LQR Control for a Mono-Propelled VTOL Drone

As autonomous landing rockets become more common in industry, Florida Rocket Lab needs a testbed to develop the guidance, navigation, and controls (GNC) for their propulsive landing vehicle ['Hopper']. A mono-propelled vertical take-off and landing (VTOL) drone ['Skipper'] is an optimal choice, as it is relatively safe, cost-effective, and mirrors the dynamics of a rocket closely. This paper first covers the derivation of the equations of motion for a mono-propelled VTOL drone. Using these equations, a linear state space is then constructed and controlled using a linear-quadratic regulator. Due to the highly nonlinear nature of the system, a gain-scheduling program is then implemented based on the error between the linear and nonlinear states. Finally, a Kalman filter is developed to account for sensor noise. The combined control system is then implemented in MATLAB/Simulink and used to control the drone in a variety of environments. This paper aims to guide other student teams in the development of similar projects.