Sandia Labs FY21 LDRD Annual Report


Low SWaP tunable hyperspectral video imager helps analyze explosions. Explosive events consist of rapidly varying spatial, spectral, and temporal information that produce a key signature to determine an explosive event’s composition and configuration. Ideal measurement of explosive events is performed with a single low size, weight, and power (SWaP) instrument, which is capable of high spatial resolution, high-speed temporal sampling,

Sandia’s low SWaP high speed tunable spectral computational imager helps to determine an explosive event’s composition and configuration.

and fine spectral resolution. This 4D dataset, also called a datacube, is not measurable with a single traditional 2D imaging system and requires compressive and computational optical methods. This LDRD project team designed, built, and tested a low SWaP diffuser-based high-speed spectral computational optical imaging system. The low SWaP spectral imager was successfully tested in a lab-based setting and during outdoor explosive testing. The team also developed image reconstruction methods, both traditional and machine learning, to reconstruct the spatial and spectral scene with high fidelity. Sandia partnered with New Mexico Tech to develop and build a prototype multichannel spectral imager as a complementary instrument to Sandia’s imager. This project showed that computational optical methods can enable measurement of critical optical signatures for Sandia and NNSA’s non-proliferation missions. (PI: John VanderLaan)

Quantifying the security of a cyber system. Although the national security community widely recognizes the need for evidence-based cybersecurity, cyber analysts largely rely on qualitative metrics and expert intuition. The Science and Engineering of Cybersecurity by Uncertainty Quantification and Rigorous Experimentation (SECURE) Grand Challenge laid a foundation for cyber modeling and experimentation to catalyze the use of quantitative metrics and analytical evidence to inform high- consequence national security decisions. The SECURE project produced Defender, an objective, verified, and validated process that quantifies the security of a cyber system. As a result, it is now possible to quantify return on cybersecurity investments, rigorously compare two proposed remediations, identify critical components

Each data point represents a combination of command and control experiment, scanning/detection experiment, and attacker/defender strategy. Defender’s goal is to push the attacker toward top left of the plot.

for improving security and model fidelity, and quantify attack consequences. The project created tools to automate and orchestrate cyber experiment scenarios, solve multi-level adversarial optimization problems, and reduce the number of expensive, high fidelity runs. The LDRD team developed the ability to optimize resilience against two cyberattacks on a power grid control system, as well as methods to quantify, compare, and rank defenses to quantify the resilience of the defender system. (PI: Ali Pinar)



Made with FlippingBook Ebook Creator