Evaluation of Numerical Solution of Stochastic Differential Equations Describing Body Sway Using Translation Error

Abstract

Although many hypotheses, such as stiffness control and muscle mechano-reflex control, have been proposed for the elucidation of a control system to maintain an upright posture, the detailed mechanism is still unclear. In this study, the characteristics of muscle activity in the upright posture were investigated to verify the stiffness control hypothesis by employing stabilometry and the simultaneous measurement of electromyographs (EMGs). For each experimental test, with the subjects' eyes open/closed during stabilometry, EMGs were recorded for the gastrocnemius muscle, soleus muscle, and tibialis anterior muscles, simultaneously. We discuss a relationship between the center of gravity and muscle activity to control the upright posture. Mathematical models to control the standing posture are also discussed in this study.

Estimating a Minimum Embedding Dimension by False Nearest Neighbors Method without an Arbitrary Threshold

“…Then, since deep learning handles a large number of parameters, the generated model can be said to be a multivariable-dependent system. On the other hand, it has been pointed out that the biological signals to be generated are not multivariable in nature, but may be relatively simple systems that depend on a few variables at most [3,4]. To investigate the variables on which the system depends, several methods have been proposed to estimate the minimum embedding dimension of the time series using attractors.…”