Robust Control of Nonlinear Systems Using Model Error Control Synthesis

The proposed work aims at developing a comprehensive theory for the robust control of both linear and nonlinear systems. A new approach will be investigated that uses a real-time nonlinear estimator to determine model-error corrections to the control input. The estimator determines the model error using a one time-step ahead approach. Also, the estimator can be used to determine state quantities. Control compensation is achieved by using the estimated model error as a signal synthesis adaptive correction to the nominal control input so that maximum performance is achieved in the face of significant model uncertainty and disturbance inputs. A significant advantage of this approach over other adaptive methods is that model parameters need not be updated online. Instead, the effect of these errors is used to update the actual control signal, which leads to a simple design strategy. The novel part of the control approach is that it combines a nonlinear estimator and a nonlinear controller in a cohesive approach to provide system robustness. The design procedure is straightforward, unlike other methods such as standard adaptive control methods that may involve extensive design procedures to guarantee performance specifications. The impact of this proposed work will be the development of a methodology that is at once of basic theoretical value and will enable a large class of system applications in engineering applied science. Additionally, Prof. Crassidis will be engaged in both international collaborations and workshops for dissemination of this research.