SVM-based system for point-to-point hand movement

A system for humanoid robot arm movement concerns about generating a human-like point-to-point (p2p) trajectory. For more than a decade, numerous systems have been devised and many of them were based on complex dynamical systems. In this paper, we introduce a simpler system, integrating support vector machine (SVM) learning model, which can achieve same p2p objective. In our experiment, we compare our system with another version where SVM model component is absent and show that in many cases, our proposed system outperforms and effective for all experiments.