Submitted by Automa??o e Estat?stica (sst@bczm.ufrn.br) on 2016-07-08T21:06:56Z No. of bitstreams: 1 CaioJulioCesarDoValeFernandesDaSilva_DISSERT.pdf: 5841265 bytes, checksum: b45c3017ad558e351b1dbe31689d6c95 (MD5) / Approved for entry into archive by Arlan Eloi Leite Silva (eloihistoriador@yahoo.com.br) on 2016-07-12T19:16:48Z (GMT) No. of bitstreams: 1 CaioJulioCesarDoValeFernandesDaSilva_DISSERT.pdf: 5841265 bytes, checksum: b45c3017ad558e351b1dbe31689d6c95 (MD5) / Made available in DSpace on 2016-07-12T19:16:48Z (GMT). No. of bitstreams: 1 CaioJulioCesarDoValeFernandesDaSilva_DISSERT.pdf: 5841265 bytes, checksum: b45c3017ad558e351b1dbe31689d6c95 (MD5) Previous issue date: 2015-12-17 / A explora??o de petr?leo em profundidades elevadas requer o uso de rob?s m?veis para realizar opera??es diversas, como manuten??o, montagem etc. Nesse contexto, o estudo do planejamento de trajet?rias e navega??o desses rob?s se faz relevante, visto o grande desafio que ? navegar em um ambiente que n?o ? totalmente conhecido. Assim, este trabalho tem o objetivo de criar um algoritmo de navega??o, que deve realizar o planejamento da trajet?ria de um rob? m?vel que se encontra em uma dada posi??o (x, y) e deve atingir a posi??o desejada (x_d, y_d), evitando, no entanto, a colis?o com qualquer obst?culo existente no caminho. Para a gera??o da rota global foi utilizado um algoritmo gen?tico (offline), que leva em considera??o apenas as coordenadas dos pontos a serem visitados. Para desviar dos poss?veis obst?culos no caminho, o rob? deve gerar rotas locais baseadas nas curvas de B?zier (online). Na implementa??o do programa n?o h? qualquer informa??o sobre a localiza??o ou o formato dos obst?culos, mesmo assim, o rob? deve evitar os obst?culos, baseadas nas informa??es dos sensores de proximidade. Esta estrat?gia ? v?lida na situa??o em que os obst?culos s?o pequenos em rela??o as dist?ncias entre os pontos de visita??o. Os resultados das simula??es e experimentos com um rob? m?vel real (Robotino) demonstraram que o rob? foi capaz de realizar o percurso definido pelo algoritmo gen?tico, desviando dos obst?culos atrav?s de curvas de B?zier e atingindo as posi??es desejadas dentro da margem de erro definida como aceit?vel. As principais contribui??es deste trabalho est?o no c?lculo online das curvas de B?zier no planejador de rotas locais, atrelado a um planejador de rotas global, com obten??o de resultados experimentais. / Oil exploration at great depths requires the use of mobile robots to perform various operations such as maintenance, assembly etc. In this context, the trajectory planning and navigation study of these robots is relevant, as the great challenge is to navigate in an environment that is not fully known. The main objective is to develop a navigation algorithm to plan the path of a mobile robot that is in a given position (?, ?) and should reach the desired position (??, ??) avoiding colision with any obstacle standing in the way. The global route was generated using a genetic algorithm, which takes into account only the coordinates of the checkpoints. The mobile robot path generation based on B?zier curves was able to dodge the possible obstacles in the way. There was no information about the obstacles?s shape or location during the implementation of the program yet the robot must generate local path based on information from proximity sensors located around the robot to be able to avoid collisions. This strategy is valid in the situation where the obstacles are small relative distances between the visited sites. The results of simulations and experiments with a real mobile robot shown that the robot was able to perform the route defined by the genetic algorithm, dodging obstacles by Bezier curves and reaching the desired positions within the margin of error defined as acceptable. The main contributions of this work are the equations used to define the control points in the online calculation of Bezier curves in the planner of local routes, linked to a global route planner with obtaining experimental results.