Sistema de identificação de direção de fonte sonora para auxiliar deficientes auditivos

• Ano de defesa: 2020
• Autor(a) principal: Machado, William
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Lavras; Programa de Pós-Graduação em Engenharia de Sistemas e Automação; UFLA; brasil; Departamento de Engenharia
• Palavras-chave em Português: Redes naturais artificiais; Detector de sinal sonoro; Somatório de diferenças absolutas; Acessibilidade; Deficientes auditivos; Direção sonora; Artificial neural networks; Sound detector; Sum of the absolute differences; Acessibility; Deaf; Sound direction; Automação Eletrônica de Processos Elétricos e Industriais
• Link de acesso: http://repositorio.ufla.br/jspui/handle/1/50285

There are thousands of people in the world with disabilities who have special needs. With urban development, difficulties arise in locomotion on public roads and access to public transport. Each type of disability requires a certain type of adaptation for the individual to be able to perform their daily tasks. A very common problem in the day-to-day life of the hearing impaired is the difficulty in detecting the direction of sound signals from sound sources. Often, the hearing impaired cannot distinguish which side (right or left) an audible alert was issued. In order to offer an alternative to this problem, the present work proposes a system that assists the hearing impaired. The system is based on the time lag of the sound signal captured by two sound sensors (microphones), where each sensor is located on the user's shoulder. The sound signals captured from the surroundings by the sensors will have a time lag between them due to the difference in distance between the sound source and the sensors (microphones). This lag will be calculated by the software running on a microcontroller that will determine the direction (right or left) of the sound source in relation to the user according to the acquired lag. The microcontroller, in use of this information, will activate the vibracall engine placed on each shoulder of the user to alert him of the direction coming of the sound.The results obtained through the simulation proved to be promising, since the Neural Network One Step Secant (OSS) presented 96.3% of acuraccy in detecting the direction of different types of sound sources. The algorithm simulation test using the Sum of Absolute Differences showed 100% acuraccy and the tests in real environments showed a acuraccy rate of up to 85.5%. Thus, the development of work values accessibility, its social character, low cost and inclusive, as it is possible to make accessible and assist the hearing impaired in locomotion in the urban environment, avoiding accidents, in addition to providing comfort and a sense of belonging to spaces in which everyone circulates.