Second-Order Correction and Numerical Considerations for the Two-Step Optimal Estimator

摘要: A modie cation of the two-step optimal e lter is presented. The an alternative to standard recursive estimators that are applied nonlinear measurement problems, such as extended and iterated Kalman lters. It improves estimate error by splitting cost function minimization into two steps (a linear rst step a nonlinearsecond )by dee ning set rst-step states combinations desired states. approximation made in time update rather than conventional methods. An extension including higher-order termsin thestateestimateerrorispresented. Previous work used rst-orderexpansion relating rst- second-step nd Terms third order retained here resulting keeping second-order corrections both state covariance. result with lower bias mean square error. root implementation algorithm also presented robustness accuracy lter. Performance verie ed using radar ranging example.