Simulation environments
Aplark Sim
Test intelligence in worlds you can control.
Aplark Sim is being designed as a simulation environment for generating test conditions that are difficult, dangerous or impossible to capture in the field — with the scene parameters recorded as first-class provenance rather than discarded at render time.
The problem
The cells you cannot fly
Coverage gaps cluster in exactly the conditions nobody wants to fly into: degraded visibility, sensor obscuration, adversarial geometry, failure modes that are expensive to stage and unethical to stage repeatedly. Those are also the cells that dominate the hazard exposure calculation.
Simulation is the obvious answer and the reason it is only partly an answer is provenance. A rendered frame whose scene parameters, renderer version and sampling seed were not recorded cannot be cited as evidence, because nobody can say afterwards what operational condition it actually represented.
How it works
How it is intended to work
Aplark Sim is being built so that scene specification is the unit of work: an ODD cell is named, and the simulator is asked for samples in that cell rather than for a scene that happens to fall there. Scene parameters, renderer build, asset digests and seeds would be emitted as a manifest alongside every frame.
The intended output is not imagery. It is imagery plus the declaration that makes the imagery admissible — or the honest statement that it is not.
Trust boundary
No privileged status inside Aplark Assure
A vendor that both generates the test data and adjudicates the sim-to-real gap in the assurance argument holds a conflict of interest that a competent certification authority will identify immediately. We have resolved it in the architecture rather than in the sales conversation.
Data produced by any Aplark generation tool enters Aplark Assure as a declared, provenance-tracked input carrying exactly the same evidentiary burden as a third-party source. It takes the same declared source type and the same default intra-cluster correlation — simulation runs sharing a scene default to ρ = 0.3, and overriding that requires evidence plus a ledger record. It carries the same mandatory sim-to-real-gap defeater in the assurance case. It meets the same provenance requirements on the generating environment as any external dataset.
No code path, configuration flag or licence tier relaxes any of these because the data came from us. Aplark Assure will never produce the thing it evaluates.
Capabilities
What it does
Cell-addressed generation
Intended to accept a target ODD cell and return samples in it, so that generated data is directed at a declared coverage gap rather than produced speculatively.
Provenance manifests
Planned to emit renderer build, asset digests, scene parameters and seeds with every output, in the format Aplark Assure ingests as a declared source.
Sensor modelling
Under design: EO/IR sensor response, motion blur point-spread functions, and noise characteristics parameterised so that a generated frame corresponds to a stated platform rate and integration time.
Who it is for
- Programmes whose coverage gaps sit in conditions that cannot be flown
- Teams already using Aplark Assure who want gap-directed generation