Optimizing Simulation for Long Distance Throw in Baseball from a 3-dimensional Captured Motion

摘要

It is possible to regard an ideal motion or a proficient motion as an optimal motion, especially, in the field of sports. We call an ideal motion generated under an artificial environment like a computer "artificial proficient motion". In this paper, we described the simulation method for obtaining an optimal motion during long distance throwing in baseball for an upper limb from a 3-dimensional captured motion by an optimizing calculation with a quasi-Newton method and a multiobjective function representing a proficient. The consequent motion by an optimizing calculation generally depends on the order of optimization for multiobjective function. We also reported the results of the simulation experiment when we changed the order of optimization for the multiobjective function, and discussed the consequences of the experiment in viewpoints of kinematics and dynamics.