A unified theory of adaptive locally optimum processing

By using the concept of deflection and treating the signal as of known or unknown structure, a unified approach to adaptive locally optimum processing can be developed. This theory brings together processing techniques which have been treated by various authors for application to the detection of either broadband or narrowband signals in undersea surveillance and communications. The theory applies in particular to the detection of weak signals in the presence of interference characterized by a multi-state model. The processing can be accomplished in either the time domain by processing complex samples of the analytical signal or in the frequency domain by processing complex Fourier coefficients. Furthermore, algorithms which process the amplitudes and phases of the samples are derived. These algorithms implement both first and second order detectors (in terms of partial derivatives of the probability density function of the interference)