Prototype Design of Deep Learning-based Voice Control Model for Smart Home

Abstract

The demand for smart home technology is increasing to help older people feel more comfortable at home. Smart home technology can support the elderly in independent daily activities. The Internet of Things (IoT) is currently one of the key platforms for data-driven smart homes. The automated process of recognizing or verifying an individual’s identification based on his speech is known as voice recognition or speaker recognition. The main challenge in adjusting to the evolution of conversations in society is that the systems generally refer to existing patterns in the database. Therefore, we propose a prototype design of a smart home’s deep learning-based voice control model. First, we develop the model based on the convolutional neural network (CNN) and deep neural network (DNN) to obtain the best accuracy. Then, we create a model-based CNN and DNN used to construct a voice recognition system independent of text and language. The simulation result shows that the proposed model could extract the voice sample. The result also indicates that the accuracy of using CNN is better than that of using DNN.