Open AccessParameter estimation of G0 distribution based on improved recursive expectation–maximisation method for clutter modelling
Author(s) -
Lu Jiaxin,
Sun Yuze,
Zhuo Bangsheng,
Yang Xiaopeng
Publication year - 2019
Publication title -
the journal of engineering
Language(s) - English
Resource type - Journals
ISSN - 2051-3305
DOI - 10.1049/joe.2019.0511
Subject(s) - clutter , computer science , algorithm , radar , estimation theory , matrix (chemical analysis) , maximum likelihood , sample (material) , distribution (mathematics) , mathematical optimization , mathematics , statistics , telecommunications , mathematical analysis , materials science , chemistry , chromatography , composite material
Modelling and simulation of clutter are important in radar signal processing, the G0 distribution is generally adopted to simulate the ground clutter in radar echoes. In order to improve the modelling accuracy of clutter modelling, the actual data should be used for model parameter estimation. However, in some special situations, the actual data sample is very small. Existing methods cannot estimate the parameters of G0 distribution efficiently. To solve this problem, an improved recursive expectation–maximisation (EM) method is proposed to estimate the parameters of clutter in this article. This method combines the expectation step and maximisation step in one equation. Through recursive method and simplification of the positive definite matrix, this proposed method can obtain maximum likelihood estimation more efficiently than the conventional EM method and recursive EM method. Simulation results show that the performance of the proposed method is better than that of the conventional methods for a small data sample.