基于多自由度机械臂的临场感遥操作技术研究

发布时间：2018-01-25 11:04

本文关键词： 点云建模聚类分析虚拟力遥操作动力学　出处：《西南科技大学》2017年硕士论文　论文类型：学位论文

【摘要】：由于应用环境复杂多变目前机器人智能化程度还不能完全实现自主完成复杂的作用任务。本文围绕遥操作主从机器人在未知非结构环境中从端机械臂精确跟随、主端操作者对从端环境感知等问题提出了基于图像虚拟力觉的多自由度主从机械臂力觉临场感遥操作技术的研究。在遥操作主从机器人控制过程中,主端操作者主要依靠从端传输的图像、力、位置等信息操作从端机械臂完成作业任务。由于从端机械臂处在远端,操作者距离从端机械臂距离较远,无法实际感受从端机械臂相对于环境的真实位姿,无法精确得知从端机器人与作业对象之间的位置关系,从端机器人任务完成的程度高度依赖操作者的熟练程度,导致主从控制系统实际操作效率低下。同时在遥操作主从控制中广泛使用的虚拟夹具、虚拟管道主要针对具体环境使用软件建模,通常需要对任务空间进行预先扫描与采样,建立环境模型后再构建相应的虚拟夹具、虚拟管道。预建模无法精确应对复杂、非结构、易变化的环境。针对上述问题,本文通过对从端环境点云信息进行聚类分析,建立点云模型,采用人工势场法建立人工势场,形成基于目标及障碍物的虚拟引力和虚拟斥力。通过构建稳定的主端力觉临场感系统,在从端机械臂运动过程中主端操作者借助力反馈设备能够充分的感受从端环境虚拟引力/斥力,实现了对操作者的力觉辅助,在力觉引导过程中操作人员能够快速接近作业对象、避开障碍物。系统在设计过程中将主端操作者的思维决策的准确性、高效性同机械臂的控制策略相结合,提高机器人的作业效率,减小操作的操作压力,降低了对操作者熟练程度的要求,实现了人-机交互的高度融合。综上,论文基于C++语言结合Open CV2.4.9和PCL1.6.0在VS2010平台及Matlab 2010b下对系统进行了设计与实现。
[Abstract]:Due to the application of complex environment the robot intelligence can not be fully realized independently task complex. This paper around the teleoperation slave robot in unknown unstructured environment from the end of the mechanical arm accurately follow the main end operator of the proposed research on multi DOF master-slave arm virtual image McGregor sleep based on tele operation technology from the end of the environmental perception problems. In teleoperation robot control process, the main terminal operator mainly rely on from the image terminal transmission power, position information operation from the end manipulator to complete the task. Because of the end of manipulator in the distal end of the mechanical arm, the operator's distance from the distance, not the actual feeling from the end of the mechanical arm relative to the true pose of the environment, not to know precisely the relationship between the position of the robot and the object from the end of operation, from the end of robot task completion process Is highly dependent on the operator's proficiency, resulting in a master-slave control system actual operation efficiency. At the same time in the virtual fixture is widely used in master-slave teleoperation control, virtual pipeline according to the specific environment for the use of software modeling, usually need to pre task space scanning and sampling, the environment model was established after the construction of virtual fixture, the corresponding virtual the pre modeling pipeline. Cannot accurately deal with complex, unstructured, easy to change the environment. Aiming at the above problems, based on the cluster analysis from the end point cloud information, the establishment of the point cloud model is build by artificial potential field with artificial potential field method, the formation of virtual gravity targets and obstacles and through the virtual repulsive force based. The main building stable end force telepresence system, the operator from the main end manipulator movement process by means of force feedback device can fully feel the gravity / virtual environment from the end The realization of the operator's repulsion force assistant, in force personnel to operate the boot process to quickly close the job object, avoiding obstacles in the system. The design process, the main end operator decision accuracy, combining control strategy with high efficiency mechanical arm, improve the working efficiency of the robot, reduce operation pressure, reduce the operator proficiency requirements, to achieve a high degree of integration of human-machine interaction. In conclusion, the combination of Open CV2.4.9 and PCL1.6.0 C++ language in VS2010 platform and Matlab 2010b of the system was designed and implemented based on.

【学位授予单位】：西南科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242

【相似文献】