Blind Adaptive Frequency-Shift Filtering for Cochannel Interference Removal

The objectives of this project are to comparatively evaluate four methods (three of which we have recently devised) for blind adaption of frequency-shift filters. These filters exploit the spectral redundancy exhibited by cyclostationary communication signals to separate spectrally overlapping signals (cochannel interference). The evaluation will be performed by simulating the techniques in various interference scenarios and using the Monte Carlo method to measure output signal-to-interference-and-noise ratio and, for digital signals, bit error rate. These techniques are potentially applicable to numerous signals intelligence problems, as well as to interference mitigation in both military and commercial communications. The result of this work will be the identification of the most promising methods for various types of applications, and a determination of the feasibility of a programmable signal-separating workstation or one or more dedicated processors for signal separation in specific applications.