GPS Based Absolute And Relative Attitude Determination Of Spacecraft

In this thesis, the method of the attitude determination by GPS is developed. The method consists of two parts: ambiguity resolution and attitude determination.The theory of ambiguity resolution is elaborated, and model of ambiguity resolution is established. An algorithm of ambiguity resolution based on genetics algorithm is proposed. By whitening the covariance matrix of flat ambiguity, correlation between ambiguity components, the number of local extrema of cost function is reduced. The search program getting into precocity is avoided by suitable encoding and genetic operation. Combining with the genetic algorithms, the right ambiguity is rapidly obtained. A integrated algorithm of ambiguity resolution is proposed. By using triple carrier phase, integer Gauss transformations, QR factorization, Cholesky factorization, and geometry constraint, the correlation between ambiguity components is reduced, and more error ambiguity can be discarded, also process of ambiguity searching getting rapidly. The ratio test combining constraint of baseline is used to fix ambiguity rapidly. The algorithm is tested by static and dynamic experiment.Theory of attitude determination by GPS is addressed in detial. Error of attitude determination by GPS is analyzed. Then a Kalman filter is developed in which taking the components of attitude matrix as state variable, Kalman filter is simplified and the veolecty of algorithm is raised. Furthermore a distributed Kalman filter is developed. In this algorithm, Kalman filter is divided into three independent parts, which is processed at the same timed. For Kalman filter will tend to instability if dynamical model is incorrect, two algorithms is developed named predictive filter, Which are based on kinematics model and dynamics model respectively. The procedure of the algorithm consists of three steps, first, model error was estimated, second, model error is used to correct system model, finaly attitude is estimated with corrected model. Simulation result indicated that predictive filter of the dynamics model may be followed the attitude movement, if disturbance moment is time variable and large relative to gradient moment of gravity, and predictive filter based kinematics issuitable if there is no information about dynamics.Two algorithms are developed to estimate relative attitude between spacecraft in formation flying, one is algorithm baesd QUEST (QUaternion ESTimator) and the other is least square estimation based on element of attitude matrix. The two algorithms do not use information of relative attitude dynamics between spacecrafts. Relative attitude is estimated by pseudo observation. Error of relative positioning is analyzed qualitatively.A test platform for three dimensions attitude and relative attitude determination is established. The algorithms and the test platform is tested. The attitude determination algorithm of predictive filter, QUEST algorithm, least square estimation based on element of attitude matrix are tested by the platform.