Adaptive learning of multi-finger motion and force control

This paper is based on a prosthetic human organ targeting human arm with simulated / manipulated fingers with a highly related tendon driven mechanism with force sensing feedbacks. The control environment is mostly common to the both real and prosthetic human arms. It is a neural feedback based mechanism which is normally in human arms and here with this project it is integrated with "stall current sensing feedback system" possibly taken as a force feedback system literally. Based on support vector machine (SVM), this paper proposed an adaptive learning procedure intending to approximate the mapping among object positon and the corresponding joint displacement. Finally the application will run as smooth as possible with respect to the given objected environment with grabbing and releasing most common objects as well as improving the realistic projection, which is manipulating the human arm prosecution with more than 75% of possibility.