Design and Validation of a Three-Channel EMG Classification System Using Indigenously Developed Dry Electrodes

Reliable and affordable electromyography (EMG) signal acquisition is a critical requirement for myoelectric prosthetic control systems. Conventional gel-based electrodes offer good signal quality but suffer from limitations related to single- use design and long-term stability while commercial EMG acquisition systems are very costly. This paper presents a comparative study of custom-designed copper-based dry electrodes and commercially available gel electrodes for sur- face EMG acquisition, along with the indigenous development of a low-cost and customizable data acquisition module as an alternative to commercial solutions. The performance of both electrode types is evaluated through EMG signals acquired from forearm muscles. Computationally efficient time-domain features are extracted and used to train an artificial neural network (ANN) for move- ment classification. The trained ANN model is deployed on a microcontroller and integrated into a custom-built active prosthetic hand, enabling real-time control without reliance on external computing hardware. Experimental results demonstrate that the proposed dry electrodes provide stable and reliable EMG signals, while the developed acquisition system supports accurate classification and embedded implementation. The presented approach validates the feasibil- ity of an indigenously developed, fully embedded, cost-effective EMG-based prosthetic control system suitable for practical and wearable applications.