Modeling of 3D Grasping of Artificial Hand Under Dynamic Load. A Novel Approach to Control the Robotic Hand Grasping Process by Using Artificial Neural Network Algorithm

      Аннотация: In this book, a novel design of fingertip to detect slipping of the grasped object under dynamic load conditions was proposed to measure the normal and tangential contact force components. The proposed design was treated with conventional sensors and has the ability to optimize the design of the contact region without the need to change the sensory system. The proposed design was modeled mathematically, starting with the derivation of the velocity kinematics for any point on the system elements (Arm, Hand and Grasped object) from the base coordinate to the contact point coordinate. Also, the kinematics of contact for the grasped object are derived to explain the relation between the contact forces on the surface of the object and the resultant external force which applied to the object. The artificial neural network was trained based on the patterns of force component ratio signals at slippage occurrence, in order to detect slippage according to understanding of the behavior of contact force component ratio with respect to the coefficient of friction at slip occurrence without the need to any knowledge about the grasped object characteristics.

Ключевые слова: Neural Network, slip detection, Tactile Sensor, Robotic Hand