M.I. Botov – Ph. D. (Eng.), Associate Professor, Military Chair, Siberian Federal University. E-mail: email@example.com
V.A. Vyahirev – Ph. D. (Eng.), Associate Professor, Deputy Chief of Military Chair, Siberian Federal University. E-mail: firstname.lastname@example.org
The basics of the theory of unbiased, efficient and consistent (enough) the estimation of parameters of radar signals in adaptation of spatial and time-frequency characteristics of the measuring system to the external correlated noise. The specifics of the approach is that the measured values of the radar signal in the adaptation of the measuring complex to the modes of interference taking power in nature: the signal / (interference + noise) is dependent on the distance between the target and source of noise on the measured parameters. In this case, the traditional measurement algorithms for maximum incomplete sufficient statistics or its derivative are biased, inefficient and inconsistent (not enough). The systematic error is due to distortion: a) pattern measuring angular coordinates the formation of gaps in it in the direction of the sources of active interference, and b) the amplitude-frequency spectrum of the signal when setting dips (ridges suppression) AFC meter Doppler frequency drive to the crests of the spectrum of passive interference to suppress it, and c) characteristics of the discriminator range tracking system in suppressing interference, which differs from the signal delay time (in particular, the suppression of impulse noise, going on the range). Ineffectiveness of (growth of the fluctuation of error) is related to its failure, since the measurement of the signal parameters against external interference involves unconditional consideration of all the elements of a sufficient statistic corresponding signal model.
The problem is solved in two ways:
a) by means of non-adaptive algorithms, which are invariant to uninformative signal parameter (in this case – to the energy of the expected signal.) Quality indicators measuring these algorithms (systematic and fluctuation errors) are determined once the errors of the signal parameter estimation uninformative;
b) using adaptive algorithms that use a smoothed estimate uninformative signal parameter. Adaptation is that the accumulation of single uninformative parameter estimates resulting algorithm on the parameters of quality close to that of the quality of the algorithm with the known non-informative parameter.
With a significant excess of the signal / (interference + noise) is a non-adaptive threshold and adaptive measurement algorithms have roughly the same quality. When approaching the signal / (interference + noise) to the threshold of measurement provides an adaptive algorithm significantly more accurate assessment of the informative parameter relatively maladaptive. Gain in accuracy increases with the decrease of the signal / (interference + noise).