U.S. Department of Energy

Pacific Northwest National Laboratory

Controlled Datasets to Inform the Study of Sensor Degradation

Principle Investigator: 

We are developing a process for studying the drift and degradation of signatures arising from sensor platforms with the express aim of improving confidence in the signature-building process. Implementation of this process will allow us to understand the effect that controllable and uncontrollable factors have on sensor performance.

Approach

The application of this methodology has accomplished or will enable:

  • accurately predicting the impact that agiven operating environment will have on measurements
  • tuning sensors or analysis to account forenvironmental change or sensor drift
  • discriminating changes (drift) in the target phenomenon from sensor or environmental variation
  • recognition of natural change versus deliberate interference with measurements

By providing ground truth data and a validation pathway, the results will provide insight into the following fundamental questions posed by the Signature Discovery Initiative:

  • How and why do underlying phenomena, the associated signature, or our perception of them change?
  • How do we define a signature that is flexible and adaptable, yet tightly correlated to the phenomenon?
  • How do we determine when to redefine a signature?

Impacts

Our methodology systematically informs the design of better, more robust and easily interpretable signatures. This approach improves measurement systems and makes data analysis systematic, rigorous and easy to understand.

Project Staff: 
Marvin Warner, Alejandro Heredia-Langner, Brian Clowers, Spencer Hays, Matthew O'Hara, Mark Phillips, Jon Wahl
| Pacific Northwest National Laboratory