Abstract
We develop a three dimensional neuro-musculo-skeletal model which can simulate human walking. Dynamics of the simulation model is expressed by 14 rigid links with 21 degrees of freedom. This simulation model can walk stable to some extent by stabilization mechanism using neural oscillator. However, in case making running motion of the child simulation model, stability is insufficient. This paper describe about new stabilization mechanism using singular value decomposition and phase oscillator. By the introduction of the new stabilization mechanism, this simulation model can adjust walking or running that has been learned. As a result, it is possible to make high speed child running motion easier and more stable.
