Optimal Estimation with Two Process Models and No Measurements
James M. Maley · 2015
Abstract : An observer is presented for combining the predicted states of 2 independent process models when no measurements are present. The observer follows a derivation similar to that of the discrete time Kalman filter. A simulation example is provided in which a process model based on the dynamics of a ballistic projectile is blended with an inertial navigation system. The results show that under the certain conditions, the algorithm provides estimates of the projectile states with less error than either individual process model.