Abstract:
Paralysis affecting only one limb, known as monoplegia, is frequently caused by brain injuries, strokes, or other neurological illnesses. Since monoplegia patients have severe difficulties performing daily tasks, assistive technology like robotic arms and prosthetic limbs are to be developed to help people with motor impairments to regain functionality and enhance their quality of life. However, these are expensive solutions that don’t give the patient much comfort or flexibility. Therefore, in order to improve patients’ movement and comfort we have developed an EMG-based robotic glove targeted for physical rehabilitation by providing natural gesture recognition. This advanced system incorporates state-of-the-art hardware components, including 24V and 5V solenoids, relays, batteries, Arduino Uno microcontrollers, Bluetooth modules, and Raspberry Pis. These elements work together to ensure precise control and smooth communication. Twelve participants, aged twenty to forty-five, performed open and closed hand gestures during experimental sessions, with data collected by the Myo armband at 50 Hz. Pre-processing methods like bandpass filtering and signal rectification were applied to improve data quality, and convolutional neural networks (CNNs) were used for real-time gesture classification. Through the combination of advanced signal processing methods and hardware components, we have developed a robust robotic glove system with intuitive gesture-based control. The use of solenoids and relays enables precise actuator control, facilitating delicate fnger movements and enhancing the glove’s dexterity. The Bluetooth module provides a wireless connection between the Raspberry Pi and the Arduino Uno, the system’s central control unit. Experimental outcomes demonstrate the efciency of this approach in identifying and executing predetermined movements, highlighting its potential applications in assistive technology, human-robot cooperation, and rehabilitation.
