Particle filter based estimation of navigation information during in-flight post ejection

In case of delivery of guided munitions, the orientation of the daughter vehicle is aligned with that of the mother, prior to ejection. A predefined trajectory is provided to the daughter for guidance purpose. In the absence of external perturbation, the velocity matching Kalman filter methods yield fairly good estimation of post ejection navigation information for the daughter. However, after being ejected, the daughter vehicle is subject to perturbation arising due to wind velocity. This prevents the daughter vehicle from accurately following the pre-determined trajectory. Moreover, the direction of wind thrust is such that it introduces additional states to the plant and makes the mathematical model non-linear, where Kalman filter is no longer applicable. Present work is an attempt to investigate the performance of a particle filter to effectively handle the presence of such nonlinearities in the post ejection scenario. Recursive Monte Carlo simulations of realistic trajectories are performed in order to establish effectiveness of the proposed scheme.