| With US Armed Forces' goal to achieve the fielding of unmanned, remotely controlled technology such that one-third of the aircraft in the operational deep strike force aircraft fleet are unmanned long overdue, the thesis investigates the major efforts and challenges still being faced by the researchers in the various disciplines involved in the multi-disciplinary task of the unmanned aerial vehicle. It discusses the disciplines concerned, their research efforts and some lingering challenges for researchers interested in the subject.;It also investigates the possibility of having a UAV without any man in the loop (autonomous UAV). This may serve as back-up strategy for the UAV control in case there is a loss in communication between the remote pilot at the base station and the aircraft putting into cognizance the high rate at which the UAV are lost or damaged. We introduce an intelligent agent that is trained to learn the flight dynamics and performs the role which the remote pilot at the control station would have done. Like the remote pilot which determines and communicates wirelessly through sensors to the UAV the necessary reference input required to obtain a particular output, the intelligent agent mimicking the action of the human pilot determines the magnitude of the required reference input that needs to be given to the UAV controller depending on the sensor data of the prevailing condition.;Considering the multi-disciplinary nature and complexity of such a project as well as the additional challenge introduced by the removal of man from the loop, this thesis limits its study to the control part of the project using a PID controller. The designed ANN-PID Controller yielded desirable transient and steady-state responses on simulation. |
