Developing an ankle-foot muscular model using Bayesian estimation for the influence of an ankle foot orthosis on muscles

The objective of this study is to develop an estimation model of foot-ankle muscular activity for designing an ankle-foot orthosis with a training function. We built a Bayesian network model [1] and chose three muscles to confirm its effectiveness. Such a model needs to include all factors affecting the gait, for example, speed reflex movements, joint angles and so forth. In an experiment, we examined the normal gait of a non-disabled subject. We measured the muscular activity of the lower foot muscles by electromyography, the joint angles by using statistical methods, and the sole pressure on each part of the sole. From this data, we obtained the causal relationship at every 10% level of these factors. Our model has three advantages. First, it can express the influences, which change throughout the gait, because we use 10% level nodes of each factor. Second, it can express the influences of factors, which are different for low and high muscular activity levels. Last, the model can compensate the missed estimations by estimating every 10% level muscle activity. In an evaluation of this model, we confirmed that this model can estimate all muscular activity level with an accuracy rate greater than 90%.