Maximum likelihood estimate of parameters of Nakagami-m distribution

Abstract

Nakagami-m distribution is well known for its ability to model a number of probability density functions, be it symmetric or asymmetric. Many Maximum Likelihood parameter estimation techniques for this distribution have been proposed that use estimated higher order moments of the data. However, the required large amount of data may not always be available. This is a drawback of using moments based approaches. In this work we propose a Maximum Likelihood parameter estimation technique for Nakagami-m distribution by giving a closed form expression for it. We demonstrate the performance of the proposed approach using certain test cases and compare the same to conventional algorithms using moments. We show that the new algorithm can model those pdfs better which may be deviating slightly/morderately from Gaussian shape and hence alleviating the need for extra mixture components.