Research on Estimation of Human Elbow Joint Rotational Acceleration Based on Mechanomyography and Broad Learning System

In the present study, we used a Broad Learning System and Mechanomyography to estimate rotational acceleration of elbow joints. To achieve our study objectives, we proposed a mechanomyography (MMG) signal isolation method based on variational mode decomposition (VMD) algorithm. In addition, we used Broad Learning System to estimate acceleration and got a precise result. Most importantly, we developed a MMG signal sensor based on pressure sensor, which is different of other MMG signals based on acceleration sensor. This sensor would not be disturbed by gravity acceleration and movement acceleration of human body limbs, compared with acceleration sensor. Further study is warranted to examine the effectiveness of this sensor in other kinds of body movements and to develop a new structure to avoid the interference caused by positive pressure from straps.