Methods of Orbit Refinement

T h e accuracy to which the position of a satellite or spacecraft can be predicted depends upon the accuracy to which the "initial" position and velocity vector, or related orbital parameters, arc known. Since these parameters can be determined only from observational d a t a which inevitably contain observational errors, the accuracy to which they can be known depends upon the nature, the quantity, the accuracy, and the distribution (in space and time) of the observational d a t a , and t h e way in which these d a t a are processed. The accuracy of the orbital parameters, and of prediction, depend also upon the accuracy to which all of the forces acting on t h e satellite or spacecraft are known and taken into account. Clearly, the term " a c c u r a c y " m u s t be taken largely in a statistical sense. Orbit refinement is essentially data smoothing for the purpose of accurate prediction. Given the n a t u r e of the observational d a t a , and the statistical properties of the observational errors, it is possible to formulate a method of d a t a smoothing and prediction which is optimum in the sense of giving predictions with the greatest possible accuracy. However, such an optimum method will, in general, be useful only as a standard of comparative performance for more practical methods. T h e reason for this is t h a t it has not been difficult to find simpler and therefore more practical methods which are nearly as accurate as the optimum method. (For example, see Ref.