Ground-Based Radar Tracking of Ballistic Target on Re-entry Phase Using Derivative-Free Filters

Radar tracking of a ballistic target in re-entry phase has been considered in this paper. The motion of the target is evaluated with the assumptions that the drag and gravity are the only forces acting on the ballistic target after it enters into the endo-atmospheric phase. With unknown ballistic coefficient, the problem is actually a case for combined state and parameter estimation. After around 1997 the central or divided difference filter was developed for nonlinear stochastic estimation, which is a derivative-free filter and takes care of the second term of Taylor series expansion. The performance of the CDF is expected to be considerably better than the EKF and close to the UKF. This paper addresses the quantitative aspects of the improvement of performance. In particular, using Monte Carlo runs, the performance of the CDF has been compared with that of the EKF, standard UKF and Square Root UKF. While the performance of the CDF is comparable with the UKF, the CDF has a higher computational efficiency compared to the UKF. These features make it one of a strong candidate for on-line implementation in ground based radar tracking