Abstract
In order to improve true error rate performance under radio frequency man-made noise enviroment, optimum reception based on the statistical characteristics of man-made noise has been proposed. In this optimum reception, the statistical characteristics of mam-made noise have to be estimated in advance. In this paper, we employ Middleton's class-A impulsive noise model as a statistical model of man-made noise, and discuss the parameter estimation of class-A model and the performance of optimum receiver constructed with estimated parameters. Numerical results show that, although the enormous amount of noise sample is required to obtain precise estimated parameters, a good error performance of optimum receiver can be obtained with parameters estimated by a certain amount of noise sample.
