FIELD: physics.
SUBSTANCE: method involves receiving signals from the current observation direction using an antenna system with integral (with number j=1), azimuth-differential (with number j=2), elevation angle-differential (with number j=3) channels and antennae q=1, 2,…Q compensation channels, coordinated processing of signals of each channel using Q+3 identical receiving channels, analogue-to-digital conversion of signals of each receiving channel and processing the digitised signals of each channel by calculating vectors of weight coefficients of the integral, azimuth-differential and elevation angle-differential channels and weighted summation of signals of the integral and compensation, azimuth-differential and compensation and elevation angle-differential and compensation channels, characterised by that, in order to calculate vectors of weight coefficients, constraint matrices for the shape of the resultant beam pattern of the integral, azimuth-differential and elevation angle-differential channels are formed first for all possible directions of observations in form of readings of values of the beam pattern of the integral, azimuth-differential, elevation angle-differential and compensation channels in r=1, 2,…, R constraint points, wherein for an odd number of constraint points, one of them lies in the given direction of observation, and the rest of the constraint points lie symmetrically about it and the direction finding planes; for an even number of constraint points, their positions lie symmetrically about the direction of observation and the direction finding planes; deviation of constraint points from the direction of observation is selected based on the minimum root-mean-square value of the error in estimating angular coordinates of the direction finding signal for the expected distributions of values of angular coordinates of arrival directions and power of the directing finding signal and noise; the constraint matrices are stored in constraint matrix memory; when processing digital signals of each channel, covariance matrices are formed for the integral, azimuth-differential, elevation angle-differential and compensation channels; the values of the constraint matrices for the integral, azimuth-differential and elevation angle-differential for the current observation direction are read from the constraint matrix memory, and to eliminate the effect of calibration errors, i=1, 2,…, I values of the regularisation parameter 
are generated in form of a geometrical progression with the quotient Δµ=2…3 and initial value equal to the level of inherent noise of receiving channels; values of vectors of weight coefficients for the integral, azimuth-differential and elevation angle-differential channels corresponding to each regularisation parameter are generated based on the constraint matrices in accordance with a defined expression; the sum of squares of magnitudes of the weight coefficients for the integral, azimuth-differential and elevation angle-differential channels is then calculated for each value of the regularisation parameter; weighted summation of signals of the integral and compensation, azimuth-differential and compensation and elevation angle-differential and compensation channels is carried out and estimates of angular coordinates of the useful signal source are calculated for all values of vectors of weight coefficients for different values of the regularisation parameter; values of squares of angular distance between the angular coordinate estimates are calculated for neighbouring values of the regularisation parameter; values of angular coordinate estimates are calculated, which correspond to the first, in ascending order of the regularisation parameter, minimum square of angular distance between angular coordinate estimates provided the sum of squares of magnitudes of weight coefficients exceed an upper threshold value, and in the absence of said minimum, the angular coordinate estimates are those values which correspond to the value of the regularisation parameter where the sum of squares of magnitudes of the weight coefficient is closest to the lower threshold value without exceeding it.
EFFECT: high accuracy of direction finding.
6 dwg
