Robust Gain-Scheduled Nonlinear Control Design for Stability and Performance

摘要

The models of control systems encountered in many naval applications are nonlinear; moreover, they are also time varying, and have uncertainties affecting them. The underlying controller design problems, beyond requiring system stability, also typically require the optimization of some performance objectives. We propose a numerical solution methodology for solving the general nonlinear controller design problem. The proposed controller architecture is gain scheduled (i.e., the controller uses the measured nonlinearities and time- variations), and optimizes the worst-case performance over the uncertainties of the system. The search for the optimal controller parameters can be reformulated as convex optimization problems involving linear matrix inequalities in several important cases. The design methods are demonstrated on models of Unmanned Combat Air Vehicles (UCAVs). We also address robust estimation problems that underlie many naval applications in control, communications and signal processing areas. Traditional estimation algorithms are based on a nominal system model without uncertainty. However, in many cases, there exist uncertainties in model parameters that may degrade the estimation performance of traditional non-robust algorithms. We present an adaptive robust Kalman filtering algorithm that addresses robustness issues in estimation problems that arise in linear time-varying systems with stochastic parametric uncertainties.