Emergence of Grasping based on tacit learning

人类的抓取行为是对环境最敏感的行为之一。我们可以根据抓取对象和所处的环境情况，形成合适的抓取姿势。为了将这种能力应用于人工手臂控制，我们使用了基于生物控制原理开发的隐性学习。仿真研究表明，所提出的控制器可以根据抓取力学习到合适的抓取姿态。我们应用我们的控制器来控制三维人工手臂，以适应剩余关节的运动，并通过实验表明，通过身体/环境的相互作用，出现了适当的行为。

Grasping by human beings is one of the most environment-sensitive behavior. We can crate appropriate grasping posture depending on the grasping object and the environmental situations. To apply this ability to the artificial arm control, we used tacit learning that is developed based on the principle of biological controls. We showed in the simulation study that the proposed controller can learn the appropriate grasping posture depending on the grasping force. We applied our controller to control the 3DOF artificial arm to adapt the motion of the remaining joints and experimentally showed that the appropriate behaviors emerged through the body/environment interactions.