CASIA OpenIR  > 毕业生  > 硕士学位论文
面向服务机器人的机械臂自主抓取规划研究
赵雄雷
Subtype硕士
Thesis Advisor谭民
2019-05-16
Degree Grantor中国科学院大学
Place of Conferral中国科学院自动化研究所
Degree Discipline控制工程
Keyword服务机器人 机械臂 自主抓取 路径规划 障碍物搬移规划
Abstract

随着人们生活水平的提高以及人口老龄化的加重,服务机器人逐渐成为机器人领域重要的发展方向。服务机器人为提供更优质的服务,应当具有操作能力,确保在导航至指定操作区域后实施有效的抓取。作为高质量抓取操作的关键因素,机械臂抓取规划的研究具有重要的意义和价值。本文面向服务机器人,围绕机械臂自主抓取规划开展研究,论文的主要内容如下:
首先,介绍了本文的研究背景和研究意义,综述了物体检测与机械臂路径规划的研究现状,给出了本文的主要内容和结构安排。
其次,搭建了服务机器人机械臂实验平台,从硬件和软件系统两方面进行了描述,介绍了机械臂运动学,在此基础上进行手眼标定,求取了相机坐标系到机械臂基坐标系的手眼关系矩阵,通过物体抓取实验进行了验证。
第三,针对服务机器人的机械臂在障碍环境下抓取路径平滑性的问题,提出了一种基于RRT-Connect与Bezier的机械臂路径规划方法。首先进行基于深度学习SSD物体检测和点云信息提取,在此基础上,利用RRT-Connect在机械臂关节空间下搜索一条无碰的路径,并在路径冗余点去除和路径转折点平滑方面,采用二分合并以及Bezier平滑加以改进,最后对各候选Bezier曲线以路径长度为指标,结合环境和自身约束进行优化选取。所提方法的有效性通过仿真和实验进行了验证。
第四,针对环境中可能存在阻碍机械臂直接抓取目标物体的障碍物情形,结合上述基于RRT-Connect与Bezier的机械臂路径规划,提出了一种障碍物搬移规划方法,通过先搬移阻碍障碍再抓取目标物体的方式加以解决。该方法以各障碍物的点云簇到目标物体抓取区点云的距离为依据建立候选待搬移障碍物集,并从中选取到目标物体距离最小的障碍物作为当前待搬移障碍物;而后借鉴人工势场虚拟力思想,构建当前待搬移障碍物引力势场栅格图,对其他障碍物以及目标物体构建斥力势场栅格图,叠加得到合力势场栅格图,确定当前待搬移障碍物的最优放置位置进而实施搬移操作,直至机械臂能够直接抓取目标物体。实验结果表明了所提方法的有效性。
最后,对本文工作进行总结,并指出需要进一步开展的研究工作。
 

Other Abstract

Service robots have become an important research direction in robotics with the improvement of people’s living standard and the aggravation of population aging. In order to provide better services, robots should possess the manipulation ability to guarantee the effective grasping of target object after it arrives at the operation area. As a key factor of high-quality grasping operation, the research on grasping planning of the manipulator is significant. This thesis focuses on the autonomous grasping planning of manipulator for service robot. The contents are as follows:
Firstly, the research background and its significance of this thesis are given. The research development of object detection and path planning is reviewed. The contents and structure of this thesis are also introduced.
Secondly, a manipulator experimental platform for service robot is built. The hardware and software systems are described. The kinematics of the manipulator is given. On this basis, the hand-eye system is calibrated, and the hand-eye relationship matrix of the camera coordinate system to the manipulator base coordinate system is obtained. The manipulator experimental platform is tested by an object grasping experiment. 
Thirdly, aiming at the smoothness problem of the grasping path for the manipulator under obstacles environment, a path planning approach based on RRT-Connect and Bezier is proposed. Object detection based on SSD and point cloud extraction are performed. On this basis, RRT-Connect is adopted to search a collision-free path in the joint space of the manipulator. Then the dichotomy merge and the Bezier curve are employed to remove redundant path nodes and smooth the turning points in the path, respectively. With the constraints of environment and the manipulator itself, an optimized Bezier path is selected from all candidate Bezier paths according to the path length. The effectiveness of the proposed approach is verified by simulations and experiments.
Fourthly, for the situation where there may exist obstacles that directly interfere with the object grasping for the manipulator, a move planning method of interference obstacles is presented with the support of above-mentioned path planning. The problem is solved by first removing the interference obstacles and then grasping the target object. The proposed method establishes a set of candidate interference obstacles based on the distance between the point cloud cluster of each obstacle and that of the target grasping area. The obstacle with the smallest distance to the target object is selected as the current obstacle to be moved. According to the artificial potential field, the grid map of a gravitational field for the current obstacle to be moved is built, and the grid maps of repulsive fields from other obstacles and target object are also constructed. Then a resultant grid map is obtained to determine the placement position of the current obstacle to be moved, after that this interference obstacle is moved by the manipulator. The proposed method is verified by the experiments.
Finally, the conclusions are given and future work is listed.
 

Pages79
Language中文
Document Type学位论文
Identifierhttp://ir.ia.ac.cn/handle/173211/23769
Collection毕业生_硕士学位论文
Recommended Citation
GB/T 7714
赵雄雷. 面向服务机器人的机械臂自主抓取规划研究[D]. 中国科学院自动化研究所. 中国科学院大学,2019.
Files in This Item:
File Name/Size DocType Version Access License
赵雄雷-面向服务机器人的机械臂自主抓取规(7076KB)学位论文 暂不开放CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[赵雄雷]'s Articles
Baidu academic
Similar articles in Baidu academic
[赵雄雷]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[赵雄雷]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.