Sub-Nyquist Classification of Frequency Agile Waveforms

In this program, TrellisWare Technologies, Inc. proposes to develop a three-stage architecture that detects, identifies, and classifies frequency agile signals across extremely wide bandwidths (e.g., 10+ GHz). Using this architecture, signals hopping periodically across extremely wide bandwidths can be classified without resorting to a brute-force, computationally infeasible narrowbanding and energy detection approach. Although TrellisWare proposes to focus on off-line processing of synthetic data in Phase I and real-world data in Phase II, the proposed sub-Nyquist signal processing techniques would all support real-time operation in an FPGA-accelerated SIGINT analysis platform.