Optimal reconstruction of ARMA signals with decimated samples under corrupting noise by use of multirate Kalman filter (Q1908850)

An optimal signal reconstruction problem is formulated and treated using multirate space state modeling. The proposed multirate Kalman reconstruction filters noisy, decimated data while producing a good estimate of the complete, clean source. Taking into consideration the analysis of optimal state estimation theory, the multirate Kalman reconstruction filter has achieved the minimum variance estimation and interpolation of noisy and missing ARMA signals. From the investigation of some numerical examples, the multirate Kalman reconstruction filter possesses good estimation/interpolation performances and illustrates significant superiority over a linear interpolator.