A support method for haptic skill acquisition using graph theory

Abstract

This paper proposes a method to assess human motion. This method supposes a skill acquisition support for trainee. By using the proposed method, trainee enables to evaluate how coincident own motions are with the motions of trainer. This evaluation is represented as trainee's point. Moreover, this method enables to define a number of motions as evaluation figure. These motions are conducted by a specific trainer. For this method, graph theory and correlation are employed. Concretely speaking, a value of each component in eigen matrix of the adjacency matrix is dealt as the score of appropriate motion. The adjacency matrix is derived from graph. Nodes of the graph are constructed by the motion trainer conducts and the score is defined as trainer's point. At that time, connection between two nodes of the graph is weighting by coefficient of correlation. Trainee's motion is assessed based on the trainer's point. The viability of the proposed method is confirmed by experiments.