: Unmanned aerial vehicles (UAV) received a lot of attention in recent years, considering latest advances in technology which boosted the possibility of vast military and commercial applications for such aircrafts. The possibility to build smaller (and more efficient) UAVs improved research interest in this field. One popular application in recent years is cooperative flight (i.e. flying of a team of UAVs). In cooperative mission, objectives and means can be assigned to multiple UAVs - instead of one UAV - and thus makes mission designing possible in situations that one UAV cannot handle the whole mission. Among the important issues in every UAV mission is trajectory generation. This trajectory should be flyable first (i.e. it should consider UAV’s maneuver limits), and second, it should not pass through obstacles. In addition, UAVs are prone to numerous uncertainties and faults during their flight course. Of the most common aircraft faults are lock-in-place fault (as an actuator fault) and icing fault (as a structural fault). These faults share similar symptoms that make theirisolation difficult. Moreover, in the case of icing on the aircraft wings and control surfaces, system states are subject to change and in extreme cases (i.e. full icing) UAV may even lose controllability which leads to crash. Actions taken to prevent icing are generally costly, heavy and require huge power consumption and are not implementable on small aircrafts. Therefore, it seems crucial to design a method to prevent crash in case of an occurrence of icing fault. In this thesis,we discuss fault detectionand isolation in UAVs cooperative flights. For this matter, we first discuss trajectory generation using Pythagorean Hodograph (PH)algorithm. This algorithm can be implemented easily and the designed route is flyable. In our thesis, we present a new trajectory generation approach through which not only the route is carried out with a suitable and suboptimal result, but also it will reduce the design time to less than a quarter comparing to regular methods. Furthermore, a controller is designed to track the route in normal and faulty situations. The controller is robust and is designed based on theorem, using McFarlane-Glover loop shaping algorithm. Finally using one Unknown Input Observer (UIO), a fault detectionand isolation scheme is proposed to detect icing faults. Our proposed method enables the observer to be sensitive against both structural faults and actuator faults, and can isolate icing faults from lock-in-place. Simulation results confirm the success of our design. Finally, the performance of this fault detection system is examined through a cooperative simultaneous arrivalmission. Keywords: Unmanned Aerial Vehicle, Trajectory Generation, Fault Detection and Isolation, Pythagorean Hodograph, Robust Controller.
