Automated Configuration of BaSyx DataBridge using Standardized Asset Administration Shell Interface Modelling

In modern industrial environments, integrating heterogeneous systems and protocols is still painfully manual. Digital twins and data pipelines promise interoperability, but bridging raw machine data to structured digital representations often means editing piles of JSON by hand—slow, brittle, and error-prone.

The Asset Administration Shell (AAS), part of the RAMI 4.0 framework, gives us a standardized way to describe, synchronize, and interoperate physical assets (robots, sensors, software) as digital twins. Eclipse BaSyx’s DataBridge is a key piece here: it connects real-world data sources to AAS SubmodelElements via protocols like REST, Kafka, MQTT, and OPC UA.

The problem? Traditional DataBridge setup requires multiple JSON files per variable—transformers, protocol configs, AAS server definitions, and routes. With a handful of tags it’s fine. With hundreds (and frequent schema tweaks), it becomes a maintenance trap.

From manual setup to model-driven automation

During the digitalization of the XXL Pilot Factory (CONVERGING project), the team needed to sync nearly 300 variables into the BaSyx AAS Server. Debugging cycles forced repeated changes across configuration files, quickly revealing the limits of manual work.

The breakthrough: model the DataBridge configuration itself as an AAS—machine-readable, standardized, and complete enough to generate all DataBridge files automatically.

 

 

 

 

 

 

The DataBridge Configurator

We built the DataBridge Configurator, a Python tool that reads a dedicated “DataBridge AAS” and auto-generates all required BaSyx DataBridge configuration in seconds.

 

 

 

 

 

 

 

 

This model relies on three submodels:

Communication Message (custom): Describes the raw data structure from the source (objects, arrays, properties) as AAS elements (SMCs, SMLs, Properties).

AID – Asset Interface Description (IDTA): Standardized definition of protocols (e.g., Kafka, HTTP), endpoints, and interaction metadata (topics/paths, content types).

AIMC – Asset Interface Mapping Configuration (IDTA): Standardized mapping between sources → message structure → HTTP/AAS destinations.

Together, these submodels define how data should flow from origin systems into the AAS. The Configurator consumes that definition and emits all the JSON the DataBridge needs (consumers, transformers, AAS server, routes).

What gets generated:

• consumer.json (e.g., Kafka subscriptions, groups)
• aasserver.json (AAS endpoints and property paths)
• JSONata files (one per variable)
• jsondatatransformer.json (index of transformations)
• routes.json (links sources, transforms, and sinks)

Results: setup time drops from hours to seconds; mappings are consistent and verifiable; changes in the AAS model regenerate the whole configuration instantly.

Repo (complete demo): https://github.com/aimenct/basyx-databridge-configurator

Demonstration & validation

At AIMEN Technology Centre’s XXL Pilot Factory, an ABB robot, operator tracking, and ergonomic monitoring publish data (joint positions/torques, tool state, body kinematics, RULA) into Kafka. The DataBridge synchronizes this data into the AAS Server.

 

 

 

 

 

 

 

 

 

 

With the Configurator:
• ~280 variables were configured automatically.
• Per-variable JSONata transformations, AAS endpoints, and routing were generated cleanly.
• Debugging and iterative changes became straightforward, update the AAS model, regenerate, redeploy.

Why this matters

Time savings: from hours of hand edits to inmediately generation.

Fewer errors: one source of truth in AAS; mappings are machine-readable.

Agile by design: changes in data structures or targets regenerate cleanly.

Standards-based: leverages IDTA AID/AIMC and aligns with IEC 63278-1.

What’s next

Upstream contribution: propose the component to the Eclipse BaSyx Databridge ecosystem.

Dynamic runtime config: move from static files to in-memory loading—update routes and transforms without restarts.

Template evolution: track AID/AIMC spec updates and propose standardization of the Communication Message submodel.

More automation: auto-generate Communication Message submodels from JSON and derive HTTP interface definitions from the asset’s own AAS.

UX improvements: dedicated UI to model faster AIMC submodel.

 

Presented at ETFA 2025

This work “Automated Configuration of BaSyx DataBridge using Standardized Asset Administration Shell Interface Modelling” was presented at the ETFA 2025 event in Oporto, showcasing AIMEN Technology Centre’s approach to scalable, automated digital-twin integration.

 

Creators

Andrés Pérez | Davinia Fernández | Emilio Costa | Lucía Alonso