Real-Time EMG Classification for Bipedal Robot Control: A Hands-On Educational Platform

A hands-on educational platform is presented for teaching biosignal processing, machine learning, and robotics through real-time control of a bipedal robot using electromyo-graphic (EMG) signals. Lower-limb EMG signals are captured using Delsys sensors and processed on a Raspberry Pi, where relevant features are extracted and classified using a K-Nearest Neighbors (KNN) algorithm. The recognized movements are then mapped to motor commands and transmitted to an Arduino-based controller that actuates the bipedal robot. The system is designed to create a demonstrative module for hands-on learning in biomedical engineering and robotics education. Experimental results demonstrate reliable physical execution of learned movements with the implemented algorithm. By integrating signal acquisition, real-time classification, and robotic control, this platform provides a practical and engaging approach for students to explore human-machine interfaces.