Absolutely Robust Control Systems

摘要

A new approach to constructing robust feedback control systems for nonlinear nonstationary SISO plants with an incomplete mathematical model is proposed. Based on the Lyapunov function method, the initial system is transformed into a degenerate linear system (reduced system) with components of a model of an indeterminate plant being involved in its right part. The robust control strategy lies in suppressing a plant's model as a potential source of parasite dynamics and reducing the reduced system to a homogeneous equation. The proposed OBS suppressor hyperobserver of a model's dynamics allows one to reduce the effect of various uncertainties to an indefinitely small value with the simultaneous preservation of the desired dynamics of the reduced system. The desired qualitative characteristics of the reduced system are set by choosing the tuning parameters of a PD controller of the (n - 1)th order. The contrastive analysis of the solutions to a benchmark problem using the Matlab/Simulink package leads to a number of conclusions that are of great practical importance.