Computationally Efficient Wideband RF Sensing Algorithm for Threat Detection

BANC3, Inc. in Princeton, NJ has developed a wide instantaneous bandwidth software defined receiver (SDRX) architecture that processes the input RF signal in both the frequency and time domains simultaneously and continuously in real time. The spectra and pulse descriptor words generated by the SDRX are processed with an intelligent data reduction technique that significantly reduces the data throughput while preserving the accuracy of the signal parameters. A computationally efficient wideband RF sensing algorithm was developed that operates on the reduced data sets to reliably detect and classify threat signals in complex RF environments. This algorithm was successfully modeled during Phase I and the results indicated a correct ID of 76% for complex signal environments with signal densities up to 763 signals/second using 10 different threat signal types.  Under Phase II, this algorithm will be integrated on to an embedded ARM processor in the Direct RF FPGA eliminating the need for external GPUs. This configuration is well suited for SWaP constrained applications deployed in austere environments (drones, combat vehicle, dismounted soldier). The algorithm employs threat signal templates that can be generated through modeling and/or in real-world collections. Based on the latest developments with integrated direct RF sampling devices, we anticipate developing and demonstrating a hardware-in-the-loop configuration which will deliver 6.4 GHz of instantaneous bandwidth over an RF range of 100 MHz to 36 GHz that can reliably detect RF enabled threat signals to improve the survivability of ground platforms. Our solution will employ an open architecture design to facilitate integration into ground platforms.   Key performance criteria to be measured during Phase II include reaction time, latency, and throughput as well as threat signal recognition accuracy, precision, and recall.