INTELLIGENT FUSION TECHNOLOGY, INC. — Department of Defense SBIR Phase I: AF181-041

Description

:In military satellite communication (MILSATCOM), the downlink transmission signals suffer from nonlinear distortions caused by the high power amplifier (HPA) of the satellite transponder, which can be characterized as amplitude modulation-to-amplitude modulation (AM-AM), and amplitude modulation-to-phase modulation (AM-PM) effects. To solve this problem, Intelligent Fusion Technology, Inc. (IFT) and its university partner California State University propose three different predistortion (PD) solutions depending on current and future MILSATCOM system configurations, and potential modification cost, which are near-term on-ground solution, mid-term on-board solution, and long-term combined on-ground and on-board solution. In this proposal, we mainly focus on the near-term on-ground solution due to its lowest cost of implementation. In this solution, a physical model-based PD and a machine learning-based PD controller will be developed to predict the nonlinearity of the HPA in the presence of the jamming signals. Various advanced techniques such as extended Salehs HPA model, PD modeling architectures, and particle swarm optimization (PSO) are applied in the physical model-based PD, while the PD controller is obtained by training a sophisticated deep and reinforcement learning (DRL) model using suitable signal and HPA characteristics datasets, which are generated by the IFT MILSATCOM emulators developed in our prior work.