Tracking procedure for non-normally distributed measurement errors.
Alexander. Kukliansky · Calhoun: The Naval Postgraduate School Institutional Archive (Naval Postgraduate School) · 1987
The Kalman Filter is a widely used procedure in tracking algorithms. When normality assumptions are violated, the Kalman Filter performance tends to degrade. In this thesis a new procedure is introduced for accommodating non-normal properties of measurement error distributions. The procedure is developed for the multi-observer situation. Simulation experiment results are presented and numerical comparisons are made between the Kalman Filter performance and that of the new procedure.