DEPS Abstract

Guest >> Sign In

DIRECTED ENERGY PROFESSIONAL SOCIETY

Abstract: 25-Systems-032

UNCLASSIFIED, PUBLIC RELEASE

AI/ML Applicability to Directed Energy Kill Chain

Recent developments in embedded GPU and FPGA processing have enabled consideration of high-performance, high-computation AI/ML-based techniques for developing automatic image processing components in the Directed Energy Kill Chain. Such techniques can include linear filters, support vector machine, and neural networks, with their own performance benefits and training/data requirements.
These techniques could each be applied to several steps of the Kill Chain, including target acquisition, coarse track, fine track, aimpoint selection, and battle damage assessment. However, the kill chain steps each has different imaging system requirements regarding field-of-view, resolution, and latency, such that the decision to use a specific AI/ML, versus non-AI/ML solutions is not always clear.
Discussion, informed by recent experiments at AFRL/RDL, will traverse the Kill Chain and consider generic imaging system requirements and potential embedded computation resources for each step, to compare algorithm costs and benefits. In addition, non-AI/ML alternatives are considered which may be considered at various steps to build comprehensive, integrated kill chains.
The talk is presented in an overview style and is directed towards technical Government researchers without specific AI or kill chain expertise. Listeners will gain understanding of the Directed Energy Kill Chain, an understanding of the costs and benefits of AI/ML, as well as applications where AI/ML might or might not apply to the kill chain.

Approved for public release; distribution is unlimited. Public Affairs release approval # AFRL-2025-4908

UNCLASSIFIED, PUBLIC RELEASE