Sistema baseado em inteligência artificial para deteção do estado de locomoção de uma pessoa

• Autor(a) principal: Silva, Carolina Costa de Jesus da
• Data de Publicação: 2022
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10400.22/21355
• topic: Artificial Intelligence; Dataset; Accuracy; Inteligência Artificial; Precisão

Currently, there are more and more systems using Artificial Intelligence (AI), since one of its characteristics is to provide the ability for systems to work in a similar way to a human. The development of various technologies [1] [2] such as sensors, Internet-of-Things (IoT), data analysis, etc. allows the application of AI through the existence of a large number of data [3]. Artificial Intelligence is used in several applications, one of the main ones being healthcare [4]. In this perspective, the project proposal presented in this dissertation aims at using Artificial Intelligence to classify a user’s motion activities (slow or fast gait), with the data obtained through a developed plantar pressure measurement system. The developed project can be divided into three parts, the first being the development of the plantar pressure monitoring system, the second the development of the AI models, and the third the processing of the input data of the models (datasets). Two datasets from different systems were used, the first being obtained through a public repository and the second obtained through the monitoring system developed for the project. The purpose of using the two datasets was to later compare the accuracy of the models for input data from different systems. At the end of the project, when analyzing the accuracy values of each of the models for the two datasets, it was possible to conclude that, with the data from the monitoring system developed for the project, greater accuracy of the classification model was obtained than with the data from the public dataset.