A Trajectory Prediction Method Based on Parametric Modeling of Trajectory Curve Evolution Patterns

Aiming at the trajectory prediction problem of hypersonic glide vehicles (HGVs), a trajectory prediction algorithm based on the parametric modeling of trajectory curve evolution laws is proposed. To address the trajectory prediction problem under the condition of inaccurate estimation of state characteristic parameters, the directly measurable target states are taken as the research object, and the trajectory prediction algorithm is designed by employing the differential geometry theory to extract the spatial motion features of the target. Firstly, the differential geometry theory is utilized to decouple the motion states of the target, and simultaneously, the maneuvering characteristic parameters describing the target’s motion laws are accurately extracted. Secondly, to prevent the mutual interference of different modal data, the ensemble empirical mode decomposition (EEMD) method is utilized to separate the trend and period terms of the characteristic parameters. Finally, the autoregressive (AR) model is employed to establish a parametric model of the target’s motion states and perform extrapolation. By combining this with the dynamics equations, the vehicle’s motion states are reconstructed to realize trajectory prediction. Simulation results demonstrate that the proposed trajectory prediction algorithm possesses high prediction accuracy.