Explorando os movimentos das mãos como interface para controlar computadores: Um protótipo utilizando o microcontrolador ESP32

Abstract

This work presents the development of a prototype system for gesture recognition as an interface to control computers. The device utilizes MPU-6050 sensors to capture motion data, which is then processed by an ESP32 microcontroller. Customized components were crafted using a 3D printer to assemble the device, a method chosen to ensure maximum comfort and flexibility for users. The data collection process involved four users performing six different hand movements, resulting in a dataset composed of 1,440 gestures. Various classifiers were evaluated to determine the system’s effectiveness, including Random Forest (RF), Multilayer Perceptron (MLP), Extreme Gradient Boosting (XGBOOST), and Support Vector Machine (SVM). The SVM classifier, configured with a Radial Basis Function (RBF) kernel and with the application of Principal Component Analysis (PCA), achieved an average accuracy of 95.14%, indicating a good generalization capability of the model and the effectiveness of dimensionality reduction in improving performance. The results obtained demonstrate the potential of the developed system for gesture-based computer control, offering a promising alternative for more intuitive and accessible human-computer interactions.