CASIA OpenIR  > 毕业生  > 博士学位论文
移动机械臂视觉伺服关键技术研究
王士军
学位类型工学博士
导师原魁
2017-05-23
学位授予单位中国科学院大学
学位授予地点北京
关键词移动机械臂 视觉伺服 轨迹规划 双轨迹控制
其他摘要
融合了移动机器人可移动性和机械臂可操作性的移动机械臂,比非操作型的移动机器人或工作范围受限的机械臂拥有更多的自由度和更广的操作空间,是近年来机器人领域的研究热点。移动机械臂并不是移动平台和机械臂两个子系统的简单叠加,因此,在解决移动机械臂的视觉伺服问题时,必须考虑系统具体的运动和控制特点。本文针对移动机械臂的运动规划方法和视觉伺服控制问题进行了深入的研究,论文的主要工作可总结如下:
第一,移动机械臂运动学建模。传统的移动机械臂建模方法没有考虑移动平台子系统和机械臂子系统灵活性和运动学量纲差异以及独立控制的特点,因而难以直接用于控制实际系统。针对这一问题,提出了一种移动机械臂双轨迹运动学建模方法,并引入球坐标控制参数描述机械臂末端执行器和移动平台之间的相对运动关系。
第二,移动机械臂轨迹规划。针对移动机械臂具有运动学冗余性的特点,在分析机械臂工作空间操作性能分布特点的基础上,提出了基于可操作度评价的移动机械臂轨迹规划方法。针对需要移动平台和机械臂协调运动的任务,提出了基于双轨迹运动学模型的轨迹规划方法,该方法可以在控制末端执行器运动轨迹的同时,控制移动平台沿预定义的轨迹运行,同时能够保持机械臂具有较大的可操作度,避免陷入奇异位形或关节限位。
第三,移动平台视觉伺服控制。针对工业应用环境对移动平台导航和定位特点的要求,提出了基于色带和MR-二维码的移动平台自主导航和定位方法。其次,对基于Hessian矩阵分解的光条中心提取算法进行了改进,并应用课题组开发的一种基于FPGA和DSP的嵌入式视觉处理板实现,开发了基于视觉测量的深度信息传感器。然后将此设备用于托盘的自动识别和定位,结合深度测量信息和码盘信息,完成了基于深度信息的托盘自动搬运任务。
第四,移动机械臂视觉伺服控制。针对由于环境条件和系统自身物理结构的限制,移动机械臂无法同时跟踪预定义的双轨迹问题,提出了基于任务优先级的移动机械臂双轨迹控制方法。并将该方法与基于位置的视觉伺服方法相结合,完成了移动机械臂的运动目标抓取任务。针对无法直接估算出目标位姿的应用,通过结合图像特征信息和深度信息,给出了一种结合了深度信息的图像雅克比矩阵获取方法,并设计了移动机械臂的视觉伺服控制策略,完成了基于图像的移动机械臂目标抓取任务。
最后,对论文中的工作进行了总结,并讨论了在本文基础上可以进一步开展的工作。
;
Combining the mobility of mobile robot with the manipulator’s operability, the mobile manipulator has more motion freedom and wider operating space than both the mobile robot without operational capacity and the manipulator with limited workspace. So the mobile manipulator receives more and more attention in robotics field. The mobile manipulator is not simply the summation of the mobile platform subsystem and the manipulator subsystem, so when deal with the problem of mobile manipulator visual servoing the system’s specific movement and control characteristic must be seriously considered. This dissertation mainly focuses on the mobile manipulator’s motion planning and visual servoing control, and the contents of the dissertation are as following:
Firstly, kinematic modeling of the mobile manipulator. The traditional modeling method of the mobile manipulator didn’t consider the mobile platform subsystem and manipulator subsystem’s flexibility and the kinematics dimensions difference, and also the independent control characteristic, so it is difficult to be directly applied in controlling the actual system. In order to solve this problem, a new mobile manipulator modeling method named dual-trajectory modeling method is proposed, besides the spherical parameters are introduced to describe the relative movement relationship between the end-effector and the mobile platform.
Secondly, trajectory planning of the mobile manipulator. Considering the kinematic redundancy of the mobile manipulator, the distribution characteristic of the operating performance of the manipulator is analyzed and then the trajectory planning method based on manipulability is proposed. Aiming at the tasks that need the coordinated motion of the mobile platform and the manipulator, the trajectory planning method based on dual-trajectory kinematic model is forwarded. This method will simultaneously control the end-effector and the mobile platform moving along the predefined trajectory, and at the same time, keep the manipulator in larger manipulability state to avoid falling into singular configuration or joint limitation.
Thirdly, visual servoing control of the mobile platform. In order to meet the industrial application environment’s need on mobile platform navigation and location features, the autonomous navigation and location method based on colored adhesive tape and MR-code is proposed. And the light center location algorithm based on Hessian matrix decomposition also is improved, and one kind of depth information sensor based on visual measuring is developed by the intelligent image processing card consists of FPGA and DSP. Then the device is applied in pallet automatic identification and location. Combining the depth measuring information with the encoder information, the automatic pallet handing tasks based on depth information is accomplished.
Fourthly, visual servoing control of the mobile manipulator. Because of the limitation of environmental conditions and the system’s physical structure, the mobile manipulator could not track dual-trajectory simultaneously, the dual-trajectory control method of the mobile manipulator based on the task priority is forwarded. Then combining this method with the visual servoing control strategy based on position, the task of moving target grasping is accomplished. Considering the application that the target’s pose cannot be estimated directly, the image jacobin matrix combined image features and depth information is presented. Then the visual servoing control strategy of the mobile manipulator is designed and the target grasping task of the mobile manipulator based on image features is completed.
Finally, the conclusions of the research are given and the future work is addressed.
文献类型学位论文
条目标识符http://ir.ia.ac.cn/handle/173211/14673
专题毕业生_博士学位论文
作者单位中国科学院自动化研究所
推荐引用方式
GB/T 7714
王士军. 移动机械臂视觉伺服关键技术研究[D]. 北京. 中国科学院大学,2017.
条目包含的文件
文件名称/大小 文献类型 版本类型 开放类型 使用许可
王士军 论文明审.pdf(3859KB)学位论文 暂不开放CC BY-NC-SA请求全文
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[王士军]的文章
百度学术
百度学术中相似的文章
[王士军]的文章
必应学术
必应学术中相似的文章
[王士军]的文章
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。