In this paper, an imitation learning approach for learning and generalizing grasping skills based on human demonstrations is presented. To this end, motions of robot hand (including finger, wrist, and elbow) with assumption that they are decoupled and each motion moves in two dimension space are investigated. Each element of the robot hand is modeled by an affine nonlinear equation and output of each motion ensure to track a prior trajectory. | Designing and implementing a controller for real robot hand grasping
