出版社:Ivan Kozhedub Kharkiv National Air Force University
摘要:There are some questions of extended Kalman filters application in cognitive radars,which provide real-time adaptation of probing signals parameters for targets and jamming scenario,are considered. A high computational cost is a well-known prob-lem of realization of the adaptation algorithms. In the cognitive radars this algorithms are based on interpolation polynomial expansions and (or) stochastic approximation technics for recurrent filters as usual. Thereby simple extended Kalman filters using can be appropriate for cognitive radar in cases of weak nonlinearity and smoothness of recurrent filtered functions. In the article,a method for taking into account the correlation moments of estimates of smoothed rectangular coordinates of a target for separated filtering of trajectory parameters based on the use of the constant coefficients proposed. The reasonability of the constant coefficients using and slow character of changes of the correlation constant coefficients determined. The choosing of correlation coefficients in dependence of targets position in space proposed. The requirements of computing resources of cogni-tive radars that use extended Kalman filters can decreased by the proposed method. The simulation modeling in case of two-dimensional radar presented. Conclusions verified by obtained simulation results. Additional consideration of correlation de-pendences of estimates allows a partially reconstruction of initial information about orientation of semi-axes of error ellipse of initial measurements. This makes it possible to increase the accuracy of information output with low computational costs and can help speed up the convergence characteristics of adaptation algorithms in the cognitive radars as well.
其他摘要:Розглянуті питання застосування розширених фільтрів Калмана в когнітивних радіолокаційних станціях. Такий підхід може бути доцільним у випадку ”слабкої” нелінійності та гладкості функцій,що підлягають рекурентній фільтрації. Запропоновано метод врахування
关键词:cognitive radar;method for taking into account the correlation moment;separated filtering of trajectory pa-rameters;extended Kalman filter.