READS: Real-time Edge AI for Distributed Systems
Description
This project will develop and deploy low-latency controls and prediction algorithms at the Fermilab accelerator complex
Detailed example
The ML outputs of the system are inferences as to the origin of beam loss in the Main Injector acclerator enclosure and also suggested regulation ramps to best improve the Spill Duty Factor in the Delivery Ring for Mu2e
AI / analytics pattern
Classical/Predictive Machine Learning: Models trained on data to make predictions or classifications based on identified patterns or relationships.
Automation level / stage
c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.
Expected benefit
READS has two sub-projects. The first project created the means to stream live Main Injector and Recycler accelerator beam loss monitor data. This data is then fed to an AI model deployed on an FPGA so that it can infer, in realtime, the origin of beam loss, either Main Injector or Recycler, for each beam loss monitor in the tunnel enclosure. The second project aimed to improve upon traditional resonant beam extraction regulation techiniques using AI for use in the Fermilab Delivery Ring and Mu2e.
Audit / financial statement impact
The use case does not have an effect on civil rights/liberties/privacy, access to education/housing/insurance/credit/employment, access to critical government resources/services, human health/safety, critical infrastructure/public safety
Controls / human review
ATO: No; PIA: Not published
Data needed
research datasets from scientific experiments