The SimReady Standardization Workflow#
The SimReady Standardization Workflow outlines the process of specifying asset capabilities within OpenUSD. It describes how to extend OpenUSD for new workflows, industries, and simulation runtime use cases. This playbook details steps for identifying domain experts, drafting asset specifications, performing data mapping and gap analysis, developing reference pipelines, and fostering industry adoption.
Goals#
Design Asset Specifications for Standardization
Identify and define new asset capabilities to extend OpenUSD and unlock new simulation workflows. Collaborate with NVIDIA and OpenUSD ecosystem partners so that they have agency in the creation process and can build assets that meet simulation needs at scale.
Create End-To-End Reference Pipelines
Produce documentation, validators, workflows, samples and software to enable successful adoption by the ecosystem partners.
Enable Industry Adoption
Create efforts to involve and encourage ISV / GSI / Partner adoption, including integration of capabilities into other DCCs and libraries.
Terminology#
Term |
Description |
---|---|
OpenUSD |
Open Universal Scene Description, an open framework for 3D data interchange. |
AOUSD |
Alliance for Open Universal Scene Description - an open, non-profit organization dedicated to fostering the standardization, development, evolution, and growth of OpenUSD. |
SimReady |
A set of 3D asset specifications for industrial digitalization and other simulation use cases. |
Domain |
An area of activity within the larger context of simulation (visualization, process planning, AV simulation, robotics etc.). |
Capabilities |
Features or use cases of a given domain which are enabled by the asset specification. |
Pipeline |
A series of processes to convert, transform, validate and load assets – from source systems to consumption in simulators or other runtimes. |
References#
Reference |
Description |
---|---|
OpenUSD Homepage |
|
AOUSD Homepage |
|
AOUSD Working Groups |
|
AOUSD Interest Groups |
|
Use Case Template Document |
|
Gap Analysis Template Document |
|
Asset Validator Documentation |
https://docs.omniverse.nvidia.com/kit/docs/asset-validator/latest/index.html |
Team / Roles#
Role |
Description |
---|---|
PIC |
Pilot In Command – Makes sure that the three goals are achieved and serves as the “chair” of the project. Works with external partners as well as the Domain and OpenUSD Experts to clarify the use case and set priorities. Responsible for project and product level documentation, positioning, alignment with related product managers and products, releases, announcement demonstrations and project planning. |
Engineering PIC |
Responsible for technical documentation and related artifacts, as well as high level planning and execution and releasing of development tasks. |
Content PIC |
Responsible for Sample Content, including publishing via partners and datasets such as NVIDIA SimReady libraries and/or physical AI Datasets on Hugging Face or similar platforms. |
Domain Expert |
Deeply familiar with related standards and use cases. Works closely with the overall PIC and OpenUSD Experts to formalize requirements and specifications. This may require several domain experts for a given project. An External Partner is preferred, with membership in AOUSD (or intention thereof). |
OpenUSD Expert |
External or internal role. A technically minded expert in OpenUSD. Should be able to pass OpenUSD certification (In Progress). Works in close collaboration with the Domain Expert. The OpenUSD Learning Path offer a great way to onboard experts. |
QA PIC |
Responsible for asset capability testing and end-to-end workflow testing. |
NVIDIA DevRel / SAM & Partner |
NVIDIA teams liaising closely with experts and partners. |
Technical Writer |
Ensures that specifications comply with technical writing standards which will be suitable for formalized standardization submissions. |
Stakeholders |
Industry partners such as OEMs or ISVs as well as NVIDIA teams. |
Process Schematic#
Process Description#
The SimReady Standardization Workflow is a phased process that helps you define and align on scope, iterate on the specification, and deliver a tested and validated candidate specification.
Phase 1: Definition & Alignment#
Defining the scope is essential for developing USD asset capabilities for simulation. Start by detailing the MVP use-case and identifying necessary domain expertise and partners early on. The use-case should clarify initial simulation goals and required 3D asset capabilities, guiding the identification of necessary asset types (e.g., environments, robots, or vehicles). Avoid overly broad scopes; focus on well-defined initial needs and prioritize competing goals to ensure clear alignment with all stakeholders before proceeding.
Specify a domain and identify a domain expert#
PIC will identify the required personas to help achieve the desired domain simulation use-case
Domain Experts need to have deep knowledge of the use case, workflows and associated standards and software
Domain Experts may need to provide sample data / assets
Domain Experts with motivation to participate in standardization through AOUSD are preferred
PIC
Project Kickoff Meeting#
Introduce project members, and associate them to the relevant roles
Create a Project folder to store related documentation. This folder will be used to collate all working documents, usually created from copied templates (see steps below)
Set up meeting schedule and invite stakeholders
PIC
Deliverable: Project Structure Setup
Specify MVP Use Case#
A detailed use case is vital to scope requirements
Include workflow drafts and functional requirements
Start with simple use cases that significantly “move the needle”
For example, determine the most important “asset class” for the domain and explain how it is embedded within the use case. Determine which systems need to interface with it in which manner, and the inputs and outputs that it needs to support.
Perform Data Mapping#
If a standard or de facto data model already exists, perform Conceptual Data Mapping.
If no data model exists, go straight to a Gap Analysis exercise below to identify the inputs and outputs needed within the simulation runtime.
Domain Expert
OpenUSD Expert
Deliverable: Data Mapping Documentation
Perform Gap Analysis#
Identify data models and capabilities that have no existing representation in OpenUSD with the provided template.
Phase 2: Development & Iteration#
Once the simulation use-case and goals are defined and agreed upon, begin prototyping and iterating within OpenUSD to meet runtime requirements. This involves adding and testing custom attributes, incorporating existing schemas (if they exist), and documenting efforts. Connecting prototype USD files to runtime may require additional engineering work and multiple iterations to achieve stability and produce the expected results.
As part of this process, all stakeholders should work to assemble the capability’s written specifications, define and build validators to ensure assets are built to the new specification consistently, and create end-to-end workflows to help others add this new capability to their assets. Thorough testing of these elements is essential for standardization consideration.
Assess Viability and Prioritize Capabilities#
Categorize capabilities into:
Those that can be standardized across simulation runtimes in a specification
Those that are not yet ready for standardization in a specification (e.g., are too specific to a given solver/runtime/software)
PIC
Domain Expert
OpenUSD Expert
Create Asset Specification Draft#
Prototype Content Specification
Prototype and test Schemas
Validate the draft specification in simulation runtimes and sandbox environments
Draft preliminary workflow documentation for stakeholder review based on provided template
OpenUSD PIC
Technical Writer
Domain Expert
Deliverable: Preliminary Specification Documentation
Deliverable: Preliminary Workflow Documentation
Iteration#
Continue to test, tune, and formalize schemas, build asset validation rules and code, and refine the specification documentation.
Finalize Content Specifications
Create sample USD assets that demonstrate specifications and can be validated
Perform a QA Pass
Plan for Internal Beta and Acceptance Criteria Checks
Once Internal Beta is complete, provide to an External Beta
Engineering PIC
OpenUSD PIC
Domain Expert
Technical Writer
Phase 3: Package Deliveries#
To be considered a SimReady specification, all prior work — defining the specification, developing schemas, creating and validating content, and documenting workflows — must be tested and validated end-to-end (from asset creation to runtime results). This ensures others can replicate the steps to deliver consistent assets that work predictably in simulation runtimes.
Engage with NVIDIA stakeholders for a review process to ensure the specification’s broad applicability across multiple runtimes and use cases. Build a checklist to capture and deliver each component to ensure a smooth review.
Candidate Specification#
Collect and polish the written specification document that was created using the spec template in Phase 2
PIC
OpenUSD Expert
Tech Writer
Pipeline & Validation#
Build an end-to-end asset workflow pipeline from source data into a simulation runtime:
Provide a Converter that implements source data and OpenUSD conversion according to Conceptual Data Mapping.
Note: Conversion does not have to be file based, conversion to/from data streams or other sources such as databases is within scope
Supply Validators that check candidate assets for compliance with defined asset specification
Note: Validators generally correspond to Asset Capability Requirements and the framework allows for deep integration
Deliver any post-export, pre-ingest, and/or at-ingest asset transformation logic to conform it to the defined specification and validate its capabilities for a given simulation runtime
Polish and provide preliminary workflow documentation that instructs creators on how to apply the capability, validate it and test it in the intended runtime(s).
Engineering PIC
OpenUSD Expert
Tech Writer
Sample Content#
Provide Physical AI Dataset(s) and/or Sample Content that demonstrate the defined asset capability and can be used as part of the workflow documentation (examples/tutorials) and to exercise the validators.
Content PIC