Implementación de un sistema de detección de patrones asociados a la escritura aérea, sobre un sistema embebido de 32 bits, para el fortalecimiento de procesos pedagógicos

Abstract

Over time, neural networks and machine learning are becoming increasingly integrated into daily activities and production processes at an industrial level. This is why the development of a prototype as a pedagogical resource is presented below, a strategy that offers an alternative to the conventional method of learning numbers. This model is based on Tiny ML and is implemented through the Edge Impulse development platform, together with the Arduino Nano 33 BLE Sense. This makes it possible to identify the numbers from zero to nine, graphed in aerial writing. The physical design of the prototype resembles a wand and is designed in AutoCAD. In addition, for better visualization, the data is transmitted through the BLE communication protocol and is displayed in a graphical user interface designed in Python using the wxPython framework.