欢迎访问《兵工学报》官方网站,今天是

兵工学报

• •    下一篇

视场约束下无人机动态降落图像视觉伺服控制

郭涵宇1,宋韬1,叶建川1*(),祝海2,3,周颉鑫2,3,朱效洲2,3   

  1. (1.北京理工大学 宇航学院,北京 100081;2.军事科学院 国防科技创新研究院,北京 100071;3.智能博弈与决策实验室,北京 100071)
  • 收稿日期:2025-02-28 修回日期:2025-05-15
  • 通讯作者: *邮箱:yejianchuan@yeah.net
  • 基金资助:
    国家重点研发计划项目(2023YFC3341100)

Image Visual Servo Control of Unmanned Aerial Vehicle Dynamic Landing under Field of View Constraints

GUO Hanyu1, SONG Tao1, YE Jianchuan1*(), ZHU Hai2,3, ZHOU Jiexin2,3, ZHU Xiaozhou2,3   

  1. (1. School of Aerospace Engineering, Beijing Institute of Technology, Beijing 100081, China; 2. Defense Innovation Institute, Chinese Academy of Military Sciences, Beijing 100071, China;3. Intelligent Game and Decision Laboratory, Beijing 100071, China)
  • Received:2025-02-28 Revised:2025-05-15

摘要: 基于图像视觉伺服(Image-Based Visual Servoing, IBVS)的四旋翼无人机(Unmanned Aerial Vehicle, UAV)系统是一个具有耦合特性的非线性系统,其在降落及跟踪过程中的姿态变化会造成地面目标脱离视场(Field of View, FOV),导致控制失效。为解决此问题,提出一种考虑FOV约束的控制方法,并设计高阶滑模观测器(Higher Order Sliding Mode Observer, HOSMO)估计集总扰动。基于虚拟图像平面和图像矩特征构建系统模型,设计基于IBVS及HOSMO补偿的模型预测控制方法,并分析稳定性。结合FOV约束条件得到姿态动态边界,进而引入FOV约束,并分析约束的有效性。仿真结果表明,所提方法实现了UAV对机动目标的稳定跟踪,高度方向的跟踪误差降低约49%,同时在FOV约束下,实现了UAV对位于FOV边缘的机动目标的稳定跟踪和准确降落,对比现有方法扩展了基于IBVS的跟踪及降落包线。

关键词: 无人机, 基于图像视觉伺服, 虚拟相机, 自主降落, 视场约束

Abstract: The quadrotor unmanned aerial vehicle (UAV) system based on image visual servoing (IBVS) is a nonlinear system with coupling characteristics. Its attitude change during landing and tracking will cause the ground target to leave the field of view (FOV), resulting in control failure. To solve this problem, a control method considering the FOV constraint is proposed, and a higher order sliding mode observer (HOSMO) is designed to estimate the lumped disturbance. The system model is constructed based on the virtual image plane and image moment features, and a model predictive control (MPC) method based on IBVS and HOSMO compensation is designed, and the stability is analyzed. The attitude dynamic boundary is obtained by combining the FOV constraint condition, and the FOV constraint is introduced, and the effectiveness of the constraint is analyzed. The simulation results show that the UAV can stably track the maneuvering target through the proposed method, and the tracking error in the height direction is reduced by about 49%. At the same time, under the constraint of the FOV, the UAV can stably track and accurately land on a maneuvering target that is at the edge of the FOV, and expand the tracking and landing envelope based on IBVS compared with the existing methods.

Key words: unmanned aerial vehicle, image-based visual servoing, virtual camera, autonomous landing, field of view constraints

中图分类号: