Conventional body-powered systems eventually suffer from the limitations of the relatively low force and dexterity these can offer while requiring strenuous exertion to operate, and advanced myoelectric systems tend to rely on complex algorithms and expensive hardware and calibration routines. These challenges present significant barriers to access of functional prosthetic hands for upper-limb amputees. This paper presents an EMG-controlled three-fingered prosthetic robotic hand based on non-invasive surface electromyography signals from healthy distal residual forearm muscles. The proposed system integrates analog signal conditioning, time-domain feature extraction (Mean Absolute Value) and threshold-based classification to enable Arduino-based real-time processing for servo-driven mechanical actuation. The system achieves 92.5% single-gesture and 88.4% multi-gesture accuracy, with a total average response time of 290 ms, confirming that simplified embedded control can deliver performance and reliability for practical prosthetic functionality in a daily-assist device.
S et al. (Thu,) studied this question.