Identification of the trajectory of a propulsed space vehicle

Numerical filtering techniques for the identification of space vehicle trajectories are discussed, notably polynomial filter, Wiener filter and classical Kalman filter techniques. The relationship between these methods and the difficulties associated with each method are described and compared. The Kalman technique is shown to be the best adapted method to solve the divergence problem frequently encountered in the restitution of the unforeseeable trajectory. By introducing a few noise spectral characteristic considerations in the design of the Kalman filter a theoretical solution to the divergence problem is proposed. The results of a computer simulation for post flight data analysis and for real time data analysis are presented.