An improved c-k class estimation of the regression parameters in aircraft magnetic interference model

Abstract

In aeromagnetic detection, the estimation of the regression parameters in aircraft magnetic interference model is the key of aircraft magnetic compensation. Taking aim at the multicollinearity of aircraft magnetic interference model, a new parameters estimation method called improved c-k class estimation with combination of wavelet threshold denoising and c-k class estimation was proposed. First, wavelet threshold denoising was used to pretreat magnetometer data in order to decrease the noise of electric equipments which will influence the accuracy of parameters estimation. Then c-k class estimation was used in the estimation of regression parameters. In a simulation example, the estimation accuracy of LS estimation, c-k class estimation and improved c-k class estimation was compared in different signal to noise ratio (SNR). The result shows that improved c-k class estimation is more accurate than other two methods, especially more adaptive in low SNR.