| In the current navigation field,X-ray pulsar navigation using pulsars as signal source has become a very promising spacecraft navigation method.The basic measurement of the X-ray pulsar navigation is the pulse time delay value,which is obtained by comparing the phase of the integrated pulsar's pulse profiles with respect to the phase of the standard pulse profiles.How to quickly and accurately estimate the integrated pulsar's pulse profiles time delay value is the key to improve the performance of the pulsar autonomous navigation.Compressed sensing is a classical method for estimating pulse time delay in X-ray pulsar navigation.Compressed sensing consists of the following aspects: sparse dictionary for signal transformation,measurement matrix sampling,and recovery algorithm for signal reconstruction.On the one hand,it is considered that the number of atoms in the dictionary is closely related to the estimation accuracy,the more the number of atoms is,the smaller the interval between atoms is,and the more accurate the estimation is,the larger the amount of calculation is.On the other hand,the measurement matrix is an important factor affecting the application of compressed sensing,the performance of the measurement matrix will affect the signal estimation effect,the limited dimensions of the measurement matrix will lead to the limited dimensions of the signal and affect the memory consumption.Therefore,from the perspective of the waveform dictionary and the measurement matrix,it is a worthwhile approach to implement integrated pulsar's pulse profiles estimation method based on compressed sensing with high accuracy and low computational complexity.This paper studies the time delay estimation from the following two aspects:(1)A time delay estimation method based on two-stage compressed sensing is proposed.Aiming at the problem that the increase of the number of atoms in the dictionary of compressed sensing brings about large computation while improving the estimation accuracy,the proposed method combines rough estimation with precision estimation as two level dictionaries.Firstly,the global phase estimation of the integrated pulsar profile is carried out by using the feature of large atomic interval of rough estimation dictionary,and the estimated delay value is obtained.Then,by making use of the characteristics of the small atomic intervals and numbers which aresuitable for local estimation of the precise estimation dictionary,the exact time delay estimation of the integrated pulsar profile can be performed.(2)Combining with intelligent optimization algorithm,an arbitrary dimension Hadamard measurement matrix selection method based on adaptive genetic algorithm is proposed in this paper,and which is applied to pulsar time delay estimation based on compressed sensing.Aiming at the problem that the limited dimension of the Hadamard matrix leads to the constraints of the pulsar signal length and the large memory consumption,this method adopts adaptive genetic algorithm to select partial Hadamard measurement matrix with any dimension according to the rules,and then uses compressed sensing to obtain pulsar time delay estimated value.In this paper,the above two methods are simulated by using numerical simulation and RXTE(Rossi X-ray Timing Explorer)real data.Theoretical analysis and experimental results show that the amount of data in the two level dictionary is two orders of magnitude smaller than traditional dictionary.The proposed method reduces the computational complexity greatly compared with traditional compression sensing method with the same time delay estimation accuracy.The dimension of the measurement matrix has been reduced to match the corresponding pulsar's cycle,and the time delay estimation effect of the patial Hadamard measurement matrix selected by the adaptive GA based on certain rules is superior to that selected randomly.Therefore,these two methods proposed in the paper can effectively improve the accuracy of time delay estimation while reducing the amount of calculation,and increase the accuracy of navigation. |
