System and method for signature and extraction, detection, discrimination, and localization

Abstract

A system and method whereby acoustic signals can be classified and identified as to nature and location of the original signal. The system and method determine from an arbitrary set of signals a signature or other characterizing feature and distinguish signals associated with a plurality of conditions by means of dictionaries comprising atoms of signals.

What is claimed is: 
     
         1 . A system for extracting, detecting, classifying, discriminating, localizing, and processing signals comprising:
 a. at least one transducer device operable to collect a plurality of signals;   b. a processing device coupled to said transducer device, said processing device executing instructions for decomposing the signals into representational spaces matched with one another, each said representational space defining a multi-dimensional set of atoms, wherein the decomposing comprises performing simultaneous sparse approximations of the signals, each signal being thereby represented by a combination of atoms selected from said representational space, each selected atom being weighted by a corresponding coefficient; and,   c. a computer memory coupled to said processor for storing said representational spaces.   
     
     
         2 . The system as recited in  claim 1 , further comprising a detection unit coupled to said processing device, said detection unit selectively determining a match between a set of said signals responsive to respective representational spaces thereof. 
     
     
         3 . The system as recited in  claim 1 , wherein said transducer device collects includes at least one sensor. 
     
     
         4 . The system as recited in  claim 1 , wherein said processing device executes a Greedy Adaptive Discrimination (GAD) decomposition operating with a dictionary of prototype atoms. 
     
     
         5 . The system as recited in  claim 4 , wherein the dictionary comprises a set of Gabor atoms together with Fourier and delta function bases. 
     
     
         6 . The system as recited in  claim 1 , wherein the said atoms are selected from a random set of atoms. 
     
     
         7 . The system as recited in  claim 1 , wherein each said atom selected for said representation of at least one signal is caused to be orthogonally related to at least one other atom in said representational space for said signal. 
     
     
         8 . The system as recited in  claim 1 , wherein said processor device generates a set of test atoms based on said representational spaces, and defines a dictionary of said test atoms, a subsequently collected signal being decomposed in terms of test atoms selected from said dictionary, whereby a degree of similarity is determined for the subsequently collected signal relative to said representational spaces of previously collected signals. 
     
     
         9 . The system as recited in  claim 1 , comprising a plurality of said transducer devices collecting signals through a plurality of channels, the signals of at least one said channel being alignment shifted in a predetermined measurement space relative to the signals of at least one other of said channels. 
     
     
         10 . A method of signature extraction comprising the steps of:
 a. establishing two or more analysis sets, each set comprising at least one signal;   b. decomposing the analysis sets into matched structure books, wherein the decomposing comprises executing a processing device to perform simultaneous sparse approximations of the signals of the analysis sets; each said structure book including at least one representation defining a multi-dimensional combination of coefficients and atoms; and,   c. storing said structure hooks in a computer memory.   
     
     
         11 . The method as recited in  claim 10 , further comprising actuating the processing device to produce a parametric mean structure book representative of a signature for each set of signals. 
     
     
         12 . The method as recited in  claim 10 , wherein a separate simultaneous sparse approximation is performed for each analysis set. 
     
     
         13 . The method as recited in  claim 10 , wherein a single simultaneous sparse approximation is performed for an aggregate collection of signals from a plurality of analysis sets. 
     
     
         14 . The method as recited in  claim 10 , wherein at least one of the signals is alignment shifted in a predetermined measurement space relative to another of the signals before simultaneous sparse approximation. 
     
     
         15 . The method as recited in  claim 10 , wherein the signals are collected at a plurality of different locations, the signals collected at each location determining at least one analysis set, a location specific signature being determined based on the resulting structure books. 
     
     
         16 . A method of detecting, classifying, discriminating, and localizing signals comprising the steps of:
 a. collecting a set of signals;   b. decomposing the sets of signals into matched structure books, wherein the decomposing comprises executing a processing device to perform simultaneous sparse approximations of the signals; each said structure book defining a multi-dimensional combination of coefficients and atoms; the simultaneous sparse approximation being executed with respect to a dictionary defined relative to previously determined signatures; and,   c. comparing the resulting structure books to structure books of the previously determined si matures.   
     
     
         17 . The method as recited in  claim 16 , wherein a single simultaneous sparse approximation is formed for an aggregate collection of signals from at least two groups of signals, each defining one of said analysis sets. 
     
     
         18 . The method as recited in  claim 16 , wherein step c is accomplished by examining the values of the coefficients of the atoms in the resulting structure books compared to typical values for similar atoms of one or more known signatures. 
     
     
         19 . The method as recited in  claim 16 , wherein the simultaneous sparse approximation is accomplished with respect to a dictionary comprising an aggregate collection of atoms present in previously extracted structure books of known signatures of interest. 
     
     
         20 . The method as recited in  claim 19 , wherein the dictionary additionally includes the aggregate collection of all atoms selected during the sparse approximation that produced the previously extracted structure books of known signatures of interest, whether or not said atoms are present in said structure books of signatures of interest. 
     
     
         21 . The method as recited in  claim 16 , wherein a plurality of signals are collected and at least one of the signals is alignment shifted in a predetermined measurement space relative to at least one other signal before the simultaneous sparse approximation operation. 
     
     
         22 . The method as recited in  claim 21 , wherein a plurality of relative shifts are executed on at least one of the signals relative to another, a simultaneous sparse approximation operation being executed upon each relatively shifted signal set, and a preferred relative shift being determined based on the simultaneous sparse approximation structure books resulting respectively from the relative shifts. 
     
     
         23 . The method as recited in  claim 16 , wherein:
 a. the collected signal or signals are sparsely approximated in aggregate with at least one previously established signal; and   b. the structure books of the previously established signal or signals are compared to those of the newly collected signal or signals in order to classify the new signal or signals as either similar or different from the previous group.   
     
     
         24 . The method as recited in  claim 16 , wherein a plurality of signals are collected at different locations and, subsequent to detection, time and phase information from the resulting structure books at different locations are used to extract directional information about the signals. 
     
     
         25 . The method as recited in  claim 16 , wherein:
 a. a plurality of signals are collected at different locations;   b. at least one signal from at least one first location is analyzed independently to detect signatures;   c. at least one atom of one detected signature is reconstructed to produce a signature component;   d. at least one other location channel is searched for the reconstructed signature component to determine relative time of arrival information therefor; and,   e. signals for at least one location channel are time shifted relative to the signals for the first location before simultaneous sparse approximation relative to said signature component atoms.   
     
     
         26 . The method as recited in  claim 16 , further comprising the step of reconstructing at least one component of a decomposed signal for restoration back to a signal domain thereof when collected. 
     
     
         27 . The method as recited in  claim 16 , wherein the comparison of structure books is performed with respect to a subset of atoms containing fewer atoms than at least one structure book in said comparison. 
     
     
         28 . The method as recited in  claim 27 , wherein said subset of atoms is determined by a comparison between the structure book atoms of at least two known signatures.