Design and implementation of multi-input adaptive signal extractions

This report covers research on design and implementation of novel adaptive digital signal processors for signal extraction (i.e., reduction of noise, interference, and distortion). The signal extractor configurations and adaptive adjustment schemes considered are extensions and generalization of prior art (viz., the FIR digital filter and the LMS adjustment algorithm) that consist of various combinations and interconnections of conventional adaptive filtering and noise cancelling configurations and associated adaptive adjustment schemes. The methods of implementation considered are based on speed-cost efficient Residue Number System arithmetic, and include Read Only Memory table look-up, and microprocessors and minicomputers for hardware control and for software implementation. The major goals of the research are improvements over prior art in signal extraction performance (as measured by signal distortion, signal-to-noise ratio, and mean squared error) and in the cost of implementation and speed of operation. Results on design, implementation, and performance of configurations and adjustment schemes are described.