Localization of Coordinates using AI and Technology

To address the Air ForceÆs need for high-fidelity 3D environment models and RF propagation patterns to develop low-latency CNNs for RF geolocation and signal type identification, Intellisense Systems, Inc. (Intellisense) proposes to develop a new Localization of Coordinates Using AI and Technology (LOCATE) system based on a custom deep neural network (DNN) trained on physics-based high-fidelity 3D modeling simulations to fuse information from in-phase/quadrature signals and 3D map data. Specifically, the innovation in training low-compute DNNs on thousands of simulations of RF propagation patterns in the specific mission-relevant environments will enable LOCATE to geolocate and characterize RF signals in low-latency DNNs in complex environments prone to high levels of multipath propagation. As a result, this system will be able to perform context-specific RF geolocation on low-SWaP (size, weight, and power) devices with high accuracy. In Phase I, Intellisense will conduct a feasibility study and develop prototype algorithms for the LOCATE system, focusing on creating RF geolocation algorithms that excel in accuracy and efficiency within a 16-hour deployment window for Class I and II unmanned aerial system (UAS) platforms. These algorithms, compatible with Zynq Ultrascale+ RFSoC hardware, will be trained using high-fidelity 3D simulations to ensure robust performance in diverse environments. Phase II will build upon this by integrating the selected algorithm into a UAS prototype payload, rigorously testing it in various geolocation scenarios, including challenging indoor contexts. Intellisense will evaluate the performance of the LOCATE prototype for direct-path, single-bounce and double-bounce geolocation contexts and will geolocate targets including those that may exist in low-impedance indoor locations, e.g., inside windows. By the end of Phase II, LOCATE aims to achieve a technology readiness level (TRL) of 5, with emphasis on mitigating risks associated with AI robustness and adaptability to different UAS platforms, thus ensuring readiness for potential DoD integration and commercial applications.