This paper proposes a new control method for the ground moving target tracking problem by a fixed-wing unmanned aerial vehicle (UAV) with a monocular pan-tilt camera. By utilizing the image-based visual servoing, the control law can be directly designed in the image plane, thereby avoiding errors caused by the 3D position calculation. Based on that, we design a control framework to integrate the control of the UAV and the pan-tilt, which enables the UAV to track the target while maintaining the feature point near the image center. Furthermore, considering that the low-cost pan-tilt camera we use has restricted characteristics, we present a deterministic finite automata model to transit the states of tracking when the pan-tilt attitude reaches saturation, thereby improving the tracking ability of the UAV for the moving target. The stability proof of the controller is given, and extensive experiments of hardware-in-the-loop (HIL) simulation and real flights are provided. The results show that the proposed method can achieve continuous robust tracking of the ground moving target.

针对固定翼无人机（UAV）与单目云台相机的地面运动目标跟踪问题，本文提出了一种新的控制方法。利用基于图像的视觉伺服，可以直接在图像平面内设计控制律，从而避免3D位置计算带来的误差。在此基础上，我们设计了一个控制框架，将无人机和云台的控制结合起来，使无人机能够在保持图像中心附近特征点的同时跟踪目标。此外，考虑到我们使用的低成本云台相机的特性受限，我们提出了一个确定性有限自动机模型来在云台姿态达到饱和时转换跟踪状态，从而提高无人机的跟踪能力。移动目标。给出了控制器的稳定性证明，并提供了大量的硬件在环（HIL）仿真和实际飞行实验。结果表明，该方法能够实现对地面运动目标的连续鲁棒跟踪。