a point event. A correspondence exists, but as a relatedMatch needing an EDOAL-style transformation, not a class equivalence.
ies:State vs hqdm:stateThe semantics are close (mapped at 0.85), but ies:State is a top-level root class while hqdm:state is subsumed under spatio_temporal_extent. Any reasoning that relies on state being a spatio-temporal extent holds in HQDM but not from the IES class graph alone: a soundness trap over a naive union.
ies:EventParticipant vs hqdm:participantNot every notable pair is a trap. Both models independently make participation a state, not merely a relation, a deep agreement inherited from the shared BORO commitment. This is why the participation correspondences are among the strongest in the set, and where cross-model reasoning is safe.
ies:PossibleWorld vs hqdm:possible_worldThe classes correspond (0.80), but the surrounding apparatus differs: how membership in a world is asserted, and whether worlds are themselves classified by a powertype. Mappings of the relations around possible worlds need case-by-case treatment.
The powertype stackThe class-of hierarchies line up at the root (ies:ClassOfElement vs hqdm:class_of_spatio_temporal_extent), but IES’s domain tree is shaped by intelligence and security use cases and HQDM’s by enterprise and engineering. Above the backbone the hierarchies stop being parallel and become a genuine matching problem.
Three of these are traps a label matcher falls into, one is a convergence worth exploiting, and two mark where the hand-curated backbone ends and the automated alignment work begins. A clean crosswalk is not one with no divergences; it is one where the divergences are named.
Outcome
An open, machine-readable crosswalk released under CC-BY-4.0, built on Tesseract's open-ontologies alignment engine and validated against the live published ontologies: every one of the 17 correspondences resolves, and the SHACL shapes conform. It ships with a worked safety case showing how one SAPIENT (BSI Flex 335) autonomous sensor node can have its behaviour grounded in IES-typed world states, connecting our CIVeX agent-verification research to the defence data standard.
- ›Data and documentation: CC-BY-4.0, free for public and commercial reuse.
- ›Upstream ontologies referenced by IRI: IES (Open Government Licence), HQDM (Apache-2.0).
- ›Candidate-for-review, and open to correction: the most welcome contribution is a new divergence.
This is an independent, self-initiated demonstration. It is not affiliated with, or endorsed by, the Ministry of Defence, the IES Working Group or the National Digital Twin programme. It is built entirely on published open standards and open ontologies.
"The autonomy money assumes suppliers can already speak the standard. The alignment layer is what lets them. Publishing an open crosswalk, with the traps named honestly, turns that from bespoke consulting into shared infrastructure a smaller supplier can build on."
Fabio Rovai, Tesseract Academy
Applied example: a PYRAMID avionics bridge
The crosswalk is not only a reference: it does work. PYRAMID (UK Defence Standard 00-134), the MOD open architecture for avionics and mission systems, deliberately has no shared data model and pushes interoperability into per-deployment "bridges" that perform semantic translation. Using this crosswalk, we ground three PYRAMID components in IES and HQDM so that the same object, a building that is both a Geographical_Feature and a Tactical_Object, resolves to one referent. It is the crosswalk's founding motivation, connecting the operational picture to the terrain, reduced to one bridge.
View the open crosswalk
SSSOM and RDF correspondences, the divergences record, SHACL shapes, and the SAPIENT safety case.
