Research On Key Technologies Of Space Target Orbit Determination Based On Astronomical Positioning

Currently,the competition for resources in Near-Earth space is becoming increasingly fierce,with an increasing number of orbiting spacecraft.While these satellites make life more convenient for people,they also contribute to the deterioration of the space environment.Determining the orbit of space targets through space-based optical surveillance is one of the key technologies for space situational awareness.The use of directional information for orbit determination is complex,so how to improve orbit determination methods,how to obtain position information for space targets,and how to use position information for orbit determination are currently some of the problems that need to be solved.This article proposes a dual-star positioning system consisting of two optical surveillance satellites to surveil the same space target and measure its position vector.By introducing the position vector,orbit determination for space targets is achieved.To implement the proposed space target surveillance scheme,two aspects are studied: star identification algorithms for space target surveillance and space target orbit determination based on position vectors.(1)This article studies the star identification algorithm for space target surveillance,which is used to assist in space target surveillance.A neural network star identification algorithm is proposed,which is efficient and reliable.For neural network star identification,a new guide star feature – Equal Frequency Binning Radial Feature(EFBRF)is proposed,which combines radial features and equal frequency binning discretization.It has rotation invariance,can resist a certain degree of noise,and is suitable for neural network classification.A three-star verification strategy based on the principle of the triangle algorithm is proposed,which maintains a low misidentification rate and improves the reliability of the algorithm.Finally,A method for identifying remaining stars is introduced,which makes it possible to identify all the stars in the star image and is suitable for auxiliary space target surveillance.(2)This article studies the orbit determination problem of space targets based on position vectors.First,a dual-satellite astronomical positioning system is constructed to measure the position vector of space targets.The system uses cameras on two optical surveillance satellites to form a binocular vision model,which can have a longer baseline due to the relatively far distance between orbiting satellites,enabling the dual-satellite astronomical positioning system to locate more distant space targets.Through simulation tests,the error range of the system under different scenarios is obtained.Then,a improved Gibbs method based on the least squares fitting plane is proposed for the orbit determination of space targets.This method can fit a more realistic orbit plane,making the calculated orbital elements more accurate.The final simulation test proves that the calculated orbit elements are closer to the true values.