基于视觉伺服的飞行机械臂抓取控制

发布时间：2018-04-24 00:38

本文选题：无人机抓取 + 轨迹规划与跟踪　；参考：《哈尔滨工业大学》2017年硕士论文

【摘要】：最近几年,随着旋翼无人机技术的成熟,无人机凭借其操作简单、运动灵活和稳定可靠的特性在很多领域大展身手,在农业、军事、气象、灾难救援等方面获得了很多的成功运用。但是,目前旋翼无人机的运用大多数是通过航拍这种方式来进行环境的监控,而不能和外部环境进行有效的互动。如果在多旋翼无人机上安装灵活的机械臂,那么无人机就可以实现外部环境物体的抓取和运输。现有的无人机抓取研究需要借助昂贵、覆盖范围有限的运动捕捉系统等外部设备实时获得无人机和目标抓取物的位置信息,或者只是基于视觉实现缓慢的飞行抓取,抓取过程冗长,也不适合大范围应用。因此,本课题的主要研究任务就是实现无运动捕捉系统的无人机高速自主物体抓取。本课题在一个旋翼无人机上搭载了一个可旋转的摄像头和一个轻质的机械臂。通过欧拉-拉格朗日方程,在笛卡尔空间对无人机-多自由度机械臂复合系统建立完整的三维动力学模型。不失一般性,本文将动力学模型简化至x-z平面,机械臂简化至单自由度,通过对无人机-相机系统进行几何分析,在x-z平面建立了目标图像位置和无人机空间位置的映射关系。针对无人机的姿态与目标在图像中的位置耦合问题,本文将图像投影至标定过的虚拟水平图像上来解耦无人机及相机姿态,然后将x-z平面的动力学模型映射到图像空间,并证明图像空间系统的微分平滑性。基于图像空间映射模型,本文分别提出了使用最小化Snap和迭代线性二次调节器(i LQR)两种方法的图像空间轨迹规划器,它们都能够将飞行过程中的各种耦合约束条件转化成最优化问题。为了跟踪轨迹,本文设计了视觉伺服控制器,反馈误差是目标在图像的像素误差,设计内外环PID控制器实现跟踪。为验证所提出的算法,本文对基于Gazebo的ROS无人机仿真系统进行扩展,建立了无人机-机械臂系统仿真平台,利用该平台验证了模型、轨迹规划器及跟踪控制器的有效性及鲁棒性。本论文是对无人机机械臂抓取这一新方向的探索与研究,实现了基于视觉伺服的无人机抓取轨迹规划和跟踪控制,对于该方向的其他研究人员是重要的参考,对于我国无人机运输的应用研究有较大意义。
[Abstract]:In recent years, with the maturation of the rotor unmanned aerial vehicle technology, the UAV has achieved many successful applications in many fields because of its simple operation, flexible movement and stable and reliable characteristics. However, most of the use of unmanned aerial vehicles (UAV) is by aerial photography. Environmental monitoring can not interact effectively with the external environment. If a flexible manipulator is installed on a multi rotor UAV, the unmanned aerial vehicle can be captured and transported to the external environment. The existing UAV grabbing research needs to be expensive, covering the limited motion capture system and other external equipment in real time. Obtaining the location information of the unmanned aerial vehicle and the target grabbing, or only the slow flight grasping based on the visual realization, the grasping process is long and not suitable for large scale applications. Therefore, the main research task of this topic is to realize the high speed self master body of the unmanned aerial vehicle without the motion capture system. A revolving camera and a light mechanical arm. Through the Euler Lagrange equation, a complete three-dimensional dynamic model of the UAV multi degree of freedom mechanical arm composite system is established in Cartesian space. Without losing the general nature, this paper simplifies the dynamic model to the x-z plane, the manipulator is simplified to the single degree of freedom, and through the unmanned aerial vehicle (UAV) Based on the geometric analysis of the camera system, the mapping relationship between the target image position and the unmanned aerial vehicle (UAV) position is established in the x-z plane. In this paper, the image is projected to the calibrated virtual horizontal image to decouple the UAV and the camera attitude, and then the dynamics of the x-z plane is then carried out. The model is mapped to the image space and proves the differential smoothness of the image space system. Based on the image space mapping model, the image space trajectory planners using the two methods of minimizing Snap and iterative linear two order regulator (I LQR) are respectively proposed. All of them can transform the various coupling constraints in the flight process to the most. In order to track the trajectory, this paper designs a visual servo controller. The feedback error is the pixel error of the target in the image, and the internal and external loop PID controller is designed to realize the tracking. In order to verify the proposed algorithm, this paper extends the ROS simulation system based on the Gazebo, and establishes the simulation platform of the UAV mechanical arm system. The platform validates the validity and robustness of the model, the trajectory planner and the tracking controller. This paper is the exploration and study of the new direction of the unmanned aerial vehicle (UAV) manipulator, which has realized the trajectory planning and tracking control of the unmanned aerial vehicle based on the visual servo. It is an important reference for other researchers in this direction, and it is not available to our country. The application of man-machine transportation is of great significance.

【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TP241

【相似文献】