An Expert System for Classification of Narrowband and Broadband Signals

Abstract

The research described in this paper presents an expert system for detection and classification of signals in the presence of noise. The pattern recognition expert system detects signals based upon the shape of the power spectrum. The rule based expert system classifies signals. The expert system extracts information for presence of signals from a smoothed power spectral density. For each frame the expert system decides on the presence of signal in a noise dominate atmosphere and the rule based system classifies signals and determine any correlation between different signals.