Robust Thrust Vector Control for Precision Rocket-Landing

Abstract

The objective of this thesis is to systematically develop the underlying theory behind and implementation of an integrated framework for analytical multibody dynamics modeling and closed-loop simulations with novel control strategies for the powered-descent and precision landing of rocket-powered space vehicles.

The thesis is organized as follows: Chapter 1 provides an introduction to the rocket-landing problem and the motivation for developing new methods and algorithms to enable future planetary landing missions. Chapter 2 describes the implementation of a globally-optimal minimum-propellant powered-descent guidance (PDG) algorithm using lossless convexification and convex optimization. Chapter 3 explains the analytical formulation of the nonlinear equations of motion for a variable-mass multibody rocket system using the extended Kane’s equations, and shows results from an open-loop simulation run with the optimal control commands obtained from guidance. Chapter 4 describes feedback control in detail, including a novel method for the design of internally stabilizing multivariable robust feedback controllers using Youla parameterization, along with its application to the underactuated lunar landing problem with feedback control only. Chapter 5 provides an algorithm for the design of internally stabilizing robust LPV controllers via Youla parameterization and applies it to the underactuated lunar landing scenario in a combined feedforward-feedback control architecture with propellant-optimal guidance, control allocation, and various actuator considerations. Chapter 6 concludes the thesis with key observations regarding the work done, the results obtained, the specific contributions, and potential directions for future research.