Hybrid motion control and planning strategies for visual servoing
Hybrid motion control and planning strategies for visual servoing
复制标题
DOI：
10.1109/tie.2005.851651
复制
发表时间：
2005-08-01
影响因子：
7.7
通讯作者：
Wilson, WJ
中科院分区：
文献类型：
--
作者：
Deng, LF;Janabi-Sharifi, F;Wilson, WJ
本文提出了两种混合策略的机器人视觉伺服。两个特定的图像约束，图像奇异性和图像局部极小值，被认为是在这两种策略。混合运动控制策略包括一个本地切换控制之间的基于图像和基于位置的视觉伺服直接避免图像奇异和图像局部极小值。混合运动规划策略包括一个人工势场为基础的全球混合轨迹规划器，其中一套完整的笛卡尔，图像，和机器人关节约束下的一个复杂的视觉伺服的情况下被认为是。在该策略中，图像奇异点通过基于阻尼最小二乘的联合轨迹规划来解决，而图像局部极小点仅沿规划的图像轨迹沿着评估，并在基于图像的轨迹跟踪中自动避免。两个全球规划方法被认为是。在第一种方法中，末端执行器的轨迹是直接规划相对于固定的目标对象框架，这提供了一个更短的平移路径相比，局部规划方法。在第二种方法中，相对于当前末端执行器帧规划目标轨迹，这最小化了图像轨迹离开相机视场的机会。仿真和实验结果证明了这两种混合策略的有效性。
This paper presents two hybrid strategies for robot visual servoing. Two specific image constraints, the image singularities and image local minima, are considered in both strategies. The hybrid motion control strategy consists of a local switching control between the image-based and position-based visual servoing for direct avoidance of image singularities and image local minima. The hybrid motion planning strategy consists of an artificial potential field-based global hybrid trajectory planner, where a complete set of Cartesian, image, and robot joint constraints under a complex visual servoing scenario are considered. In this strategy, the image singularities are resolved using the damped-least-square-based joint trajectory planning, while the image local minima are evaluated only along the planned image trajectories and automatically avoided in the image-based trajectory tracking. Two global planning methods are considered. In the first method, the end-effector trajectory is directly planned with respect to the stationary target object frame, which provides a much shorter translational path compared with the local planning method. In the second method, the target trajectory is planned with respect to the current end-effector frame, which minimizes the chances of image trajectories leaving the camera field of view. Simulation and experimental results are given to demonstrate the efficiency of the two hybrid strategies.