Signal processing by function elimination filters

In many applications of signal processing (in areas of communication, pattern recognition, measurement) it is necessary to detect (or classify) the signals in the absence of certain a priori information about its parameters. One of the effective methods of solving such problems - the Function Filtration Method - is discussed. Method of the determination of the generating equations form for the process with unknown parameters is proposed. A general approach to the design of algorithms for measurement under invariance conditions on the informationless parameters of the signal, whose instability is equivalent to noise, is given. This method allows the synthesis of highly effective filters for the elimination of signals with known function while their parameters are a priori unknown. Applications of this method are given for some special cases.