Skip to content

Technology

Spatial Resolver

Tag Resolver finds the object. Spatial Resolver finds what must be reviewed around it.

It turns document clauses and engineering requirements into the objects and spatial neighborhood to review now — by interpreting Spatial Graph relationships through qualified spatial rules. It is not nearest-neighbor search.

Which objects and spatial relationships should be reviewed right now?

Pain Point

Finding the object still leaves the review neighborhood empty.

Resolving a tag to a 3D object does not finish the engineer’s job. Replacement, removal, access, clearance, and impact all require the right surrounding relationships before design review can start. A bag of nearby objects is not that answer.

What it does

It turns engineering intent into the neighborhood to review.

Spatial Resolver takes a document clause or engineering requirement and a seed object, then applies Spatial Graph relationships and qualified spatial rules to select Relevant Objects and a Spatial Neighborhood. It also keeps which rule produced the neighborhood and why.

  • Inputs: engineering intent, seed object, Spatial Graph, qualified spatial rules
  • Interpretation: relationship domains such as containment, support, access, clearance, zone, adjacency
  • Outputs: review neighborhood, relationship subgraph, rule attribution, uncertainty

Example flow

From document clause to Evidence Snapshot

Design review does not stop at tag matching. It continues from object to Spatial Graph to neighborhood to AI-supported review to an evidence package.

  1. 01

    Document Clause / Engineering Requirement

  2. 02

    Engineering Identity

  3. 03

    Tag Resolver

  4. 04

    Actual 3D Object

  5. 05

    Spatial Graph

  6. 06

    Spatial Resolver

  7. 07

    Relevant Objects / Spatial Neighborhood

  8. 08

    AI-based Design Review

  9. 09

    Evidence Snapshot

Not proximity search

Listing nearby objects is not choosing what to review.

Pulling everything inside a radius creates review noise. Spatial Resolver is not nearest-neighbor search or scene-graph traversal. It selects relationships that match engineering intent and review rules. Distance is only one primitive among many.

Proximity alone

Every object within 2 m of this valve

Spatial Resolver

The neighbors that matter for isolation, support, access, and clearance when replacing it

Proximity alone

Follow parent–child hierarchy only

Spatial Resolver

Select process, structural, access, and zone relationships by intent

Tag Resolver

Identity resolution comes first. Spatial interpretation comes next.

Tag Resolver is necessary. Spatial Resolver is the differentiator on top of it. When identity is ambiguous, candidates and confidence are passed forward — Spatial Resolver does not invent objects.

Tag Resolver

Which real 3D object does this tag point to?

Spatial Resolver

Which objects and spatial relationships should be reviewed right now?

Spatial Graph

Spatial Resolver reads the Spatial Graph.

The Spatial Graph carries object identity, hierarchy, position, engineering relationships, and spatial context. Sources are not equal in richness.

  1. 01 · Smart3D

    Primary path — richest object, hierarchy, relationship, metadata, and spatial context

  2. 02 · AVEVA RVM/ATT

    Alternative engineering-model path

  3. 03 · Navisworks

    Simplified properties/hierarchy fallback

AI design review

AI reviews the same neighborhood. Engineers keep authority.

The neighborhood and relationships Spatial Resolver selects become the scope for AI-based design review. AI identifies candidate issues and supports review; authorized engineers make final engineering decisions.

AI authority is support and evidence only. Spatial Resolver proposes context; engineers decide. L3D Agentic Robotics is outside this page’s scope.

Evidence Snapshot

The output of review is an Evidence Snapshot.

When spatial reasoning contributes to a review, the Evidence Snapshot preserves objects, revisions, and clauses together with rule attribution. It is not a screenshot — it is a portable evidence package humans and AI can trace.

Evidence package

Evidence Snapshot

decision basis
Seed object Authoritative or candidate target from Tag Resolver
Neighborhood Intent-qualified review neighborhood
Spatial rule Which rule produced this set, and why
Clause / req. Document clause or engineering requirement
Revision Model and drawing revision context
Uncertainty Candidates, gaps, model quality

Evidence Snapshot records the basis for judgment. It does not replace approval.