Scalable Autonomous Operations of Unmanned Assets

Although there have been great theoretical advances in the region of Unmanned Aerial Vehicle (UAV) autonomy, applications of those theories into real world are still hesitated due to unexpected disturbances. Most of UAVs which are currently used are mainly, strictly speaking, Remotely Piloted Vehicles (RPA) since most works related with the flight control, sensor data analysis, and decision makings are done by human operators. To increase the degree of autonomy, many researches are focused on developing Unmanned Autonomous Aerial Vehicle (UAAV) which can takeoff, fly to the interested area by avoiding unexpected obstacles, perform various missions with decision makings, come back to the base station, and land on by itself without any human operators. To improve the performance of UAVs, the accuracies of position and orientation sensors are enhanced by integrating a Unmanned Ground Vehicle (UGV) or a solar compass to a UAV; Position sensor accuracy of a GPS sensor on a UAV is improved by referencing the position of a UGV which is calculated by using three GPS sensors and Weighted Centroid Localization (WCL) method; Orientation sensor accuracy is improved as well by using Three Pixel Theorem (TPT) and integrating a solar compass which composed of nine light sensors to a magnetic compass. Also, improved health management of a UAV is fulfilled by developing a wireless autonomous charging station which uses four pairs of transmitter and receiver magnetic loops with four robotic arms. For the software aspect, I also analyze the error propagation of the proposed mission planning hierarchy to achieve the safest size of the buffer zone. In addition, among seven future research areas regarding UAV, this paper mainly focuses on developing algorithms of path planning, trajectory generation, and cooperative tactics for the operations of multiple UAVs using GA based multiple Traveling Salesman Problem (mTSP) which is solved by dividing into $m$ number of Traveling Salesman Problems (TSP) using two region division methods such as Uniform Region Division (URD) and K-means Voronoi Region Division (KVRD). The topic of the maximum fuel efficiency is also dealt to ensure the minimum amount fuel consumption to perform surveillance on a given region using multiple UAVs. Last but not least, I present an application example of cattle roundup with two UAVs and two animals using the feedback linearization controller