|Ph.D Student||Shaferman Vitaly|
|Subject||Cooperative Tracking and Guidance for Autonomous Aerial|
|Department||Department of Aerospace Engineering||Supervisor||Professor Tal Shima|
|Full Thesis text|
In nature, predators and also prey cooperate in hunting and mutual defense to increase their chances of survival. Conversely to the teamwork witnessed in nature, current day autonomous aerial vehicles (AAVs) hardly cooperate or share any information. Cooperation has a vast potential for improving AAVs performance: Cooperative guidance can be used to saturate the target's defenses, limit the target's evasive possibilities, and even lure the target into a trap; cooperative tracking can improve target state estimation, via measurement sharing and different viewpoints at the target; and cooperative task assignment can be used to optimally allocate the team's resources. Inspired by the cooperation witnessed in nature, realizing that such cooperation in current day AAVs is extremely scarce, and comprehending its vast potential, this research explores the possible synergy available through cooperation in three modes of cooperation: in guidance, in tracking, and in task assignment; and also in combinations of these modes. We develop novel guidance laws and task assignment algorithms using optimal control, differential games, and evolutionary algorithm methodologies, which enable AAV teams to employ cooperative strategies to improve performance, overcome weakness, like favorable opponent dynamics, and different constraints, like occlusions, airspace limitations, and intercept angle requirements. Novel cooperative tracking algorithms based on multiple model Kalman and particle filters are also developed to harness the potential of cooperation available via measurement sharing and the different look angles at the target. The algorithms are developed in three representative scenarios: 1) cooperative missile interception, 2) cooperative target protection by a defender missile, and 3) cooperative uninhabited aerial vehicles tracking in an urban environment. These scenarios were chosen to represent a wide range of parameters, like AAV missions, AAV types, team compositions, modes of cooperation, system dynamics, and possible constraints, which facilitates exploration of different solution approaches. The cooperative guidance, tracking, and task assignment algorithms are developed, analyzed, and extensively evaluated analytically and in Monte Carlo simulations, demonstrating substantial improvement over non-cooperative approaches. It is therefore inevitable that in the future cooperation will be an important strategy in AAV systems and that guidance, tracking, and task assignment algorithms similar to those proposed in this research will play an important role in these systems.