The purpose of this dissertation is to develop a guidance and control algorithms for stable and automatic operation of a fixed-wing unmanned aerial vehicle on a moving aircraft carrier. So, this paper investigates that the sea state has analyzed conditions for suitable operating a UAV. The sea state is divided into nine levels, and each level determines the significant wave height and sustained wind speed. This paper has considered ship motion in sea state level five. This sea state level five account for about 80% of the total sea environment. In this conditions, the movement of the ship is simulated by using the MSS hydro toolbox. It is assumed that the ship moves at the 15 m/s in a forward direction velocity under the wave incident angle of 30 degrees, using the S-175 ship model. In these conditions, the landing point is constantly changing due to the ship’s heave and pitching motions. Also, the position of the landing point is offset from the center of gravity of the ship. Therefore, it should be calculated considering the center of gravity of the ship and the offset of the landing point. It is applied to the simulation considering the precondition. The simulation is performed using the MQ-9(reaper) class UAV model. This UAV is modeled based on the X-plane simulator MQ-9 model. At this time, aerodynamic coefficient data are obtained using DATCOM. Using aerodynamic data and geometry parameters, a 6-DOF nonlinear model is constructed and using this model make a linear model. The linearization model is used for the design of the linear controller and the dynamic inversion controller.
The guidance and control are conducted based on the movement of the current landing point to the continuously moving ship. However, the probability of landing failure increases if the abrupt movement occurs of the ship just before landing. Therefore, prediction of the landing point is proposed for UAV guidance in order to cope flexibly with the sudden ship motions using NARX. The NARX learns the input and output relationship of time series data, which performs the prediction of ship motions.
The guidance and controller are designed to land at the predicted point. As the inner loop control law, the LQTI linear optimal controller and neural network adaptive controller are designed. The LQTI controller is based on the LQR controller with a tracker term and an integrator term added. The neural network controller based on dynamic inversion with the adaptive signal compensates for the inversion error. The guidance law is designed for each axis on the longitudinal and lateral axis. The longitudinal guidance law applies the line-of-sight and the glideslope guidance. Also, lateral guidance law applies the nonlinear path following guidance and the track guidance. Comparative simulations are performed by applying the designed various guidance laws and controllers.
The integrated simulation environment considers wind condition, ocean environment (sea state), ship dynamic motions, prediction of the landing point, guidance laws and controllers. A total of thirty-two conditions is compared. Each case consists of thirty times of Monte-Carlo simulations with changing the initial conditions. From the simulation results, the landing success rate and landing point accuracy are analyzed. When the prediction guidance is applied, simulation results manifest that the landing success rate and landing accuracy are increased.