Adaptive-Gain Tracking Filters Based on Minimization of the Innovation Variance
Naum Chernoguz · 2006
A kinematic tracking filter is considered in the context of gain adaptation problem. The study suggests a simple adaptive-gain tracker based on minimization of the innovation variance. This is shown to provide the optimal Kalman gain. Accordingly, the innovation-based adaptive Kalman-like filter is constructed. The adaptive scheme is associated with a recursive MA-parameter estimator. With proper links for the optimal gain-vector components, the multiple-parameter adaptive filter reduces to a constrained single-parameter version. The simulation study justifies the filter performance for a wide range of conditions