Research On Detection And Identification Technology For X-ray Pulsar Signal Based On S Transform

With the intensive competition in space field,the realization of the spacecraft autonomous navigation is of great engineering and strategic significance.X-ray Pulsar-Based Navigation(XPNAV)can provide navigating information with high precision for the near-earth,deep-space and interplanetary spacecrafts by exploiting the periodic signal from X-ray pulsar.XPNAV has many advantages,such as low cost,high reliability,strong anti-interference feature,and wide range of application prospect.Therefore XPNAV has huge potential in the spacecraft autonomous navigation fie ld and is one of the pioneer fields in the spacecraft autonomous navigation research.This paper discusses the key problems of XPNAV such as the pulsar signal detection,time delay measurement,integrated profile recognition and de-noising technique.The main work and achievements of this paper are summarized as follows:1.In order to improve the quality of pulsars‘ integrated profile and the accuracy of X-ray Pulsar-based Navigation System(XPNAVS),a novel filtering method in S-transform domain is proposed based on chi-square distribution.First,the definition of S-transform(ST)power spectra of pulsar signal is presented.Then,we derive a linear relationship between the mean ST power spectra of white noise and the frequency.After that we draw the conclusion that the local ST power spectra of Gaussian White Noise(GWN)follows chi-square distribution with two degrees of freedom.We have verified by performing the theoretical analysis and simulation experiment.Then pulsars‘ integrated profile is filtered by the distribution characteristic of the local ST power spectra of GWN.The simulation result has verifieds the effectiveness of the filtering algorithm,and the proposed method is superior to a state-of-the-art method based on the Gaussian distribution in improving the pulsar profile and enhancing the accuracy of XPNAVS.2.In order to overcome the shortcomings of traditional X-ray pulsar signal detection algorithms such as the large computing amount and poor detection performance in low signal-to-noise ratio(SNR),a novel Constant False Alarm Rate(CFAR)detection algorithm is proposed based on S-transform.Firstly,threshold filtering is performed on the ST power spectrum of the cumulative signal according to the distribution characteristic of ST power spectrum of GWN.Then,the sum of the time-frequency power spectral after threshold filtering is used as the test statistic and the distribution characteristic of the test statistic is analyzed theoretically.It can be concluded that the test statistic follows the Gaussian distribution,which has also been verified by the Monte Carle method.Based on this,the decision threshold is calculated using the probability density function(PDF)of the test statistic so as to realize the CFAR detection.Theoretical analysis and simulation results have revealed that the proposed method has low computational complexity and it can improve the detection probability effectively.Moreover,the time delay of pulsar signal can be obtained under certain accuracy.3.Focusing on the X-ray pulsar signal detection in high background noise,a class of CFAR detection algorithms is proposed based on Time-Frequency Entropy(TFE).Firstly,from the view of energy distribution,the conclusion that there are differences in quality from TFE of noise and signal is verified by theoretical derivation and experimental simulation.Then the distribution characteristic of TFE of pulsar signal is analyzed theoretically.The theoretical analysis result is further verified by Monte Carle method.We detect the observation signal is based on CFAR with the test statis tic designed by TFE and we briefly discuss the application of the algorithm on time delay measurement.We have also verified the effectiveness of this method by simulation experiment.The simulation results show that the time-frequency R é nyi entropy detection algorithm based on ST generates the optimal detection performance.4.Focusing on the problem that the traditional Singular Value Decomposition(SVD)filtering algorithms can not well handle the noise and signal details at the same time,a novel SVD filtering algorithm based on ST is proposed.The algorithm separates the diagonal matrix into a signal subspace.Then a noisy subspace and a pure noise subspace,the optimal filter parameters are determined by solving the cost function.After that the real signal from background noise is extraced.The simulation results have revealed that the filtering algorithm can efficiently reduce the noise and well preserve the signal detail.5.A new algorithm based on S-transform is proposed for accurate and fast recognition of the pulsars‘ signals.We find that the time-frequency distribution of pulsars‘ integrated profiles based on S-transform is basically invariant with different observed frequencies and signal-noise ratios.This has been verified by simulation results.Based on these constant signatures,first the minimum standardized Euclidean distance classifier is employed to classify pulsars‘ integrated profiles.Then the recognition rate expression of algorithm is presented.Finally,the experimental results show that the proposed method has good noise immunity performance,higher recognition rate and fast processing speed.