Introdução ao controle adaptativo a eventos discretos em base de dados RFID com planejador automático aplicado a sistemas robóticos

• Autor(a) principal: Souza, Gabriel de Almeida
• Data de Publicação: 2020
• Tipo de documento: Trabalho de conclusão de curso
• Idioma: por
• Título da fonte: Repositório Institucional da UFU
• Texto Completo: https://repositorio.ufu.br/handle/123456789/30558
• topic: Sistemas ciber físicos; RFID; PRD; GISP; Planejamento automático; Aplicação de tempo real; Cloud computing; Controle a eventos discretos adaptativo; CNPQ::ENGENHARIAS::ENGENHARIA MECANICA::PROJETOS DE MAQUINAS::CONTROLE DE SISTEMAS MECANICOS; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO::METODOLOGIA E TECNICAS DA COMPUTACAO::ENGENHARIA DE SOFTWARE; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO::SISTEMAS DE COMPUTACAO::ARQUITETURA DE SISTEMAS DE COMPUTACAO

Industry 4.0 technologies integrate devices and data, bringing flexibility and efficiency, derived from decentralization of information sources and processing, which is fundamental to further advance applications. This work aims to introduce Cyber Physical Systems (CPS) working with decision affecting passive objects, a system with distributed data, and automatic planners, to achieve a self-sufficient process manipulator that does not require external goal insertion and can self-adjust given an exception, in real-time. This solution is more flexible and autonomous than state machines. Applying the technologies introduced in Industry 4.0 and methods that were previously treated separately, such as symbolic artificial intelligence and robot kinematics, the system can perform perception, planning and actuation processes. This system is capable of extracting information inside passive passive entities in the physical domain by using Radio Frequency IDentification (RFID) to acquire predicates, data, about each object current and objective states using the Predicate inside RFID Database (PRD) tool. This data is treated to produce a domain snapshot, by joining distributed information and generating a problem definition, through the Grouped Individual State Predicates (GISP) methodology. This problem definition may then be fed into a planning module, implemented on an Edge or Cloud server, where discrete-action and trajectory planning are concatenated to output control references, using a generic symbolic planner and a numeric trajectory generator. Then, the active agent may actuate, verify for exceptions and update the passive objects information if the obtained state is perceived with no exceptions, else it must reiterate to satisfy the global goal. This work structures the adaptive discrete event control architecture with a RFID database containing parts of predicate logic.