Mão biônica controlada por visão computacional

Abstract

Esta monografia apresenta o desenvolvimento de um protótipo de mão biônica controlada por visão computacional baseada em inteligência artificial. Voltado para aplicação industrial, este projeto conta com recursos da atual indústria 4.0 e explora as áreas de robótica, engenharia e automação industrial. Para a realização dos objetivos propostos no projeto, foram conduzidas pesquisas detalhadas por artigos, monografias e outros meios escritos a fim de criar uma base teórica técnico-científica sólida. Esse desenvolvimento teórico foi estruturado em quatro etapas fundamentais: a revisão bibliográfica, consulta aos materiais e resultados obtidos, planejamento e execução, com supervisão constante do docente orientador. O projeto se destaca pelo uso de tecnologias de ponta já citadas, aliadas aos recursos possibilitados pela impressão 3D, como a personalização e a adição de sustentabilidade possibilitada pelo PETG, material usado, reciclável e resistente. O funcionamento deste protótipo é baseado na captura de imagens via webcam, processadas em tempo real por um algoritmo de I.A. e o envio de dados para a placa microcontroladora ESP 32, responsável por movimentar as partes das mãos de maneira intuitiva e responsiva. Com potencial integração às demandas industriais mais recentes, como parte de um sistema de controle de gêmeos digitais, além disso, o projeto abre caminhos para futuras aplicações nas áreas da medicina, sobretudo em tecnologias assistivas e reabilitação. Os resultados obtidos, o desenvolvimento da estrutura de antebraço e mão biônica capaz de espelhar movimentos naturais de uma mão humana por meio da captura de imagens combinado com recursos como impressão 3D sustentável e a introdução de IA no projeto, demonstraram viabilidade e inovação por meio de um projeto acessível, inovador, inclusivo, sustentável e tecnológico, com impacto significativo nos campos da biomecânica e robótica, e grande impacto no âmbito industrial. O desenvolvimento do projeto foi valioso para o grupo, contando com obstáculos relacionados à detecção, superado e à manipulação de objetos, enriquecendo o aprendizado e apontando possibilidades para futuras melhorias e expansões do projeto.

This monograph presents the development of a prototype of a bionic hand controlled by computer vision based on artificial intelligence. Aimed at industrial application, this project relies on resources from the current Industry 4.0 and explores the areas of robotics, engineering and industrial automation. To achieve the objectives proposed in the project, detailed research was conducted using articles, monographs and other written media to create a solid technical-scientific theoretical basis. This theoretical development was structured in four fundamental stages: bibliographic review, consultation of materials and results obtained, planning and execution, with constant supervision by the supervising professor. The project stands out for the use of cutting-edge technologies already mentioned, combined with the resources made possible by 3D printing, such as customization and the addition of sustainability made possible by PETG, a recyclable and resistant material used. The operation of this prototype is based on capturing images via webcam, processed in real time by an AI algorithm and sending data to the ESP 32 microcontroller board, responsible for moving the parts of the hands in an intuitive and responsive manner. With potential integration into the latest industrial demands, as part of a digital twin control system, the project also paves the way for future applications in the areas of medicine, especially in assistive technologies and rehabilitation. The results obtained, the development of the bionic forearm and hand structure capable of mirroring natural movements of a human hand through image capture combined with resources such as sustainable 3D printing and the introduction of AI into the project, demonstrated feasibility and innovation through an accessible, innovative, inclusive, sustainable and technological project, with a significant impact in the fields of biomechanics and robotics, and also a great impact in the industrial environment. The development of the project was valuable for the group, counting on obstacles related to detection, overcoming and manipulation of objects, enriching the learning and pointing out possibilities for future improvements and expansions of the project.

This monograph presents the development of a prototype of a bionic hand controlled by computer vision based on artificial intelligence. Aimed at industrial application, this project relies on resources from the current Industry 4.0 and explores the areas of robotics, engineering and industrial automation. To achieve the objectives proposed in the project, detailed research was conducted using articles, monographs and other written media to create a solid technical-scientific theoretical basis. This theoretical development was structured in four fundamental stages: bibliographic review, consultation of materials and results obtained, planning and execution, with constant supervision by the supervising professor. The project stands out for the use of cutting-edge technologies already mentioned, combined with the resources made possible by 3D printing, such as customization and the addition of sustainability made possible by PETG, a recyclable and resistant material used. The operation of this prototype is based on capturing images via webcam, processed in real time by an AI algorithm and sending data to the ESP 32 microcontroller board, responsible for moving the parts of the hands in an intuitive and responsive manner. With potential integration into the latest industrial demands, as part of a digital twin control system, the project also paves the way for future applications in the areas of medicine, especially in assistive technologies and rehabilitation. The results obtained, the development of the bionic forearm and hand structure capable of mirroring natural movements of a human hand through image capture combined with resources such as sustainable 3D printing and the introduction of AI into the project, demonstrated feasibility and innovation through an accessible, innovative, inclusive, sustainable and technological project, with a significant impact in the fields of biomechanics and robotics, and also a great impact in the industrial environment. The development of the project was valuable for the group, counting on obstacles related to detection, overcoming and manipulation of objects, enriching the learning and pointing out possibilities for future improvements and expansions of the project.