Instrumentação de um colchão com sensores de fibra óptica plástica para prevenção de úlceras por pressão

• Ano de defesa: 2023
• Autor(a) principal: Natali, Anny dos Santos
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Espírito Santo; BR; Mestrado em Engenharia Elétrica; Centro Tecnológico; UFES; Programa de Pós-Graduação em Engenharia Elétrica
• Palavras-chave em Português: Fibra óptica poliméricas; Colchão instrumentado; Sensores; Úlceras por pressão; Engenharia Elétrica
• Link de acesso: http://repositorio.ufes.br/handle/10/12699

According to the Brazilian Institute of Geography and Statistics (IBGE) and the Ministry of Human Rights and Citizenship (MDHC), through sociodemographic analysis in 2022 released by IBGE and MDHC, it indicates that 18.6 million people, which corresponds to 8 .9% of the population in this age group aged 2 or over has a disability in Brazil. Bedridden people are more likely to be affected by pressure ulcers or skin sores, which appear as a result of prolonged pressure generated by a weight force on the skin and underlying tissues that are located between bony prominences of the body and an external surface, like the surface of a mattress. This work’s main objective is to prevent the appearance of pressure ulcers, with the use of force sensors that will allow identifying in which position a volunteer is lying on the mattress and thus allowing, in the future, with the use of data obtained through the sensors, the development of an alert system. The alert system will check whether the patient has been lying in the same position for a long time and if it identifies that this has occurred, it will generate an alert to move the patient to avoid the appearance of skin damage. The entire process of developing force/pressure sensors based on low-cost, easy-to-implement and scalable plastic optical fiber is demonstrated. The sensors, after manufactured, were instrumented on a mattress to allow the identification of three different positions of the body of a person lying down, the positions carried out were supine, prone and lateral. To identify the positions, classification algorithms known as SVM, MLP and RF were applied, and through the evaluation metric F1-score which is the combination of precision and recall metrics it is possible to evaluate the performance of models, which presented satisfactory results with average F1-score rates above 90%, allowing the identification of the positions held by the volunteers.