A Review on EMG-based Pattern Identification Methods for Effective Controlling of Hand Prostheses

The ability of amputees to do daily duties is significantly restricted by upper limb amputation. The myoelectric prosthesis uses impulses from the surviving muscles in the stump to gradually restore function to such severed limbs. Such myosignals are unfortunately tedious and challenging to gather and employ. The process of transforming it into a user control signal after it has been acquired often consumes a significant amount of processing resources. By modifying machine learning strategies for pattern recognition, the factors that influence the traditional electromyography (EMG)-pattern identification approaches may be significantly minimized. Although more recent developments in intelligent pattern recognition algorithms could discern between a variety of degrees of freedom with high levels of accuracy, their usefulness in practical (amputee) applications was less obvious. This review paper examined how well various pattern recognition algorithms for hand prostheses performed. Finally, we discussed the current difficulties and offered some suggestions for future research in our paper’s conclusion.