This research will focus on the synthesis of controllers for a class of nonlinear systems subject to randomly varying disturbances of large amplitude. The particular class of problems to be considered is that where the plant has a finite number of nominal operating points and makes infrequent transitions between these points. The goal of the work is to generalize existing analytical procedures so that more realistic constraints can be included in the system model without creating computationally intractable design equations. Of particular interest is the inclusion of sensor noise and the weakening of the Markov hypothesis on the primary exogenous disturbance by replacing it with a renewal process structure. It is expected that by extending the class of permissible plant models, it will be possible to improve the initial design thus avoiding ad hoc modifications in later stages of the synthesis process.
