Robust control of quasi-linear parameter-varying L2 point formation flying with uncertain parameters

• Main Author: Wang, Feng
• Other Authors: Tsourdos, A.
• Published: Cranfield University 2012
• Subjects: 629.13
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.545503

Robust high precision control of spacecraft formation flying is one of the most important techniques required for high-resolution interferometry missions in the complex deep-space environment. The thesis is focussed on the design of an invariant stringent performance controller for the Sun-Earth L2 point formation flying system over a wide range of conditions while maintaining system robust stability in the presence of parametric uncertainties. A Quasi-Linear Parameter-Varying (QLPV) model, generated without approximation from the exact nonlinear model, is developed in this study. With this QLPV form, the model preserves the transparency of linear controller design while reflecting the nonlinearity of the system dynamics. The Polynomial Eigenstructure Assignment (PEA) approach used for Linear Time-Invariant (LTI) and Linear Parameter-Varying (LPV ) models is extended to use the QLPV model to perform a form of dynamic inversion for a broader class of nonlinear systems which guarantees specific system performance. The resulting approach is applied to the formation flying QLPV model to design a PEA controller which ensures that the closed-loop performance is independent of the operating point. Due to variation in system parameters, the performance of most closed-loop systems are subject to model uncertainties. This leads naturally to the need to assess the robust stability of nonlinear and uncertain systems. This thesis presents two approaches to this problem, in the first approach, a polynomial matrix method to analyse the robustness of Multiple-Input and Multiple-Output (MIMO) systems for an intersectingD-region,which can copewith time-invariant uncertain systems is developed. In the second approach, an affine parameterdependent Lyapunov function based Linear Matrix Inequality (LMI) condition is developed to check the robust D-stability of QLPV uncertain systems.