A typical mission of a UAV is characterized by the requirement of navigating through multiple waypoints as dictated by the mission requirements. These waypoints are usually known a priori, and the mission must be completed satisfying certain criteria such as minimum time, minimum fuel or/and minimum danger exposure as dictated by the mission scenario. Such mission demands a high level of stealth, which has a direct implication on the safety and success of the mission. Assuming the domain of interest is fully characterized a priori, an offline trajectory planning optimization can be used to derive a satisfactory solution. However, in the case of UAVs flying over radar prone environment, UAVs necessarily require more advanced navigation and guidance capabilities. In this thesis we first propose a radar-aware path planning method for UAV in radar zones. This method focuses on trajectory generation algorithms that take into account the stealthiness of autonomous UAV; generating stealth based paths through a region laden with radars. It involves generating a preliminary cost-effective path consisting of a series of straight-line segments. Upon the optimal path being represented as a sequence of straight line segments, we smoothen it by using of a series of B-spline path patterns. For application in UAV path-planning, the smoothened B-spline path patterns are incorporated to replace parts of the path while preserving continuity to obtain a feasible path. To this end, an optimum path traceable by the UAV with radar-aware capability is generated. It has been demonstrated that the UAV manages to accomplish the mission with minimal radar exposure. For cooperative control of multiple UAVs, we developed an efficient evolutionary framework to optimize their operations in radar zone. The proposed approach is capable of solving cooperative mission-planning problem for multiple UAVs. Taking advantage of memetic intelligence, the proposed planner is competent of handling large-scale optimization problems for multiple UAVs. To achieve robustness, a cooperating individual learning strategy is proposed in which local search strategies or memes, each having different learning roles and search features, work together to accomplish the shared optimization goal. Next, our work focuses on aspect of optimizing a multi-criteria mission-planning problem for multiple UAVs, catering to the requirement of different mission constraints. The objective is to compute an optimum solution for each UAV to complete its mission, taking into account the pre-specified conditions. We demonstrate the effectiveness of the proposed approach by yielding a broad range of Pareto optimum solutions. This approach could be easily extended for different scenarios by appending different terms in objective function and constraints. Lastly, the proposed framework is experimentally validated through a flight simulation environment, showing that the applicability of our proposed approach on the actual system. DOCTOR OF PHILOSOPHY (EEE)
