A Stochastic Drag Model for Improved Uncertainty Realism of Select LEO Orbit Predictions

Atmospheric drag is a principal cause of uncertainty in predictions of low Earth orbits and a key factor concerning the degree of uncertainty realism achieved by prediction models. The stochastic effects of atmospheric drag have most often been accounted for by considering the atmospheric density or another related quantity, such as the inputs to an atmospheric density model or the ballistic coefficient, as a random variable. Instead, this paper investigates a novel approach in which an uncertainty factor \(\kappa\) $\kappa$ is appended to the drag equation and considered the lone random variable of air drag. Time series of \(\kappa\) $\kappa$ values are estimated from observational data and analyzed as second-order Gauss–Markov processes. These random processes are then parameterized and used as the stochastic foundation of two stochastic orbit prediction models which prove capable of achieving a high degree of uncertainty realism in the prediction of certain low Earth orbits when validated against the results of an empirical prediction covariance analysis.