擬線形状態推定についての一考察

概要

論文の詳細を見る
The optimum estimate of a signal (state vector) is conventionally defined by the conditional expectation concerned with the observed signals (received signals). It is assumed that the observed signals are the nonlinear vector function of two signals (state vectors) and observed noises, and that each of two signals is taken for the solution of the stochastic nonlinear differential equations with gaussian white noise inputs. In the nonlinear systems, the optimum estimate is physically unrealizable because it requires the information about the infinite dimensional statistical moments. Therefore, quasi-linearization methods are applied to the above nonlinear systems. As the optimum estimate for one of two signals is obtained, it is demonstrated that such an optimum estimate is the solution of the simultaneous differential equations.
明治大学の論文