引用本文: 仵沛宸, 帅威, 陈小平, 高杨, 洪文, 崔国伟. 抓取任务中的融差控制方法[J]. 机器人, 2022, 44(5): 589-600.
WU Peichen, SHUAI Wei, CHEN Xiaoping, GAO Yang, HONG Wen, CUI Guowei. The Rong-Cha Based Control Method in Grasping Task[J]. ROBOT, 2022, 44(5): 589-600.
 Citation: WU Peichen, SHUAI Wei, CHEN Xiaoping, GAO Yang, HONG Wen, CUI Guowei. The Rong-Cha Based Control Method in Grasping Task[J]. ROBOT, 2022, 44(5): 589-600.

## The Rong-Cha Based Control Method in Grasping Task

• 摘要: 依据“融差性思维”，提出了无需精确感知依旧可以在一定范围内有效工作的融差控制方法。具体分析了融差抓取方法如何运用相同控制量实现不同抓取任务的工作原理，这一原理使得融差抓取方法在面对一大类抓取任务时，不需要知道物体的具体参数，只需要知道这一大类物体的边界条件。进一步分析了融差抓取方法在欠驱动手爪上的适用性，并发现了欠驱动手爪的局限性。实验表明，在控制量设定不变的情况下，依据融差抓取方法，柔性手爪可以抓住且不抓坏宽度范围为5~45 mm的嫩豆腐，且能够成功抓取宽度范围为5~60 mm的硬质长方体；弹簧关节欠驱动手爪可以抓住且不抓坏宽度范围为20~40 mm的嫩豆腐，且能够成功抓取宽度范围为5~60 mm的硬质长方体。这体现了融差抓取方法的通用性和欠驱动手爪在抓取柔性物体时的局限性。最后，展示了柔性手爪使用融差抓取方法在桌面抓取应用中以简单的控制策略成功抓取不同形状、不同材质的物体。这充分说明了融差抓取方法不依赖于精确的对象感知及物体模型，能够简化控制策略。

Abstract: Based on the principle of Rong-Cha, a novel control method is proposed, which can work effectively within a certain range without precise perception. The working principle of how to use the same control input to achieve different grasping tasks by the Rong-Cha based grasping method is analyzed. Based on this principle, the grasping method can cope with a large class of grasping tasks without knowing the specific parameters of the objects, only the boundary conditions of this large class of objects are needed to know. The applicability of the Rong-Cha based grasping method to underactuated grippers is analyzed, and limitations of underactuated grippers are unveiled. The experiments show that by the Rong-Cha based grasping method and with the unchanged control inputs, the soft gripper can grasp the soft tofu without damage within the width range of 5~45 mm, and can successfully grasp the rigid cuboid within the width range of 5~60 mm. The spring-jointed underactuated gripper can grasp soft tofu within the width range of 20~40 mm without damage, and can successfully grasp the hard cuboid within the width range of 5~60 mm. The results demonstrate the versatility of the Rong-Cha based grasping method and the limitations of the underactuated gripper in grasping soft objects. Finally, it is shown that the soft gripper uses the Rong-Cha based grasping method to successfully grasp objects of different shapes and materials with a simple control strategy in a desktop grasping application. This fully demonstrates that the Rong-Cha based grasping method doesn't rely on the accurate object perception and the corresponding object model, and can simplify the control strategy.

/

• 分享
• 用微信扫码二维码

分享至好友和朋友圈