Tecnologia vestível aplicada à segurança de pessoas com surdez: proposta de um dispositivo inteligente para reconhecimento de sons de emergência e emissão de alertas visuais e táteis

Abstract

This study proposes an artificial intelligence-based device designed to enhance the safety of deaf individuals by recognizing emergency sounds and issuing visual and tactile alerts. The objective is to develop a solution capable of identifying critical sound events and converting them into accessible forms of communication. The theoretical framework covers deafness, accessibility, assistive technologies, wearable devices, and embedded artificial intelligence for sound recognition. The methodology is characterized as applied research employing a mixed-methods approach and exploratory, descriptive, and experimental objectives; it involves a literature review, the definition of the solution's architecture, and the development of a classification model based on convolutional neural networks. The results demonstrate that the model achieved an accuracy of 86.40%, highlighting its ability to distinguish emergency sounds from ambient noise. It is concluded that the solution contributes to improving accessibility and safety for deaf individuals, serving as a foundation for future implementations and validations in real-world environments.