No abstract

No abstract

No abstract

No abstract

We consider a correlation functional determined for the complex-valued normal Markov process and obtain analytical expressions describing the statistical properties of such a functional. We present plots for the probability density of the random values of the correlation functional.1. Solution of applied radiophysical problems is in particular based on laws of mathematical statistics. For example, a widespread method for obtaining information on the state of the ionosphere consists in recording of the correlation function of incoherent scattering of microwave radiation and in subsequent estimation of the ion and electron temperatures of the ionospheric plasma [1,2]. For such an approach, the estimate of the additive functional of the correlation (autocorrelation) function is an information parameter. The desired estimates of the autocorrelation function are also random because of the random nature of the experimentally obtained sets of elements of the sequences (readouts) {z 0 , z 1 , . . . , z N }. In turn, this results in statistical straggling of the estimated ion and electron temperatures of the ionosphere. Developing the statistical estimation algorithms assumes that information on the distribution law of the estimated parameter or the related quantity is known.Let us consider the additive autocorrelation functional of the form:where {z 1 , z 2 , . . . , z 2N } is the sample of size 2N , extracted for the general set of readouts of the normal complex-valued Markov process z(t), τ is the subscript denoting the discrete delay step, τ = 0, 1, . . . , N, and the asterisk denotes complex conjugation. Functional (1) can be associated with the following integral analog determined from the continuous realization z(t) of the normal Markov process:where T is the time interval of observation and τ is the time interval of delay. By virtue of the statistical relation among the values of the normal Markov process z(t), the distribution of the correlation functional K τ has a significantly different form depending on N (or T ), τ , and the parameters of the normal Markov process. Therefore, we have the problem on the degree of influence of the *