EQUIVALENCE OF FINITE MEMORY FILTERS

The well-known Jazwinski's limited memory filter, Schweppe's finite memory filter, and Kwon's optimal finite impulse response (FIR) filter are compared in the filter structures, system models, and optimality criterions, and are shown to be equivalent on condition that they are applied to the discrete system with no process noise and unknown prior information of the system state. The different properties such as stability and computation burden are briefly discussed. Kwon's optimal FIR filter is shown to have some advantages in terms of stability and modeling constraints.