Controle preditivo não linear aplicado ao Sistema de Piloto Automático de mísseis

Abstract

This dissertation addresses the development and application of Model-Based Predictive Control (MPC) in missile autopilot systems. Detailed mathematical modeling of axisym- metric missiles controlled by fins is presented, including practical simplifications that maintain the necessary accuracy for effective control across the flight envelope. Based on this modeling, a Nonlinear MPC controller is developed and tested through simulations. Its performance is compared with another controller, which utilizes an advanced robust control method where the plant’s nonlinearities are modeled as uncertainties and described by Integral Quadratic Constraints (IQC). Thus, the viability and effectiveness of MPC in complex and nonlinear applications such as missile autopilot control are demonstrated. The simplicity, intuitiveness, and accuracy of MPC are emphasized, especially when applied to systems with well-defined mathematical modeling. The performance of the MPC controller proves to be superior in comparison to the IQC controller in nominal simulations, showing robust stability to parametric variations and the capability for real-time implementation, given the rapid optimization solutions achieved. Significant contributions include the formulation of an approximate equation for dynamic pressure in compressible fluids, for different Mach numbers and altitudes, and the proposition of an MPC controller with two internal models, one for prediction and another for state updating. Identified limitations pave the way for future research, suggesting the inclusion of robust control techniques in MPC, the application of more realistic missile aerodynamic data, and the consideration of full 6 Degrees of Freedom (6-DOF) dynamics. This work not only advances knowledge in the field of Nominal MPC control but also establishes a solid foundation for future research in the area of Nonlinear and Robust MPC control systems.