Abstract:

In X-ray pulsar-based navigation, strong X-ray background oise leads to a low signal-to-noise ratio (SNR) of the bserved profile, which consequently makes it very difficult to btain an accurate pulse phase that directly determines the navigation recision. This signifies the necessity of denoising of the bserved profile. Considering that the ultimate goal of denoising s to enhance the pulse phase estimation, a profile denoising lgorithm is proposed by fusing the biorthogonal lifting wavelet ransform of the linear phase characteristic with the thresholding echnique. The statistical properties of X-ray background noise fter epoch folding are studied. Then a wavelet-scale dependent hreshold is introduced to overcome correlations between avelet coefficients. Moreover, a modified hyperbola shrinking unction is presented to remove the impulsive oscillations of the bserved profile. The results of numerical simulations and real ata experiments indicate that the proposed method can effectively mprove SNR of the observed profile and pulse phase estimation accuracy, especially in short observation durations. And it also outperforms the Donoho thresholding strategy normally used in combination with the orthogonal discrete wavelet transform.