A new finite dimensional approximation filter is derived for nonlinear systems. For the finite dimensional approximation, a statistical linearization method is employed. The proposed filter is compared with other statistical linearization filter and the extended Kalman filter using an example.
