The Robust Estimation for the Parameters of a two-dimensional chirp signal model

Abstract

In this paper, the robust estimate of the parameters of the non-linear regression model (Two-Dimensional Chirp Signal Model) was found. This is by employing the restriction ( ) estimates method. Which is briefly known as (CM) estimates. For the purpose of knowing how efficient this method is to find the robust estimate of model parameters, Compared to the least squares method (LS) by using the simulation technique, by assuming both the normal and Cauchy distribution of the error. Use default parameter values, different levels of standard deviation, and sample size thus, compare the methods of estimation using a statistical measure mean square error (MSE). It was found that the (CM) estimates were close to the smaller least square estimates (LS) if the error is normally distributed. As for Cauchy’s error distribution, the least squares method collapsed in finding the estimate of parameters for this model. While the (CM) method has proven to be a robust to finding parameter estimation, giving the results of parameter estimates close to the actual values, and (MSE) values to estimate parameters. As well as the model is too small compared to (MSE) values to estimate parameters and the model when (LS) is used in estimation.