Industry Model · AEO Stage 1

Industry Model Spec v1.1

Spec v1.1 (normative)
Document type: Specification · Status: v1.1 · Role inside architecture: World Model

industrymodel.html is not a marketing showcase. It is a normative reference point — a Proof Object that demonstrates: how a real industrial system thinks, decides, and blocks. Inside the broader virtauto architecture, this page acts as the World Model: the explicit representation of factory reality that later feeds the Decision Layer, Decision Contracts, and Authority Graph.

Stage 1 definition (bounded decision space)

AEO Stage 1 means: one domain, one decision class, one authority path, and an explicit, explainable BLOCK case. This page is decision-first and audit-first — PR-driven, not UI-driven.

Domain Bodyshop — Doorline Front Left (TVL)
Core decision class (Decision Product A) May this door be released / handed over?
Outcome space ALLOW · HOLD · BLOCK
Normative reference Infrastructure-first Decision-first Audit-first Block-capable PR-driven Holarchic scaling

Principle: Infrastructure → Decision Products (not “Use Case first”)

The model is structured as an infrastructure stack (Data Access → Knowledge Layer → Deterministic Verifiers → Orchestration), and only then exposes decision products on top. Energy optimization is a bounded advisory decision product. In v1.1, enterprise scaling adds a Resource Orchestrator to prevent cross-line conflicts.

Data Access Knowledge Layer Deterministic Verifiers Agent Orchestration CWD pattern Resource Orchestrator

Jump to: Industrial Decision Architecture · Proof Object · Infrastructure Stack · Decision Products · Production Process · Energy Optimization · CWD Orchestration · Resource Orchestrator · Governance · Audit · BLOCK

0.5) Industrial Decision Architecture

The Industry Model is the World Model inside virtauto’s broader industrial decision architecture. It sits between Simulated Reality and the later layers that evaluate decisions, formalize them, and govern who may approve or execute them.

Five-layer proof-of-concept path
1 Simulated Reality 2 World Model 3 Decision Layer 4 Decision Contracts 5 Authority Graph
Reality → structure Structure → decision Decision → contract Contract → authority

1) Simulated Reality

Synthetic operational context for a BIW Doorline. This layer creates safe, inspectable factory states without connecting to live production systems.

Open simulated context →

2) World Model

This page. It structures the factory into entities, stages, signals, evidence, and constraints so later decision logic can reason over an explicit industrial reality.

This page

3) Decision Layer

Evaluates operational options under constraints and system intent. Example: energy optimization in advisory mode.

Explore decision layer →

4) Decision Contracts

Formalizes industrial decisions as governed objects with intent, context, constraints, impact, and execution mode.

View decision contracts →

5) Authority Graph

Defines who may approve, veto, escalate, or execute industrial actions. No autonomy without explicit delegation.

Inspect authority graph →

0) Why this page exists (Proof Object)

This is the first World-Model Spec you can point to. Its purpose is to prove one thing under real constraints: autonomy can fail correctly.

What this proves

  • An industrial decision can be bounded (domain + decision class)
  • Decisions have explicit authority and enforced constraints
  • Every decision is auditable (append-only trace, evidence refs, ruleset)
  • BLOCK is legitimate: explainable and stable, not “AI error”
  • Scaling requires explicit handling of shared resources (v1.1 Resource Orchestrator)

What this prevents

  • “Smart agent did something” without accountability
  • Silent execution (state changes without trace)
  • UI storytelling without governed artifacts
  • Local optimization that creates enterprise-level energy conflicts
No decision without trace No bypassing Guardian PR-only governed change
Operational reality

We run our own website like we want industrial agents to run: PR-only, no direct writes, no silent actions, governed artifacts, required checks. The website is not a demo playground — it is our controlled test field for governance.

1) Infrastructure Stack (normative)

The Industry Model follows an infrastructure-first approach. Industrial decisions are only allowed if the underlying information layers provide reliable, verifiable inputs.

Four-layer infrastructure
1 Data Access 2 Knowledge Layer 3 Deterministic Verifiers 4 Orchestration
Sensors first Structured state Deterministic checks Fail-safe default

2) Knowledge Layer

The knowledge layer defines the semantic structure of the factory environment. It describes entities, relationships and operational context.

Core entities

  • Door (BIW component)
  • Production station
  • Robot / tool
  • Material flow
  • Energy consumer
  • Sensor

Relationships

  • Door → Station
  • Station → Robot
  • Robot → Tool
  • Station → Energy consumption
  • Sensor → State update

3) Deterministic Verifiers

Before any decision is executed, verifiers evaluate whether the system state satisfies defined constraints.

Geometry check

Validate dimensional tolerances using structured measurement data.

Presence verification

Confirm that required components are physically present before joining operations.

Surface inspection

Detect dents, scratches or imprints prior to final release of the component.

4) Orchestration Layer

The orchestration layer coordinates the flow of information and actions between production stations.

  • Scheduling production steps
  • Ensuring process order
  • Handling exceptions
  • Forwarding verified state to decision logic

Production Process — Doorline (Front Left)

The production flow below describes a simplified BIW door manufacturing process. Each stage updates the world model with verifiable state.

Stage 1 — Inner Panel Setup

1 assembly
  • Inner panel positioned in fixture
  • Clamps secure geometry
  • Variant verified via barcode

Stage 2 — Reinforcements

2 structure
  • Hinge reinforcement added
  • Lock reinforcement placed
  • Crash beam inserted

Stage 3 — Outer Panel Joining

3 hemming
  • Outer panel aligned
  • Adhesive applied
  • Hemming process executed

Stage 4 — Structural Completion

4 joining
  • Remaining welds completed
  • Clinch and rivet points verified
  • Structural integrity confirmed

Stage 5 — Quality Gate

5 inspection
  • Geometry measurement
  • Gap / flush check
  • Surface inspection

Stage 6 — Release

6 handover
  • Door classified (OK / REWORK / HOLD)
  • Audit record created
  • Part released to downstream process

Sensors and Signals

Sensors provide the raw signals that update the factory world model.

Identification

  • Barcode scanners
  • QR readers
  • RFID tags

Position and Geometry

  • Structured light scanners
  • Laser line scanners
  • Optical measurement systems

Joining Verification

  • Weld monitoring
  • Force sensors
  • Torque sensors

Energy Optimization (Advisory)

In addition to product quality decisions, the system can provide advisory recommendations regarding energy consumption.

Advisory mode

Energy optimization recommendations do not directly execute operational actions. Instead they provide suggestions that can be accepted, rejected, or escalated.

recommendation review blocked if unsafe

Audit Trace

Every classification decision must leave a verifiable trace.


{
  "door_id": "TVL-2026-001284",
  "timestamp": "2026-01-09T10:58:12Z",

  "stage": 5,

  "checks": {

    "geometry": "PASS",

    "gap_flush": "PASS",

    "surface": "PASS"

  },

  "decision": "ALLOW",

  "governance": {

    "guardian_check": "PASS",

    "policy": "release_requires_complete_evidence"

  }

}

Explicit BLOCK Case

Autonomous systems must be able to fail safely. If constraints are violated, the system must produce a clear BLOCK outcome.

Example

If surface inspection detects a dent exceeding tolerance, the door cannot be released.

geometry fail surface fail decision BLOCK