Compound digital filter

Abstract

This paper describes the identification by the compound digital filter with two estimators involved in the Kalman filter, the first corresponds to stochastic gradient describing the transition function and the second obtains the innovation process gain through fuzzy logic. The filtering process has time constraints for your response occurring before the reference system changes to another state. The results are formally described throughout the document. Auto-Regressive Moving Averages model was proposed (ARMA (1, 1)) as the reference system. The filtering algorithm was developed in MatLab® software getting the integrated filter operations illustrated with its graphics of convergence and response, concluding with brief an analysis of the results.

Keywords

• signal srocessing
• estimation
• identification
• real-time
• fuzzy logic
• digital filters