Skip to main content
U.S. flag

An official website of the United States government

Official websites use .gov
A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS
A lock ( ) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.

Machine Learning Driven Self-correcting Autonomous Metrology Systems (SAMS)

Summary

The NIST on a Chip program's autonomous metrology project is developing machine learning systems that extend the usability of advanced quantum and photonic sensors. Our key innovation is applying physics-informed machine learning to observe and model latent variables in NIST-on-a-Chip (NOAC) sensors, enabling calibration stability over dramatically extended or potentially infinite time horizons. This work aims to deliver laboratory-grade measurements in complex operational environments without frequent recalibration, effectively embedding NIST expertise into field deployments.

Description

While NIST-on-a-Chip has produced quantum and nanophotonic measurement solutions that can significantly reduce calibration frequency, their transition from laboratory to marketplace depends on seamless integration into user workflows. Our research addresses a fundamental challenge: maintaining calibration accuracy over extended timeframes by using machine learning to track unobservable system dynamics.

Our approach seeks to democratize advanced metrology by removing the "PhD-in-the-loop". We're building self-correcting quantum and photonic sensor networks through:

  1. Machine learning models that continuously monitor and compensate for latent variable drift
  2. Machine-controlled measurement systems enabling closed-loop experiment design
  3. Active learning and Bayesian optimization for adaptive calibration
  4. Physics-informed models that incorporate device and material properties

Our primary focus is quantum (NV diamond) and photonic sensor networks for thermodynamic metrology (temperature, pressure, force and humidity). We have recently demonstrated that ML models can reduce temperature uncertainties in NV diamond sensors by as much as 10X.

Opportunities

If you are interested in joining our team as a post-doc, guest researcher, collaborator, or student volunteer send us an email.

Major Accomplishments

Patent Application Number 17/113,222
Hysteresis Compensation in Temperature Response of Fiber Bragg Grating Thermometers Using Dynamic Regression

Provisional Application Number 63/683,468
NV-DIAMOND BASED RADIATION DOSE RATE SENSOR

Created December 9, 2021, Updated February 28, 2025