基于视觉的双臂机器人协调运动研究

发布时间：2018-01-26 17:04

本文关键词： NAO 机器人双臂机器人协作 D-H 图像处理视觉伺服　出处：《青岛科技大学》2017年硕士论文　论文类型：学位论文

【摘要】：双臂机器人在多变环境以及复杂任务处理等情形中具有很强的适用性和可靠性,因此双臂机器人在工业领域、服务领域有很好的应用前景。双臂机器人的协调运动是机器人系统的关键技术之一,也是机器人领域的研究热点。本文调查了国内外双臂机器人及协调运动的研究现状后,以NAO机器人作为实验平台对基于视觉的双臂机器人协调运动进行了研究,主要研究内容如下。首先介绍了NAO机器人的软件和硬件参数,并对本文所涉及的头部、左臂、右臂三个运动链进行了DH参数建模,同时分析了各运动链的正运动学过程并给出了左臂和右臂的逆运动学解析解及推导过程。其次,分析了视觉系统的数学模型,实现了机器人相机参数的标定,并分析了手眼标定的数学原理,针对运动时NAO机器人本体不稳定的情况对每个相机使用大量运动状态采集并结合最优化方法完成了手眼标定。然后,针对如装配等松协调环境,提出使用基于图像的视觉伺服方法来实现协调运动,同时提出了一种Mark标记点的识别定位方法,并对协调过程进行了实验和仿真。结果证明了该协调运动方法的可行性。最后针对NAO机器人独特的视觉系统和连杆结构,提出由两个NAO机器人组成的紧协调方案来模拟一特定任务。先利用两NAO机器人头部下相机的公共视野及工件上标定图案来相互定位,再利用头部上相机来识别棍状工具,在工件坐标系内进行三维直线重建,同时提出了一种具有较高适应性使用改进的骨骼细化方法的直线识别方法。针对该任务工具直线与工件平面的垂直约束提出了一种规划方法,并针对NAO机械臂仅有5个自由度的情况,对不可达的空间位姿通过放弃绕末端上一轴转动的数值求解策略实现了多机械臂的协调运动。并用两个NAO机器人和MATLAB进行了实验及仿真,验证了该方案的可行性。
[Abstract]:The dual-arm robot has strong applicability and reliability in changeable environment and complex task processing, so dual-arm robot is in the field of industry. The coordinated movement of dual-arm robot is one of the key technologies of robot system. It is also a research hotspot in the field of robot. This paper investigates the research status of dual-arm robot and coordinated motion at home and abroad. The coordinated motion of dual-arm robot based on vision is studied with NAO robot as the experimental platform. The main contents are as follows. Firstly, the software and hardware parameters of NAO robot are introduced. The DH parameters of the head, left arm and right arm are modeled. At the same time, the forward kinematics process of each kinematic chain is analyzed, and the inverse kinematics analytical solution and derivation process of the left arm and right arm are given. Secondly, the mathematical model of the vision system is analyzed, and the calibration of the camera parameters of the robot is realized. The mathematical principle of hand-eye calibration is analyzed. Aiming at the instability of NAO robot in motion, a large number of motion state acquisition and optimization method are used to complete the hand-eye calibration for each camera. Then. Aiming at loose coordination environment such as assembly, a visual servo method based on image is proposed to realize coordinated motion, and a method of recognition and localization of Mark marker points is proposed. Experiments and simulations are carried out on the coordination process. The results show the feasibility of the coordinated motion method. Finally, the unique vision system and connecting rod structure of the NAO robot are discussed. A compact coordination scheme consisting of two NAO robots is proposed to simulate a specific task. Firstly, the common field of view of the camera under the head of the two NAO robots and the calibration pattern on the workpiece are used to locate each other. Then using the camera on the head to identify the stick shape tool, the three-dimensional straight line reconstruction is carried out in the workpiece coordinate system. At the same time, a linear recognition method with high adaptability to use the improved bone refinement method is proposed, and a planning method for the vertical constraints between the straight line and the workpiece plane of the task tool is proposed. And aiming at the NAO manipulator with only 5 degrees of freedom. For the unreachable spatial posture, the coordinated motion of the multi-manipulator is realized by giving up the numerical solution strategy of turning around the upper axis of the end, and two NAO robots and MATLAB are used to carry out experiments and simulations. The feasibility of the scheme is verified.
【学位授予单位】：青岛科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242

【参考文献】