Suboptimal Nonlinear Filters for Tracking a Ballistic Target
Sihua Liu · Acta Simulata Systematica Sinica · 2005
It is studied the problem of tracking a ballistic object in reentry phase from radar observations. A model with highly nonlinear state and measurement equations is considered and the theoretical Cramer-Rao lower bounds (CRLB) of estimation error are derived. In our work three suboptimal filters are designed and their error performances are compared with CRLB. Besides frequently used filters in nonlinear filtering problems as the extended Kalman filter and the unscented Kalman filter, a new filter combining reduced sigma point unscented transformation with classical Kalman filter is presented. The simulation results favor the last that keeps the balance of good accuracy and tolerable computational complexity.