Filter sensitivity in exoatmospheric target vehicle tracking

Abstract

An analysis of the orbit determination capabilities of the miniature seeker technology integration (MSTI) program was conducted. In the analysis an extended Kalman filter and an adaptive extended Kalman filter were used to estimate the position and velocity of an exoatmospheric target vehicle. A number of cases were investigated, including measurements from up to three spaced-based observers using angles-only measurements; angles plus range measurements; and angles, range, and range-rate measurements at each observation time. The target was tracked from the burnout point at 59 sec after launch through 180 sec after launch and final estimates for position and velocity were computed at 180 sec. Both the extended Kalman filter and the adaptive extended Kalman filter were found to be unsuitable for estimates using one satellite and angles-only measurements. The estimates were greatly improved when adding range measurements or a second observation satellite. The addition of a third observation satellite did not greatly improve the estimation results. It was also found that adding range-rate measurements improved the velocity results, but degraded the position estimates. In addition, the extended Kalman filter was determined to give good overall results, while the adaptive extended Kalman filter produced excellent results in certain cases and poor results in other. The results are graphically presented and conclusions are provided concerning use of the filters.