Force Estimation From Surface-EMG Using Element Description Method

There are two main contributions in this article. One of them is to have generated the interpretable model about the relationship between sEMG and force. The other is to have conducted on estimating force from sEMG with the same level accuracy as the conventional method. As the above, we proposed the effective modeling method to estimate the human force from surface-electromyography (sEMG) in this article. A sEMG is one of the human biological signal and it indicates muscle contractions. In the conventional research, the force estimation from sEMG has been conducted. However, the calculation process between sEMG and force is unclear because those methods are the machine learning such as the DNN, etc. From the above, it could not be considered about the relationship between input and output based on the model. Then, we proposed the element description method (EDM) which can generate the model whose calculation process is not black box for the force estimation from sEMG in this article. We compared the conventional method (DNN) with the EDM in this article. As the result, the root mean square error with an EDM was same degree with the DNN. Moreover, the model with an EDM was more effective than the DNN because the calculating process of the model by an EDM was interpretable. From the above, we could show the effectiveness of the proposed method in this article.