Universal Signal Matching for RF Threat Classification
In a successful Phase I program for Navy SBIR topic N092-113,"Universal Signal Matching for RF Threat Classification,"Michigan Aerospace Corporation demonstrated robust techniques for Specific Emitter Identification that combine advanced time-frequency analysis to generate feature vectors for individual pulses with cluster analysis for de-interleaving. Features extracted from this analysis are classified by Ensembles of Decision Trees, which provide robust identification as well as known/unknown detection. Feature vectors of unknown emitters are added to the library and additional examples of known emitters are added to supplement the higher-order statistics of those clusters in feature space. Phase II will focus on extending these capabilities and implementing the PerSEIve platform for adaptive signal processing, envisioned as an enabling framework for development, comparison and combination of algorithms to achieve high accuracy and allow for rapid processing of huge amounts of data for validation. PerSEIve will utilize subcontractor Optillel"s hybrid GPU/CPU techniques and will also benefit from the Hybrid Real Simulator of our other partner, General Dynamics-AIS, which creates synthetic data streams as observed from platforms of interest. Furthermore, PerSEIve will support an incremental migration path from prototype algorithms to any target field unit.
Small Business Information at Submission:
Michigan Aerospace Corporation
1777 Highland Drive Suite B Ann Arbor, MI -
Number of Employees: