Adaptive Robust Visual Servoing/Force Control for Robot Manipulator With Dead-Zone Input
Adaptive Robust Visual Servoing/Force Control for Robot Manipulator With Dead-Zone Input
复制标题
DOI：
10.1109/access.2019.2930074
复制
发表时间：
2019-07
期刊：
影响因子：
3.9
通讯作者：
Yu Zhang;C. Hua;Junlei Qian
中科院分区：
文献类型：
--
作者：
Yu Zhang;C. Hua;Junlei Qian
研究了不确定机器人机械臂的自适应鲁棒视觉伺服/力控制问题。考虑机器人在未知约束曲面上工作，假设图像平面上特征点的未知深度是时变的。对未知约束曲面进行线性参数化，设计新的自适应律对未知参数进行在线估计。在基于图像的视觉伺服控制中，未知时变深度在整个雅可比矩阵中呈现非线性，且不能与其他不确定的运动参数一起自适应，具有特殊的作用。为了解决这一问题，提出了一种补偿深度无关的交互矩阵框架和相应的自适应律来补偿闭环动力学中的深度。此外，通过设计鲁棒补偿项和自适应律，考虑并处理了死区输入约束。利用所提出的控制方案，证明了图像位置和力跟踪误差渐近收敛于零。最后，数值模拟和实验结果验证了该方法的有效性。
In this paper, we investigate the adaptive robust visual servoing/force control problem for the uncertain robot manipulator. The robot is considered to work with an unknown constraint surface and the unknown depth of feature point in image plane is assumed to be time-varying. The unknown constraint surface is linearly parameterized, and new adaptive laws are designed to estimate the unknown parameters online. In image based visual servoing control, the unknown time-varying depth plays a special role as it appears nonlinearly in the overall Jacobian matrix and cannot be adapted together with other uncertain kinematic parameters. To handle this problem, a compensatory depth-independent interaction matrix framework and the corresponding adaptive laws are proposed to compensate for the depth in the closed-loop dynamics. Moreover, the dead-zone input constraint is considered and handled by designing the robust compensatory terms and adaptive laws. With the proposed control scheme, it is proved that the image position and force tracking errors converge to zero asymptotically. Finally, numerical simulation and experimental results illustrate the effectiveness of the proposed approach.