Modeling and taylormade training method using neural network for specific muscle of the upper limb

In this paper, we present modeling and taylormade training method using neural network for specific muscle of the upper limb. Validity of neural network model is demonstrated by comparing with the previous musculo - skeletal model in actual experiments. Applying the neural network model, an optimum path maximizing the activity of an agonist muscle and minimizing that of other muscles is designed by using an optimization method. The degree of muscles revitalization is evaluated by the amplitude of EMG signal, and the effectiveness of the proposed method is demonstrated.