Simulation Scenario
Describe a reusable, versioned situation to simulate, including intended use, setup, behavior constraints, observations and qualified expected outcomes.
Bundle → Layer → Finding → Questions Filled
3 bundles · 3 layers · 6 findings · 12 questions
Scenario definition What situation is set up and why.
Purpose and setting
The objective, environment and initial conditions.
Objective
The question the scenario is meant to answer or the skill it is meant to train.
- What question or training goal does the scenario serve?
- Which outcomes count as success or failure?
Environment and initial state
The environment, map, weather and starting state of all elements.
- Which environment and initial conditions does the scenario set?
- Which parameters are fixed and which are varied?
Actors and events Who takes part and what happens.
Actors and storyline
Actors with their behaviour and the scripted or random events.
Actors
The actors, agents or vehicles with their roles and behaviour models.
- Which actors take part, and how is each one's behaviour defined?
- Are any actors controlled by humans in the loop?
Events and triggers
Scripted events and the conditions that trigger them.
- Which events occur, and what triggers each?
- Which events are stochastic, and with which seeds or distributions?
Validity and reuse How far the scenario can be trusted and reused.
Provenance and coverage
Sources, version and the range of conditions covered.
Scenario provenance
Where the scenario comes from, such as real data, expert design or generation.
- Was the scenario derived from real recorded events, designed by experts or generated?
- Which version is it, and who approved it?
Coverage
Which part of the operating domain the scenario represents.
- Which conditions of the operating domain does the scenario cover?
- Which important conditions are not covered by any scenario in the set?
Classifiers Filled
- Family
- World Models
- Category
- Information and virtual systems
- Entry kind
- entity
- Navigation path
- NAV.INF.VRT.SCN
- Domain
- INF.VRT.SCN
- Industry
- Cross-industry
- Tags
- simulationscenarioinf.vrt.scn
What it is Filled
A simulation scenario is the specification of one situation to be simulated: its inputs, actors, environment, initial conditions, events, objectives and expected outcomes. It is reusable across runs and simulators; the simulation model that computes behaviour and the results of a particular run are separate records.
In scope
- Scenario identity, revision and intended-use envelope
- Actor roles or explicit actorless applicability, external environment and model bindings, setup values and variation constraints
- Event and stop requirements, planned observations and optional outcome criteria
- Derivation, review, release and reuse restrictions of the scenario definition
Out of scope
- Simulation engine, solver implementation, execution scheduling and physical actuation
- Authoritative actor, environment, map, dataset and computational model lifecycles
- Run telemetry, actual observed outcomes, result evaluation and certification decisions
- Whole scenario-library coverage management and universal domain coverage claims
Why it exists Filled
Describe a reusable, versioned situation to simulate, including intended use, setup, behavior constraints, observations and qualified expected outcomes.
Distinguishing features Filled
- It specifies a situation to simulate, while the simulation model computes behaviour and a run produces results.
- It can be reused across simulators and runs when written in a neutral format.
- Expected outcomes are part of the scenario, which makes it usable as a test case.
- A game world is meant for open-ended participation; a scenario has defined objectives and bounds.
What robots and AI may and may not do Filled
Must not
- Present simulation results as real-world evidence without stating validity limits.
- Change a scenario after a run without versioning it.
- Hide failed scenarios from a test or certification report.
- Use personal data from recorded events in scenarios without a lawful basis.
- Design scenarios that rehearse attacks on real people or infrastructure.
Only with a human decision
- Accepting a scenario set as sufficient for safety or certification claims.
- Approving scenarios used to train people for safety-critical tasks.
- Releasing scenarios derived from real incidents.
May
- Author scenarios with stated objectives, actors, events and expected outcomes.
- Run scenarios in a simulator and compare results with expected outcomes.
- Vary parameters systematically and record seeds for reproducibility.
- Report gaps in scenario coverage.
Moral aspects Filled
- Safety claims for autonomous systems rest on how well scenarios cover real conditions.
- Scenarios derived from real incidents can expose the people involved.
- Biased scenario sets can hide failures that affect specific groups or places.
Who is affected
- People affected by systems validated in simulation
- Trainees and operators
- Regulators and certifiers
- People whose recorded data underlies scenarios
Owners Filled
Steward
The engineering, testing or training team that authors and maintains the scenario set.
Roles
- Scenario steward
- Owns namespace, definition scope and revision decisions.
- Scenario author
- Records setup, constraints, assumptions and provenance.
- Domain reviewer
- Assesses fitness for the declared use and unresolved limitations.
- Execution custodian
- Owns external executor authorization and run records; accepts only supported profiles.
- Data custodian
- Controls asset access, rights, retention and disposition.
Master systems
- Scenario libraries and databases
- Simulation and test management tools
Links to other meta-models Filled
references
- WM-VRT-003 - Candidate registry-parent binding to an external simulation definition or context. Pin target revision and instance semantics before use; no inheritance or runtime ownership is inferred.
aligned
- https://www.asam.net/standards/detail/openscenario-xml/ - Optional driving-scenario profile only. Pin XML version; no universal scenario equivalence or lossless conversion claim.
- https://sed-ml.org/ - Optional experiment-description concepts; exact schema mapping and version conformance remain deferred.
- https://fmi-standard.org/docs/3.0.2/ - Optional external model-interface requirements; engine, solver and execution lifecycle remain external.
- https://www.w3.org/TR/prov-o/ - Optional provenance vocabulary mapping for revision and evidence links; attribution is not permission.
neighbor
- WM-VRT-003 Simulation - Registry parent denotes domain context. Use a candidate reference to the executing simulation specification; do not infer scenario is-a simulation or copy its runtime lifecycle.
- Simulation run and result records - One definition can be used by zero or many runs. Expected outputs are prescriptions or hypotheses; actual outcomes and evaluator verdicts remain external observations.
- Actor and environment masters - Scenario-specific role, initial-state overrides and dependency versions are local; actor identity, environment geometry and behavior-model implementations remain externally mastered.
- Test specification and scenario library - A scenario may reference acceptance criteria or coverage coordinates. It is not necessarily a test, and no single scenario establishes coverage of the whole library or fitness of a system.
parent
- WM-VRT-003
What else AI and robots need to interact with it Filled
Identity and identifiers required Filled
- A scenario is identified by its library identifier and version.
- Scenario files are referenced by path or content hash in the scenario library.
- Runs reference the scenario identifier, version and random seed.
Direct properties not applicable Not applicable
Not applicable
The subject is an institutional or informational record, not a physical object with measurable properties.
Recognition optional Filled
- A scenario names an environment, actors, events, initial conditions and expected outcomes.
- Often confused with the simulation model, a simulation run or its results.
Capabilities and actions required Filled
- Scenarios can be parameterised, versioned and run repeatedly.
- Scenario sets can be analysed for coverage of an operating domain.
- Results can be compared with expected outcomes automatically.
Hazards and failure modes required Filled
- False confidence from scenarios that miss real-world conditions.
- Irreproducible results when versions or seeds are not recorded.
- Privacy breaches from scenarios built on recorded personal data.
Standards and interfaces required Filled
- ASAM OpenSCENARIO for driving scenarios.
- ASAM OpenDRIVE for road networks used in scenarios.
- IEEE 1516 High Level Architecture for distributed simulation.
Context of use required Filled
- Used in testing autonomous systems, training, emergency planning and research.
- Scenario-based testing supports safety cases in automotive, aviation and robotics.
Sources Filled
- ASAM OpenSCENARIO XML overview - ASAM
- ASAM OpenSCENARIO XML - 8.4 Storyboard element execution - ASAM
- SED-ML introduction and conceptual overview - SED-ML community
- Standard for Models and Simulations - National Aeronautics and Space Administration
- Functional Mock-up Interface Specification - Modelica Association Project FMI
- PROV-O: The PROV Ontology - World Wide Web Consortium
- Date and Time on the Internet: Timestamps - Internet Engineering Task Force
- ASAM OpenSCENARIO, Association for Standardization of Automation and Measuring Systems
- IEEE 1516 Standard for Modeling and Simulation High Level Architecture, IEEE
Open questions
- Verify source release packages, claim-level support, current applicability and licensing, distinguishing historical execution semantics from current overview metadata.
- Develop domain profiles and executable schemas with actorless, exploratory, conflicting-trigger, interrupted-run, stochastic and lossy-export fixtures.
- Pin external simulation and result bindings and test compatibility and repeatability claims with qualified executors.
- Restore independent external review before any canonical or publishable-draft promotion.
- Nested field schemas, requiredness by representation level and executable constraints are not implemented.
- Independent external provider review is absent; both providers were explicitly skipped.
- Direct HTTP checks are unattempted under the owner-reported sandbox restriction; no response statuses were measured.
- Full source-package and latest-version verification, licensing review and domain applicability remain incomplete.
- Cross-tool equivalence, stochastic replay fixtures, hardware interaction and human-participant profiles require further work.
Machine files
Provenance
world-models research · reviewable-draft
Built from: models/wm-vrt-012-simulation-scenario/spec.yaml, ver-cy/world-models/card-supplements/wm-vrt-012-simulation-scenario.json