Automated Aerial Refueling ( AAR) is one of the necessary capabilities for future Unmanned Aerial Vehicles ( UAV). A simulation system of UAV aerial refueling was established, which contained the models of UAV and refueling plane, vision sensor, aerodynamic effect and scene system. The feature point recognition and pose estimation algorithm of machine vision were studied. The design of smooth docking maneuver under the presence of wake effects and turbulence was accomplished. Desirable performances were achieved with trajectory generator and tracking controller at the last stage of refueling. The simulation results were presented, which showed that the monocular vision can obtain precise relative pose and the controller can guarantee stabilize docking of AAR under disturbance.%自动空中加油是未来无人机(UAV)需具备的重要能力之一.建立了包括UAV和加油机模型、视觉传感器、气动影响及视景的飞行硬管式无人机自动空中加油仿真系统；研究了单目视觉特征点识别与位姿估计算法,考虑大气扰动及加油机尾流的影响；设计了通过轨迹跟踪实现UAV与加油管会合对接的控制策略,并对加油末段的对接机动进行了仿真.结果表明,单目视觉可实现精确的相对位置测量,轨迹跟踪控制器超调量小,保证了自动空中加油的平稳对接.
展开▼
