Orbit determination based on particle filtering for maneuvering target

The non-Gaussian distribution of state error and measurement error is an important element for accurately modeling the dynamics in satellite orbit determination problem, causing limitations in applying the traditional filtering methods. The sampling-importance-resampling(SIR) particle filter is used in this paper to solve the mentioned problem. The dynamics including the J₂ perturbation is considered in the design. To follow the target's abrupt motion, a multi-model method is used. The residual error is also applied to detect the maneuvering of the target. The model of the target is switched to the current thrust model when the residual error intersects the model aggregates. Finally, a simulation is given to estimate the orbit of a maneuvering target.