Game theory approach to optimal linear state estimation and its relation to the minimum H/sub infinity /-norm estimation

摘要

A possible game theory approach to optimal state estimation is presented. It is found that in a certain differential game, the minimizer's policy is identical to the one obtained by optimal estimation in the minimum H/sub infinity /-norm sense. This interpretation of H/sub infinity /-optimal state estimation provides better insight into the mechanism of H/sub infinity /-optimal filtering, especially in the case where the exogenous signals are not energy bounded.