Fuzzy Digital Filter Type III

摘要

A digital filter interacts with a reference model signal into a real process in order to obtain the best corresponding answer, having the minimum filter output error using the mean square criterion. A fuzzy mechanism into the filter structure was added obtaining an intelligent filter, selecting and emitting an answer decision according to the external reference signal changes. This is in order to actualize the best correct new conditions updating a dynamic process. The fuzzy filter makes an interpretation of the input signal level to select the best parameter values from a set of membership values into the knowledge base (KB), updating the filter weights giving the approximation answers according to the reference signal. This fuzzy stage improves the filter answers, minimizing the filter error criterion with a classification of its operation levels. The filtering process requires that all answers of the error criteria are probabilistically bounded, considering the Nyquist and Shannon assumptions, having the fuzzy filter simulations using Matlab tools into the Kalman structure.