Xinyu Jiang Chenfei Ma Kianoush Nazarpour ^*

• University of Edinburgh, Edinburgh, United Kingdom

Myoelectric control systems translate different patterns of electromyographic (EMG) signals into the control commands of diverse human-machine interfaces via hand gesture recognition, enabling intuitive control of prosthesis and immersive interactions in the metaverse. The effect of arm position is a confounding factor leading to the variability of EMG characteristics. Such variability can normally be mitigated by training a highly generalisable but massively parameterised, computationally complex, high-capacity model, using a large training dataset collected at diverse arm conditions from a target user. Developing a simple, explainable, efficient and parallelisable model, with the model characteristics and performance invariant across postures, could largely promote the translation of myoelectric control into real world practice. Here we propose a self-calibrating random forest (RF) model which can (1) be pre-trained on data from many users, then one-shot calibrated on a new user and (2) self-calibrate in an unsupervised and autonomous way to adapt to varying arm positions. Analyses on data from 86 participants (66 for pre-training and 20 in real-time evaluation experiments) demonstrate the high generalisability of the proposed RF architecture to varying arm positions.