Loading...
Publication
Optimização dos parâmetros de um robô hexápode através de algoritmos genéticos
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 3.46 MB | Adobe PDF |
Authors
Abstract(s)
Os robôs com pernas apresentam vantagens significativas quando comparados com os
veículos tradicionais que apresentam rodas e lagartas. A sua maior vantagem é o facto de
permitirem a locomoção em terrenos inacessíveis a outro tipo de veículos uma vez que
não necessitam de uma superfície de suporte contínua. No entanto, no estado de
desenvolvimento em que se encontram, existem vários aspectos que têm que ser
necessariamente melhorados e optimizados.
Tendo esta ideia em mente, têm sido propostas e adoptadas diferentes estratégias de
optimização a estes sistemas, quer durante a fase de projecto e construção, quer durante
a sua operação. Entre os critérios de optimização seguidos por diferentes autores podem-
-se incluir aspectos relacionados com a eficiência energética, estabilidade, velocidade,
conforto, mobilidade e impacto ambiental. As estratégias evolutivas são uma forma de
“imitar a natureza” replicando o processo que a natureza concebeu para a geração e
evolução das espécies.
O objectivo deste trabalho passa por desenvolver um algoritmo genético, sobre uma
aplicação de simulação de robôs com pernas já existente e desenvolvida em linguagem C,
que permita optimizar diferentes parâmetros do modelo do robô e do seu padrão de
locomoção para diferentes velocidades de locomoção.
Legged robots have significant advantages when compared with traditional vehicles using wheels and tracks. Their biggest advantage is that they allow the locomotion on terrains inaccessible to other type of vehicles because they don’t need a continuous support surface. However, in their actual stage of development, there are several aspects that must necessarily be improved and optimized. With these ideas in mind, different strategies have been proposed and adopted for the optimization of these systems, either during their design phase and construction, or during their operation. Among the different optimization criteria followed by different authors, it is possible to find issues related to energy efficiency, stability, speed, comfort, mobility and environmental impact. Evolutionary strategies are a way to "imitate nature" replicating the process that nature designed for the generation and evolution of species. The objective of this project is the development of a genetic algorithm, running over a simulation application of legged robots, already developed in C, which allows the optimization of various parameters of the robot model and of its gaits for different locomotion speeds.
Legged robots have significant advantages when compared with traditional vehicles using wheels and tracks. Their biggest advantage is that they allow the locomotion on terrains inaccessible to other type of vehicles because they don’t need a continuous support surface. However, in their actual stage of development, there are several aspects that must necessarily be improved and optimized. With these ideas in mind, different strategies have been proposed and adopted for the optimization of these systems, either during their design phase and construction, or during their operation. Among the different optimization criteria followed by different authors, it is possible to find issues related to energy efficiency, stability, speed, comfort, mobility and environmental impact. Evolutionary strategies are a way to "imitate nature" replicating the process that nature designed for the generation and evolution of species. The objective of this project is the development of a genetic algorithm, running over a simulation application of legged robots, already developed in C, which allows the optimization of various parameters of the robot model and of its gaits for different locomotion speeds.
Description
Mestrado em Engenharia Electrotécnica e de Computadores
Keywords
Robôs com pernas Optimização de parâmetros Algoritmos genéticos Algoritmos evolutivos Padrão de locomoção
Pedagogical Context
Citation
Publisher
Instituto Politécnico do Porto. Instituto Superior de Engenharia do Porto