Abstract: An acoustic signal processor that samples acoustic data, performs a spectrum analysis of the sampled acoustic data, stores the acoustic data in corresponding frequency bins, and then filters acoustic data stored in a frequency bin by applying a power-law arithmetic operation to the frequency bin acoustic data, such that the power-law arithmetic operation is adaptively changed based on an order dependent determination of whether the frequency bin acoustic data is representative of noise or clutter, or whether it is representative of a signal of an object.

Abstract: An acoustic signal processor that samples acoustic data, performs a spectrum analysis of the sampled acoustic data, stores the acoustic data in corresponding frequency bins, and then filters acoustic data stored in a frequency bin by applying a power-law arithmetic operation to the frequency bin acoustic data, such that the power-law arithmetic operation is adaptively changed based on an order dependent determination of whether the frequency bin acoustic data is representative of noise or clutter, or whether it is representative of a signal of an object.

Abstract: A method for characterizing degree of phase fluctuation of a signal and a signal processor for accomplishing such characterization, is described, in which a time series of complex data vectors are received, initial spectral processing is accomplished, an estimate of the excess phase rotation is made, and a quantity WSC equal to WSC = [ 1 M ⁢ ∑ i = 3 N ⁢ B ⁡ [ C ⁢   ⁢ Φ i ] L ] is calculated.

Abstract: A novel method and apparatus for signal processing are described which discriminate between signals and noise based on the magnitude of the phase fluctuations. In an embodiment of the invention, the governing equations of the process adaptively change in response to the phase and amplitude characteristics of the signals being received for processing. In an embodiment of the invention, an automatic detection method and apparatus is described which has the capability to alert an operator of the presence of signals having phase fluctuation characteristics which match a set of preset conditions (e.g. small fluctuations, medium fluctuations, or large fluctuations).

Abstract: The invention involves a computationally simple and fast method for aligning the phases of complex signals (real (in phase) and imaginary (quadrature) components) before performing a coherent (in phase components) or vector sum, thus achieving the high gains of in phase summations without pre-processing overhead to determine an optimum progression of phase shifts to enable in phase summations. The method of the invention for aligning the phases, without preprocessing to get an optimum set of phase shifts, uses the temporal relationship that exists in the phase angles of successive vectors that occur evenly spaced in time.

Abstract: An apparatus and method for filtering the effects of large amplitude signal fluctuations from elements of the cross-spectral density matrix of a detor array. The outputs of the array are sampled T times, and the values of each element of the matrix formed by using: ##EQU1## where Z.sub.li is the complex output of an sensor of the array at time i; Z*.sub.mi is the complex conjugate of the output of an array sensor at said time i; and k is an integer greater than 1. Because of the exponent k, small, relatively constant, terms will dominate, reducing the effect of large amplitude fluctuations. Preferably, the above value will be scaled according to: ##EQU2## which, besides being similarly filtered by the exponential terms, the resultant matrix element will have a magnitude corresponding to the actual power of the detected signal, after filtering.

Abstract: A filter, and method of the filtering, for increasing the signal to noise tio of relatively small and constant amplitude signals in the presence of relatively large and varying amplitude signals. The filter produces a filtered power level a.sub.z proportional to: ##EQU1## where x.sub.i are power realizations of the signal in a selected frequency bin, and i=1, 2, . . . , N, and z is a non-zero real number not equal to -1, 0, or 1, and is preferably positive. For positive z, because filtered power level a.sub.z depends on a sum of (x.sub.i).sup.-z, sum a.sub.z disproportionately favors smaller, more stable, signals.

Abstract: A method for enhancing signal-to-noise ratio and resolution of amplitude stable signals wherein underwater acoustic data is first collected with an array of hydrophones, and then the data is digitally sampled. After producing spectra of sequential time snapshots of the digitally-sampled data, the spectra are beamformed for a single frequency. Next the low resolution beamformer response is deconvolved from the data by use of a calculated beam response pattern for the hydrophone array, so that many high-resolution estimates are created for each time snapshot. Finally, the resulting high resolution estimates are reduced to a single estimate for each spatial bin across all of the time snapshots, and the high resolution, high gain results are displayed.

Abstract: A scheme for filtering noise from a signal, in which the noise fluctuates significantly in phase. The signal is sampled to produce N members of a time series and the phase of each signal extracted in the form of a magnitude and a phaser, i.e. for a time series .vertline.x.sub.1 .vertline.e.sup.i.theta..sbsp.1, .vertline.x.sub.2 .vertline.e.sup.i.theta..sbsp.2, . . . , .vertline.x.sub.N .vertline.e.sup.i.theta..sbsp.N, where .vertline.x.sub.n .vertline. is the magnitude of the nth member of the N time samples, and e.sup.i.theta..sbsp.N the phaser for the nth sample. The phasers are added to form a complex value s such thats=.SIGMA..sub.N=1.sup.N ?Re(e.sup.i.theta..sbsp.n)+iIm(e.sup.i.theta..sbsp.n)!.Phase noise which fluctuates strongly will tend cancel itself in the sum s, whereas phase resulting from a stable signal will tend to reinforce. Thus the scalar .vertline.s.vertline.

Abstract: A nonlinear technique for high-resolution data processing produces high-rlution power spectra from the output of a conventional processor. First, a "guess" of the spectral estimate is formed. Then the spectral estimate is convolved with the system response function to produce estimated output levels. The differences between the estimated output levels and the measured output levels are used to modify the spectral estimate to produce a new "guess". This process is repeated until certain criteria are satisfied.