垂直起降无人机基于图像的目标跟踪控制  

Image-based target tracking control for vertical take-off and landing UAVs

作　　者：刘锦涛[1] 高丽[1] 吴文海[1] 李静[2] LIU Jin-tao;GAO Li;WU Wen-hai;LI Jing(Naval Aeronautical Engineering Institute, Qingdao Branch, Qingdao Shandong 266041, China;Naval Aeronautical Engineering Institute, Department of Strategic Missile Engineering, Yantai Shangdong 264200, China)

机构地区：[1]海军航空工程学院青岛校区,山东青岛266041 [2]海军航空工程学院战略导弹系,山东烟台264200

出　　处：《控制理论与应用》2017年第6期723-731,共9页Control Theory & Applications

基　　金：国家自然科学基金项目(11461075)资助~~

摘　　要：针对安装有惯性测量单元和摄像机的低成本四旋翼无人机,研究无位置、速度、航向测量情况下的机动目标基于图像的跟踪控制方法.首先,结合无人机的动力学方程在图像空间中推导了系统的误差方程.其次,为克服无航向测量的问题,设计了一种位置控制器,使用图像矩作为反馈输入并输出油门和姿态指令.最后,针对缺少图像速度测量问题,设计了一种super-twisting滑模观测器和控制器,生成的期望姿态和拉力指令无颤振,并通过李雅普诺夫理论证明了控制系统的稳定性.最终无人机通过调整倾斜姿态实现了跟踪飞行,且避免了响应慢的航向调整.跟踪机动目标的仿真结果验证了所提出方法的有效性.The low-cost quadrotor is equiped with inertial measurement unit(IMU)and camera,but there is no position,velocity or heading measurement.For this problem,an image-based control method is studied to track the maneuvering target.Firstly,the system error equations are proposed in the image space,which include the motion equations of quadrotor.Secondly,a position controller is designed,which takes image moments as inputs and attitude and thrust commands as outputs.Finally,for the problem of no measurement of image speed,a super-twisting sliding mode observer and controller are designed.The controller is robust to the unknown target maneuver and its output is chattering-free.The system stability is proven by Lyapunov theory.The unmanned aerial vehicle(UAV)completes the target tracking flight by fast tilt attitude change,not slow yaw change.The simulation results demonstrate the efficiency of proposed control scheme for quadrotor target tracking.

关 键 词：四旋翼无人机 跟踪控制 基于图像视觉伺服 滑模控制 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] V249[自动化与计算机技术—计算机科学与技术] V279[航空宇航科学与技术—飞行器设计]