Target tracking and source seeking using unmanned aerial vehicles with input constraints

Abstract

The potential applications of unmanned aerial vehicles (UAVs) in both military and civilian fields have spurred the research on UAV control. Although great progresses in this area have been made by researchers worldwide, practical and theoretical challenges are still abundant. This thesis focuses on the problems of target tracking and source seeking using UAVs subject to input constraints. Ground target tracking is an important application of UAVs. Due to UAV inherent motion constraints, the fixed-wing UAV generally can not stay on top of the ground target. Standoff tracking is one possible solution to target tracking. In the first part of the thesis, a single UAV is used to achieve moving ground target tracking. A guidance law is introduced to generate a desired relative course for the UAV. Based on the desired relative course, a saturated heading rate controller is proposed to achieve standoff target tracking. A rigorous global stability proof of the system with the proposed controller is provided. As the ground target velocity may be unknown, a bounded adaptive observer is designed to estimate the unknown constant target velocity and background wind. Ground target could be friendly or adversarial. For adversarial target tracking or hazardous target tracking, safety and tracking performance are two critical issues for the UAV. In order to achieve adversarial target tracking, the initial condition of the UAV is analyzed, and two types of tracking scenarios are defined: exposure avoidance and minimum exposure time. A bang-bang controller is proposed to achieve target tracking in both scenarios. Advances in sensing, communication and control systems make it possible for multiple UAVs to carry out missions cooperatively. A circular formation strategy for target tracking is proposed in this thesis. Firstly, a guidance law based heading rate controller is designed for circular tracking. A notion called temporal phase is introduced for multi-UAV cooperation. Based on the temporal separation, a bounded variable airspeed controller is developed to achieve equal temporal separation. Source seeking is a potential application of multi-UAV cooperation. In the past decades, single vehicle is successfully used for source seeking. However, the time cost and extreme movement are two challenging issues to be addressed. Cooperation of multiple UAVs makes it possible to attain better source seeking performance. This thesis proposes a leader-follower formation strategy for source seeking. Based on the measurement of all UAVs, the scalar field gradient at the leader location is estimated using least-squares method. After that, an adaptive observer is designed to estimate the source velocity. By using the estimated gradient and source velocity, a heading rate controller for the leader is developed to…