| UAVs are becoming more prevalent in military operations as they are typically small enough to be invisible to radar, cheap enough to be used as sacrificial vehicles, and powerful enough to carry arrays of sensors. This degree of versatility enables UAV's to be used in a wide range of applications including surveillance and resonance, meteorology, crisis management, and locating lost or jeopardized victims. Naval forces currently implement vessel based UAVs to increase their situational awareness. The principal logistical problem for ship-based UAVs is the safe and reliable recovery of the aircraft after mission completion. These aircrafts are most often under actuated, and limited in their ability to quickly change vehicle velocity resulting in UAV systems that are slow to respond to changes in the location and orientation of dynamic landing surfaces such as a ship's deck.;If landing on dynamically moving objects is a mission requirement, the UAV's control system must be able to timely manipulate the vehicles velocity with larger rates of change than the landing surface is capable of. If this is not possible, another method of control must be implemented to ensure that the UAV will not damage itself upon landing.;This work describes the development of algorithm's used to predict the position and location of a landing surface that is moving at constant velocity within two of its translational degrees of freedom(represented by a straight line vector on a surface), and is also experiencing periodic displacements in surface height as well as all of its rotational axes. These requirements have been chosen in an effort to most closely mimic the motions of naval vessels. The frequency, amplitude, and relative phase of the oscillations have been chosen in light of data supplied within [II by Defense Research and Development Canada. In contrast to the work by the Office of Naval Research [2] which is aimed at instructing a hovering vehicle when to land, the findings of this study are expected to be of use within the field of conventional horizontal takeoff and landing vehicles.;Algorithms have been developed and ran in simulations to predict the future state vector (position, velocity and orientation) of periodically moving bodies. Periodic motions were chosen to mimic those of frigate sized vessels within Sea State 4 conditions.;Since military grade UAV's are not easily available for testing purposes, a quadrotor UAV test platform has been designed for algorithm testing. Algorithms have been separately tested in simulations via MATLAB/SIMULINK. This work describes the theory and implementation of these non-platform specific algorithms, as well as the work that has been done to date in the design and construction of a quadrotor test platform involving a quadrotor UAV, a ground control station, and a mobile landing platform. This system is designed to be fully scalable to the computational needs of future control software, as is it based on a model which allows for distributed computation on an array of host computers using UDP and TCP/IP sockets as their primary means of inter process communication. The fully functional quadcopter test platform is scheduled for completion by the end of spring 2011. |
