In this paper, an innovative cooperative navigation method is proposed for multiple Unmanned Air Vehicles (UAVs) based on online target position measurements. These noisy position measurement signals are used to estimate the target's velocity for non-maneuvering targets or the target's velocity and acceleration for maneuvering targets. The estimator's tracking capability is physically constrained due to the target's kinematic limitations and therefore is potentially improvable by designing a higher performance estimator. An H-infinity filter is implemented to increase the robustness of the estimation accuracy. The performance of the robust estimator is compared to a Kalman filter and the results illustrate more precise estimation of the target's motion in compensating for surrounding noises and disturbances. Furthermore, an adaptive guidance algorithm, based on the seeker's field-of-view and linear region, is used to deliver the pursuer to the maneuvering target. The initial guidance algorithm utilizes the velocity pursuit guidance law because of its insensitivity to target motion; while the terminal guidance algorithm leverages the acceleration estimates (from the H-infinity filter) to augment the proportional navigation guidance law for increased accuracy in engaging maneuvering targets. The main objective of this work is to develop a robust estimator/tracker and an adaptive guidance algorithm which are directly applicable UAVs.

The ability to gather and process information regarding the condition of forest fires is essential to cost-effective, safe, and efficient fire fighting. Advances in sensory and autopilot technology have made miniature unmanned aerial systems (UASs) an important tool in the acquisition of information. This thesis addresses some of the challenges faced when employing UASs for forest-fire perimeter surveillance; namely, perimeter tracking, cooperative perimeter surveillance, and path planning. Solutions to the first two issues are presented and a method for understanding path planning within the context of a forest-fire environment is demonstrated. Both simulation and hardware results are provided for each solution.