Identification of Class a Noise Parameters via Least Square Gradient Method

Abstract

The Middleton Class A interference model is a statistical-physical and parametric model for man-made and natural electromagnetic (EM) interference. In this paper, the efficient estimation of the Class A model parameters based on least square gradient method is derived. The considered estimator converges fast and low-complexity with performance approaching theoretical optima for large data samples. Simulation of this estimator with three unknown parameters indicates that this technique is efficient. Index Terms—Middleton Class A Model. Impulsive Noise. Parameter Estimation. Non-Gaussian Noise. characteristic function has simple form(12). In this paper we proposed a method for parameter estimation based on the characteristic function spectrum estimation from observation samples. Our method is well suited not only for two-parameter estimation of Class A model like Zabin's work(10), but also for estimation of full three-parameter estimation and adaptive to track changes for channel noise. The later is critical to the implementation of signal detection and estimation algorithms in non-Gaussian noise environment.