${H} _\infty $ Based Estimation of Nonlinear Systems

This letter presents a H
_∞
criterion-based recursive filter of nonlinear systems with nonlinear dynamics and measurement. The filter is based on the deterministic-statistical approach to estimation. The solution is based on three components: 1) the mean square error criterion; 2) the state dependent coefficients form; and 3) calculus of variations. The computation of the estimator's gains needs the solution of a non-symmetric differential Riccati matrix equation. A simulation example is presented.