A newly funded project in the School of Aerospace and Mechanical Engineering makes use of close collaboration between researchers in different fields to improve a critical technology for national defense. Dr. Keith Walters and Dr. Andrea L’Afflitto (now a faculty member at Virginia Tech) are combining their respective expertise in aerodynamics and controls to address a difficult challenge for unmanned aerial vehicles (UAVs).
It is well known that UAVs are increasingly being used for both commercial and military applications. The United States Department of Defense (DoD) currently employs multi-rotor helicopters (quadcopters) for remote sensing missions, such as surveillance and search and rescue. In the future, they will support troops by performing tactical tasks, such as picking up and dropping off payloads and surveying cluttered environments. Of particular interest are vehicles that operate autonomously, that is without any direct control by human pilots. These vehicles use onboard computers and mathematical control algorithms to perform necessary aerial maneuvers, travel to desired locations, avoid obstacles, and perform whatever tasks are required of their mission. The development of new and improved control algorithms is, therefore, an active area of research with the potential for substantial impact on next-generation UAVs.
This project focuses on the development of improved control algorithms specifically designed for the landing of UAVs on U.S. Navy ships. Shipboard landing is a complex task for UAVs because 1) the deck is highly unsteady in rough seas; 2) adverse sea conditions are often accompanied by adverse weather and high winds; 3) the superstructure of a moving ship induces a wake in the air, which further perturbs the UAVs landing on its deck; 4) near hard surfaces, the ‘ground effect’ alters the thrust produced by the propellers; and 5) UAVs returning from a mission may be damaged. To land on the deck of a ship, a UAV’s control system regulates the thrust forces of each propeller so that the aircraft approaches the ship with some desired relative velocity and orientation, leading to (hopefully) a gentle touch down in the appropriate location.
The primary objective of this research is to design a robust adaptive control system for multi-rotor UAVs that allows precise landing on the deck of moving ships. The work builds on prior research by former AME faculty member Andrea L’Afflitto and will make use of a model reference adaptive control (MRAC) architecture. Such an approach guarantees robustness of the closed-loop feedback system to both uncertainties in system parameters and unknown state-dependent disturbances that affect the control inputs, such as wind gusts or the swinging of an attached cargo payload.
The control algorithm will also be improved by adopting more realistic functional relationships between propeller rotational speed (RPM) and the generated thrust. Currently, it is assumed that thrust is simply proportional to RPM squared under all conditions. While this is often nearly true when a UAV is hovering in calm air, it does not hold during complex aerial maneuvers, under the action of strong wind disturbances, or when the vehicle is close to a solid surface such as the deck of a ship. Keith Walters and his students will perform computational fluid dynamics (CFD) simulations of quadrotor propellers to more accurately determine the relationship between thrust and RPM under these conditions. The simulations will be used to develop an analytical function that will be included in the control algorithm developed by Dr. L’Afflitto.
The scientific advances made by this project will be disseminated in the technical literature and will provide opportunities for graduate students to participate in national or international conferences. The improvement to UAV performance during shipboard landing will be critical to increasing the value of these vehicles to U.S. Navy missions, and the technology can be translated to other branches of the armed forces to improve design and operation of their next-generation UAV systems. Eventually, the research may be adopted by the commercial sector to improve, for example, the use of UAVs for package delivery or remote sensing in adverse weather conditions.