Uso de Técnicas de Inteligência Artificial para o Desenvolvimento de Jogos Sérios Aplicados a Estrutura Robótica para Reabilitação do Tornozelo Humano

• Autor(a) principal: Freitas, Matheus de Souza
• Data de Publicação: 2024
• Tipo de documento: Trabalho de conclusão de curso
• Idioma: por
• Título da fonte: Repositório Institucional da UFU
• Texto Completo: https://repositorio.ufu.br/handle/123456789/43570
• topic: reabilitação do tornozelo humano; human ankle rehabilitation; jogos sérios; serious games; inteligência artificial; artificial intelligence; estrutura robótica; robotic structure; protocolo de comunicação; communication protocol; CNPQ::ENGENHARIAS

This Bachelor's Thesis (TCC) involves the use of Artificial Intelligence (AI) techniques to the creation of serious games aimed at the rehabilitation of the human ankle. Its objective is to use AIs to help the patient in real time in carrying out their therapy, guiding them to a foot movement appropriate to their condition while playing, favoring the feeling of accomplishment when achieving a good score in game. Primarily, the creation of games and serious games was studied, searching for what factors must be considered to entertain the player, and what characteristics must be present to allow a engaged rehabilitation for the user. The scenario of AI in games and physiotherapy, as well as the types of existing equipment and feedback provided by patients were also taken into account. A game was then developed for each of the three main movements of the human ankle, seeking an approximation, and easy association, of the movement to be performed by the user and its outcome on the game screen. Subsequently, the establishment of a communication between hardware and software was done, taking the movement carried out on the existing test bench to the game character (in Unity), enabling a responsive and fast transmission of information based on the communication protocols employed. This phase also consists of the integration process, centralizing the main data processing and analysis steps in Matlab in order to facilitate communication and future changes. Afterwards, the most appropriate artificial intelligence technique was then selected and applied to the existing system and then trained, while changing configurations and searching for an Artificial Neural Network (ANN) capable of playing by itself, which means, taking control of the character and reaching the best possible score. At last, the entire system was integrated, a method was developed to let the AI manipulate the electric motor and it was possible to make it play, consistent with the initial objective of the study, thus demonstrating the potential for the developed serious games to be used in ankle rehabilitation procedures. With these results, there is potential for the game in aiding the user by guiding them as needed to complete the desired actions for character movement in the games.