Smoothly Constrained Extended Kalman Filter

For nonlinear state estimate, the EKF is an effective technique the base point error should be noted. The paper proposes a smoothly constrained extended Kalman filter for the nonlinear Gaussian mode. The statistical noise of the system is modeled as a constrained mathematical model, the objective function is established based on the maximum posterior probability estimate criterion, the feasible area is approximated by numerical methods, wherein the feasible set is selected and propagated. The calculation of the high-order Hessian matrix and the irreversible ill-conditioned matrix that may occur during the state update process can be avoided, and the calculate stability is enhanced. Simulation experiments verify the effectiveness of the proposed algorithm.