Advanced Low Probability of Intercept/Low Probability of Detection Radar (LPI/LPD) Techniques Using Artificial Intelligence Driven Methods

ARTEMIS, Inc. had developed a next-generation software-defined radar called the SlimSDR.  We have demonstrated the system with a variety of low-probability-of-intercept / low-probability-of-detection (LPI/LPD) waveforms and operational modes.  Additionally, ARTEMIS brings Machine Learning (ML) experience in optimizing waveforms and interference mitigation.  The goal of this SBIR project is to design and develop advanced LPI/LPD radar techniques using artificial intelligence (AI) driven methods. The use of LPI/LPD radar techniques in adversarial environments has become common in recent years.  Many techniques and waveforms have been used, such as wide-bandwidth, frequency-agility, adaptive control of signal power, sidelobe reduction, binary phase codes, polyphase codes, Barker codes, Frank codes, and Polytime codes.  Many techniques are now used to counter LPI/LPD operation, some of which rely on machine- and deep learning-based detection and localization algorithms.  These can dramatically reduce the effectiveness of conventional LPI/LPD radar techniques. Highly-adaptive, advanced, AI-based LPI/LPD radar techniques need to be developed for the Navy to maintain operational advantage in near-peer conflicts.  ARTEMIS proposes to develop and demonstrate multiple synergistic LPI/LPD radar techniques that can quickly adapt in response to the tactical environment.  We will leverage our LPI/LPD waveform and AI / ML experience, developed through years of work on our current and past radar projects, including our next-generation SlimSDR radar system development (Army C5ISR Airborne Radar Branch, SBIR, Phase II Sequential and AFRL/RWWI Phase II SBIR), to meet the requirement of this SBIR topic. In Phase I, ARTEMIS will create modeling and simulation (M&S) tools for exploration of LPI/LPD radar techniques. Additionally, counter-techniques will be employed to measure the ability of the newly-developed techniques to avoid detection and classification, aiding in the development of quickly adapting waveforms and operational modes.  AI and/or ML frameworks will be developed to guide the waveform and system operation through AI-based decision engines to facilitate this high level of adaptability.  ARTEMIS will show how the developed LPI/LPD techniques counter the advanced ML and deep-learning-based detection systems. ARTEMIS will also evaluate the computational requirements of the proposed solutions, as well as algorithm training requirements, as radar platforms vary greatly in processing capacity.  The radar requirements will also be evaluated, as few radar systems are capable of operating in way that is compatible with highly adaptive LPI/LPD techniques.  Also, employing these concepts will have an impact on radar performance, ARTEMIS will quantify this impact and use it as one of the metrics for measuring the AI-based LPI/LPD concepts.  The SlimSDR will be used in Phase I as a stand-in radar to demonstrate the adaptive LPI/LPD techniques.