Detection of signals in impulsive first order moving average noise

Abstract

In this thesis, we consider the detection of weak signals in additive impulsive noise described by the first order moving average (FOMA) model. Specifically, we derive decision regions of the maximum-likelihood (ML) and suboptimum ML (S-ML) detectors in the FOMA noise model, and obtain specific examples of the ML and S-ML decision regions. We apply the ML and S-ML detectors in the antipodal signaling system, and compare the bit-error-rate performance of the ML and S-ML detectors in impulsive noise. Numerical results show that the S-ML detector exhibits practically the same performance as the optimum ML detector. The performance gap between detectors for FOMA and independent and identically distributed noise becomes larger as the degree of noise impulsiveness increases.