61
Adaptive Hierarchical Bayesian Kalman Filtering
Figurę 3. Process Data and Kalman Estimate for Example 1.
are
a- * +
P' = P+ + o2
An example of an application of a Kalman filter based on this model is given in Figurę 4. The process data in Figurę 4 is 100 samples randomly generated according to the above model with o^= 0.032, (Ą = 0.25, anda^ =
1.0. The autocorrelation in the process data is obvious. For the values of the variance components chosen for this example, the gain converges to a steady value of 0.3. Hence, the data does have a continual impact on the State estimate