Structuring a Model to Estimate Human Joint Motion from Surface-Electromyography

In this paper, the structure of a model to estimate human motion from surface-electromyography (sEMG) was defined mathematically and physically. Especially, an infinite impulse response (IIR) filter and an element description method were focused to generate the model in this paper. Human motion can be estimated from sEMG. However, the relationship between sEMG and human motion has a complex characteristics such as a non-linear relationship. The cause of the above complex relationship can be thought the muscle contraction velocity. Then, the IIR filter was applied to consider about the previous sEMG value in the estimation. In addition, some conventional models have been a black-box model and it is not easy to interpret the model mathematically and physically. From the above, an element description method (EDM) was applied to generate the model in this paper. EDM is one of the system identification method and it is easy to interpret the model mathematically and physically because EDM can generate the block diagram of the model.In this way, the proposed method could solve two conventional methods about the muscle contraction velocity and the black box model. Then, it was possible to define the structure of the model by the above in this paper. In this paper, the models about a wrist and a fingertip motion were generated and considered about the structure. As the results, the structure of models was defined with a hyperbolic tangent function. It will be possible to apply the model to various motions by just identifying the parameters about the relationship between each joint and muscle because the amount of muscles and range of sEMG are deferent from each muscle.