Application of a Local Additive Approximation Method for Evaluating the Efficiency of Maximum Likelihood Algorithms for the Joint Estimation of Regular and Discontinuous Information Process Parameters

An asymptotic method for evaluating the efficiency of the joint estimates of the observed information process parameter vector is considered, in view of these estimates are synthesized using the maximum likelihood method. It is assumed that the specified vector of parameters includes an arbitrary number of regular parameters and an arbitrary number of discontinuous parameters. Based on the additive-multiplicative representation of the moments of the decision determining statistics, the general asymptotic expressions for the probability densities and the first two moments of the estimates are found. It is shown that, under the conditions of a high a posteriori accuracy, the estimates of regular and discontinuous parameters are statistically independent. The application of the study results is illustrated by the example of determining the characteristics of the estimates of the time of arrival (the discontinuous parameter) and the central frequency (the continuous parameter) of a pulse signal with a Gaussian random substructure. #COMESYSO1120.