Multi-vector Integration And Estimation Of Celestial Velocity Measurement Navigation In Deep Space Exploration

Nowadays,deep space exploration is one of the most important space activities in the world.It is also an embodimentof comprehensivenational power and innovationcapability.Deep space missions involve relatively high technology risks.In comparison with near-earth satellites,deep space probes have their characteristics such as large communication delay,unknown flight environments,one-chance orbit insertion maneuver,higher requirement of onboard autonomous management,etc.Therefore,continuous,real-time and high-precision navigation becomes one of the key technologies of deep space missions.In the background of the first Chinese Mars orbiting and rovering mission as well as the celestial velocity measuring navigation method proposed in recent years,this dissertation focuses on the navigation principle,system modeling,navigation target selection,velocity vector integration,error analysis and state estimation,etc.The main contributions are summarized as follows:The geometry observation model of celestial velocity measurement navigation system is established based on the Doppler effect of light.The probe velocity vector estimation method is provided without the orbital dynamic model.To further improve the navigation precision,the orbital dynamic model for deep space celestial velocity measurement navigation system is established and model error characteristicand its influence on velocity measurement navigation system are also analyzed.The observability of velocity measurement navigation system is estimated based on the index of observability order and observability matrix condition number.The influence model of solar global activities on characteristic spectrum is established based on the Yunnan Observatory New Vacuum Solar Telescope(NVST)observation data.The impact of solar source instability on the performance of velocity measurement navigation system is also studied.Using the stability of radial velocity as an evaluation index,7 radial velocity measurement reference stars and 35 sets of three-star combination are proposed based on ground telescope observations.The error generation mechanism of a space heterodyne spectroscopy is studied.The error model is established for celestial spectrum detection instrument and the impact of instrument errors on the performance of velocity measurement navigation system is also studied.Focusing on the propagation of velocity determination error of a three-star velocity measurement navigation system,the mapping relationship between the measurement error in three directions and velocity determination error is given.By theoretical analysis,it is proved that the influence of measurement error on velocity determination error reaches a minimum when three-star directions are orthogonal to each other.The probability density function of velocity determination error is derived when measurement errors in three directions are subject to zero-mean Gaussian distribution.Focusing on the velocity measurement navigation target selection problem,a target selection method is proposed based on matrix column orthogonal degree.Moreover,the variation tendencies of orthogonal degrees of different three-star combinations are obtained using numerical simulation.Focusing on the over-determine equation problem of multi-velocity vector integration for spacecraft celestial navigation,a least square optimal estimation method is proposed.In this method,multi-velocity vector integration problem is equalized to linear estimation including measurement noises.The quadratic sum of difference between actual and estimated measurements is used as an optimal index.The theoretical optimal estimation expression is obtained when the number of measurements is larger than 3.Using the trace of state estimation error covariance matrix as an evaluation index,the optimal star distribution on celestial sphere is derived and proved when the number of measurements equals 3.Using numerical simulation,traversal search is performed to give approximate optimal solution of star distribution on celestial sphere when the number of measurements is from 4 to 8.A space-ground combined navigation method based on celestial velocity measurement is proposed.The radial velocity between probe and star is introduced as a new observation in addition to the ground radio ranging and velocity measurements.Navigation filter is applied to estimate the state of the spacecraft.Compared with traditional ground-based radio navigation,the position and velocity estimation accuracy in the Mars capture phase can be improved effectively when the proposed method is utilized.Furthermore,the accuracy improvement conforms to the prediction of Markov optimal estimation theory.The larger flight distance is,the more performance advantages will show of integrated celestial navigation,and the position estimation accuracy will be at least two times better than the ground radio navigation.Finally,a hardware-in-the-loop simulation system is established to verify the performance of the celestial velocity measurement navigation system.The simulation results show the effectiveness of the proposed methods.