Sandia Labs FY21 LDRD Annual Report


Machine learning for correlated intelligence. Developing and characterizing the capability of machine learning (ML) algorithms in the electro-optical (EO) imaging space is vital for correlated intelligence applications. Correlated intelligence discovers coincidence in multiple modalities (radioisotope, electro-optical, radio frequency, synthetic aperture radar, etc.). The research team used commercial open source data sets to develop several ML algorithms for detecting objects in overhead imagery. An automated, distributed benchmarking framework was used to evaluate performance. Successful implementation of ML for overhead imagery, combined with automated benchmarking and continuous improvement frameworks, is a critical part of deploying effective, real-time, correlated intelligence applications. The use of continuous improvement, automated benchmarking methods attracted a number of externally funded projects across the Department of Defense. (PI: Emily Moore)

Germanium telluride chalcogenide switches for radio frequency (RF) applications. RF systems are used across the gamut of national security applications, so any device that improves SWaP (size, weight, and power) and signal quality would have a high impact on current national security goals for improved communication systems and communication technology supremacy. This LDRD project

team developed prototype germanium telluride switches that can allow for highly reconfigurable systems, including antennas, communications, optical systems, phased arrays, and synthetic aperture radar. The demonstration of the successful germanium telluride RF switches showcased a critical element necessary for a single chip RF communication system. (PI: Gwendolyn Hummel)

Indirect heating device with 6.0 micrometer gap. S-parameter measurement with insets showing optical images before (OFF state) and after (ON state) heating process; dB values were taken at displayed frequency.



Made with FlippingBook Ebook Creator