Study On The Signal Processing And Navigation Filtering Algorithm Of Pulsar

With the development of space technology in recent years,pulsar has been widely concerned in the field of spacecraft autonomous navigation.As a new type of autonomous navigation technology,pulsar navigation can provide abundant navigation information for near space,deep space and interplanetary spacecraft.In addition,as a natural object,pulsars are far from the Earth,which helps the pulsar autonomous navigation technology shows the advantages of strong reliability and anti-interference ability,along with shows great potential for the development and strategic significance of Chinese space industry.This thesis focuses on the research of signal processing method and navigation filter algorithm in pulsar navigation,including the following aspects:In view of the importance of the profile quality of the pulsar signal for the navigation algorithm,a nonparametric regression analysis method is utilized to denoise the pulsar signal profile.Kernel regression is a common method in nonparametric regression,and compared with wavelet method and empirical mode decomposition,this method could obtain better results by denoising the profile after epoch folding.However,improper selection of bandwidth will affect the denoising.Therefore,the kernel regression algorithm is improved to reduce the sensitivity of the denoising effect to the bandwidth.In addition,the method of kernel density estimation is also proposed to deal with the photon arrival time series directly.This method is a more effective method to estimate the arrival time information as much as possible compared with the kernel regression method.Finally,the experimental results show that the improved kernel regression and kernel density estimation can improve the profile quality of the pulsar signal.Phase is an important measurement in a pulsar navigation system,and it's change can reflect the position changes,while the position changes are closely related with the velocity.Based on the detailed analysis of the relationship between phase and Doppler frequency,the Doppler frequency information is extracted from the phase time curve by digital phase locked loop,and the simulation is verified by RXTE measured data.As for the navigation filtering algorithm,the measured value of the phase can be utilized as the input information of the extended Kalman filter,and then the motion state of the spacecraft is corrected according to the basic principle of the incremental estimation method.Based on the relationship between phase and Doppler frequency,the Doppler frequency information is further introduced into the extended Kalman filter algorithm.Since the multi-pulsar navigation system has a completely observable line,the input information of the filter is increased after the introduction of Doppler frequency information,and the system is still completely observable.Through the simulation analysis,the proposed system could achieve good convergence results.