This paper demonstrates that the H∞ controller of Doyle .  can be reinterpreted as an observer-embedded L2-gain controller. The embedded observer is similar to the conventional Luenberger observer except having an additional state-dependent calibration term that compensates for exogenous inputs and model uncertainties. Based on this fact, we derive mixed Linear Matrix Inequalities, Hamiltonian Matrix, and Linear Parameterization to provide solutions to feasible observer-embedded L2-gain controllers.
All Science Journal Classification (ASJC) codes
- Applied Mathematics