Development and Validation of a Low-Cost Open-Source Force Myography Band for Hand Gesture Recognition

Wearable systems for hand gesture recognition have gained significant attention in research and clinical applications due to advances in many hand-centric control interfaces for technologies such as upper limb prostheses, hand exoskeletons, and virtual reality for rehabilitation. However, common hand gesture recognition techniques are often associated with high costs and proprietary software which can reduce the accessibility of these devices. Thus, in this work we developed, and feasibility tested a low-cost open-source alternative in the form factor of a wearable forearm band. Our system leverages force myography (FMG) which recognizes patterns in the radial muscle forces at the skin's surface on the forearm when different hand gestures are performed. Our FMG band was validated through a participant-based study ($\mathrm{N}=15$) where ablebodied individuals were instructed to perform repetition of 10 hand gestures. Offline classification analysis of the resulting muscle force data using a linear discriminant analysis showed the band achieved an average of 94.28 % accuracy among all gestures across all participants. These results suggest that despite the FMG Band's low cost, its performance is comparable to similar gesture recognition technologies currently available for research, clinical, and consumer applications.