Design of a Hybrid Micro Aerial Vehicle Concept with Multicopter and Vectored Thrust Modes of Flight

Abstract

In recent years, the speed and agility of micro-aerial vehicles (MAVs) have been greatly improved by advances in embedded computing and control theory. Controllers utilizing techniques such as incremental nonlinear dynamic inversion have enabled tracking of aggressive trajectories and high rate cameras and inertial sensors --- combined with compact FPGA and embedded graphics technology --- allow the creation of estimators that can support this aggressive flight. This leap in performance suggests the possibility that autonomous aerial vehicles may be able to compete directly with the best human pilots. Indeed, considering quadrotors specifically, autonomous drone racing competitions have already been organized with the goal of pushing the autonomous quadrotor technology to a level beyond any human pilot. With this goal in mind, and considering the prevalence of the quadrotor as the lingua franca of the field of autonomous MAVs, the performance limits of the traditional brushless outrunner motor quadrotor dynamics may be considered a barrier to the continued development of control theory and embedded computing for fast and agile MAVs.

This thesis seeks to address this limitation by designing a MAV platform prototype --- the rocket-enhanced aerial vehicle with extendable rotors (REAVER) --- which allows for significantly greater acceleration than a quadrotor while maintaining agility and the station-keeping abilities of a quadrotor. REAVER is a hybrid vehicle equipped with both a quadrotor-like mode of flight and a rocket-like mode of flight controlled via thrust vectoring. We discuss the mechanical design of the REAVER prototype, including an initial trade study on propulsion methods, the mechanical design of the vehicle, and its simulated aerodynamic performance. We also evaluate the performance of the design through a trajectory planning study using pseudospectral optimization. Time optimal trajectories which pass through a series of gates are found for the REAVER vehicle dynamics, simulating completing a racecourse. The results are compared to the performance of current autonomous drone airframes in similar tasks. Finally, the development of the jet vane thrust vector control (TVC) system used to steer REAVER during rocket-like flight is discussed. A novel small-scale jet vane design utilizing additive manufacturing is presented and conjugate heat transfer finite element simulation is performed to estimate the jet vane performance. A test stand for verifying the simulations is also presented.