Comparative study of several nonlinear stochastic estimators

In this paper we investigate the relative performance and design procedures of several nonlinear stochastic estimators. The filters that we are comparing are: Lyapunov-based, covariance assignment, extended Kalman filter, and state-dependent Riccati equation estimator. First we provide an overview of these estimators and then we will compare their performance using first- and second-order nonlinear stochastic systems. The discussion will include convergence property, difficulty level of the design, computational time, and overall performance, based on absolute error and mean square error of the estimation.