European XFEL / DESY
ALPHA
A new sense for a remarkable machine.
Building an intelligence layer for the DOOCS control environment: from autonomous metadata discovery to the next generation of acquisition and anomaly analysis.
THE CONTEXT
Alpha in development. Not in operational use. DAQ rollout is planned as the next step.
More than individual signals.
An accelerator is a system of systems. Its signals describe magnets, RF, vacuum, cooling and much more. Finding a property is only the beginning. Understanding what it means, where it belongs and how it relates to other signals is the real challenge.
A system map that keeps learning.
Our alpha work at DESY starts with discovery. Background agents aggregate available property metadata and use names, descriptions and related information to infer candidate systems and their hierarchy. The aim is a living system map that gets more useful with each investigation, with its assumptions available for review.
AUTONOMOUS DISCOVERY
From identifiers
to understanding.
Individual property descriptions become candidate relationships. Relationships form a hierarchy: from the facility, through its subsystems, down to the signal.
FACILITY / DEVICE / LOCATION / PROPERTYIllustrative inferred mapping — inspectable and correctable.
The project’s planning estimate for the property space to explore. Discovery and selective acquisition scale separately. This is not a measured throughput or a count of simultaneously subscribed channels.
THE PATH TO SYSTEM INTELLIGENCE
Connected. Step by step.
Discover the property space.
Explore DOOCS facilities, devices, locations and properties. Collect available metadata and turn disconnected identifiers into searchable context.
Infer the system behind the signal.
Background agents connect descriptions and naming patterns to candidate systems and relationships. Inferred mappings remain inspectable and can be corrected.
Acquire what matters.
The next step is a staged DAQ rollout. Dedicated workers support ZeroMQ subscriptions and RPC reads, configurable sampling and batched transfer for selected channels.
Learn behavior. Investigate change.
Connect acquired data to forecasting, reconstruction and representation-based analysis. The goal is evidence-backed findings and agent-led investigations, with people setting the scope.
TECHNICAL CONTEXT
DOOCS provides ZeroMQ-based communication and also supports RPC. Anomx builds on this with dedicated workers. The goal is efficient, selective acquisition and a traceable connection between data, models and the physical system.
Alpha-stage research and development. This case describes the Anomx integration being built, not a production deployment or an official product endorsement by DESY or European XFEL.
THE NEXT ERA
Give your system
a new sense.
Build the future of autonomous systems with us.
Request early access