Voice processing with Internet of Things for a home automation system

Abstract

This paper presents a voice processing for a domotic application in the development of the Internet of things. The recognition was based on a computational intelligence implemented in the Raspberry Pi B + embedded system, which is controlled by voice using the "Comand and Control" method from a mobile application created in Android Studio, which in addition to receiving and sending the commands also allow to monitor the real-time states of the connected devices through a database. We developed and implemented an Acoustic Model and a Language Model for the people who live in the Norte de Santander region. The voice processing integrates three devices that allow controlling the lighting of the house, in addition to the measurement of temperature, humidity and presence, applying bidirectional radiofrequency communication with the Raspberry Pi. The evaluation of the system consisted of a point-to-point comparison with a licensed system considering the characteristics required by the user and obtaining a reduction in costs of 57 % including labor and a commercial profit of 50 %.