Wireless Hand Gesture Robot Control by Voice Recognition using Machine Algorithm

This research focuses on developing a hand gesture-controlled robot using machine learning (ML) algorithms and an accelerometer, with the goal of creating a more intuitive human-robot interaction. Initially centered on signal processing, the project evolved to integrate both gesture recognition and voice command identification for robot control. The system utilizes an ADXL335 accelerometer connected to an Arduino Mega and an L293D motor driver to detect hand movements and translate them into directional commands (left, right, forward). A machine learning model was developed to classify spoken commands ("front," "back," "right," "left") using the Mel-frequency cepstral coefficients (MFCCs) for feature abstraction of the data and support vector machines (SVM) for training along with testing, achieving a 92.59% accuracy. Despite the success of the ML model, time constraints prevented full integration with the robot's physical control system. This project lays the groundwork for future research aimed at combining voice and gesture control, contributing to more seamless human-robot interaction.