Метод определения оптимальной частоты среза для оценки сигнала по его измерениям

Abstract

A typical procedure for flight test data processing is the measurements smoothing to obtain the estimate of the true signal measured in flight. If the data processing is carried out in the frequency domain, it becomes necessary to estimate the frequency transform (Fourier transform) of the true signal by its measurements. A method is proposed for constructing the simplest estimate of the signal Fourier transform in the form of the low frequency part of the Fourier transform measurements. Such estimate is determined by the single parameter – the cutoff frequency. The indicator for the estimate quality is the mean square of its deviation from the Fourier transform of the true signal. The task of constructing the estimate of the signal Fourier transform is considered as the task of finding the optimal cutoff frequency for which the criterion is minimal. An approximation of the criterion dependence on the cutoff frequency is constructed as the function of measurements only. The estimate of the optimal cutoff frequency is found by minimizing the approximation dependence. The proposed method for estimating the optimal cutoff frequency does not contain adjustable parameters. In the application to the tasks of identifying aircraft aerodynamic characteristics, it is shown that estimates of cutoff frequencies obtained by the proposed method provide identification accuracy comparable to the identification accuracy at optimal cutoff frequencies.