Worst-case filter design for a class of uncertain dynamical systems

The authors address the problem of designing a filter that guarantees both robust stability and a prescribed worst-case performance in the presence of parameter uncertainties and unknown nonlinear perturbations. The worst-case filtering problem considered is referred to as robust H_∞ filtering. It is shown that the robust H_∞ filtering problem can be reduced to a scaled H_∞-like control problem that incorporates unknown initial state and involves neither parameter uncertainties nor unknown linear functions