Boost-Phase Filtering Options: Is Simpler Better?

Abstract

It is well known that a template-based Kalman filter with correct a priori information can outperform a linear polynomial Kalman filter that does not require such information. It is also well understood that a template-based Kalman filter with slight errors in the template will require additional process noise to prevent filter divergence. For the boost-phase filtering problem, this paper will show that when template information is in error, a polynomial filter will outperform a template-based Kalman filter.