Loading...
Publication
Trajetórias de contorno de obstáculos
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 13.95 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
A crescente demanda por soluções de automação e robótica aplicadas à indústria,
especialmente no setor logístico e de armazém, tem vindo a impulsionar o desenvolvimento
de técnicas que otimizem o funcionamento e eficiência dos processos de
transporte e manuseio de mercadorias. Neste contexto, os Autonomous Mobile Robot
(AMR) ganham destaque pela sua capacidade de movimentação independente,
sem necessidade de adaptação do ambiente ao seu redor e por oferecerem soluções
inovadoras para tarefas repetitivas e de alto risco.
Este projeto aborda o desenvolvimento de uma solução capaz de aprimorar os
algoritmos de navegação utilizados em AMR, especificamente aplicados a um empilhador
autónomo. A necessidade de desenvolvimento deste projeto prende-se com
as limitações do veículo em reagir a alguns obstáculos inesperados e a ambientes de
dimensão reduzida. A solução proposta visa melhorar a adaptabilidade do AMR em
ambientes de armazém, onde é fulcral assegurar manobras em espaços estreitos e o
replaneamento de rotas alternativas pelo ambiente.
Para suprir estes objetivos, desenvolveram-se novos algoritmos de contorno de
obstáculos, com capacidade para planear e ajustar trajetórias de forma autónoma,
incorporando funcionalidades como marcha atrás para lidar com restrições de mobilidade.
Este algoritmo foi implementado em C++ e integrado na arquitetura ROS da
empresa, utilizando o simulador 3D Gazebo para testes preliminares. A estratégia
de desenvolvimento incluiu uma fase inicial de análise de algoritmos de planeamento
de trajetória e uma posterior, de implementação de técnicas compatíveis com mapeamento
2D e com as restrições cinemáticas da empilhadora.
Os resultados obtidos mostram que as soluções idealizadas melhoram significativamente
a capacidade de navegação do robô em ambientes complexos, funcionando
como uma abordagem inicial promissora para o problema, a ser aprofundado no
futuro. Desta forma, este trabalho contribui como um primeiro passo para o avanço
das tecnologias de robótica industrial, oferecendo soluções de planeamento de trajetória
alternativas que tornam a navegação mais fluída e autónoma.
The growing search for automation and robotics solutions applicable to the industry, especially in the logistics and warehouse sector, has been pushing for the development of techniques that optimize the operation and efficiency of transportation processes and the handling of goods. In this context, the Autonomous Mobile Robot (AMR) stands out as it’s capable of independent movement, without the need to adapt to its current environment, and because it offers innovative solutions for repetitive high risk tasks. This project addresses the development of a solution capable of substantially improving the navigation algorithms used for AMR, more specifically applied to an autonomous stacker. The need to develop this project is bound by the vehicle’s limitations in reacting to certain unexpected obstacles and small environments. The proposed solution aims to improve the AMR’s adaptability in warehouse environments, where it is vital to ensure manouverability in tight spaces and the replanning of alternate routes through that same environment. To achieve these goals, new obstacle avoidance algorithms were developed, capable of planning and adjusting trajectories autonomously, incorporating features such as reversing, to deal with mobility restrictions. This algorithm was implemented in C++ and integrated in the company’s ROS architecture, utilizing 3D Gazebo as a simulator for preliminary testing. The development strategy included an initial phase for the analysis of trajectory planning algorithms and a later one, to implement compatible techniques with 2D mapping and with the stacker’s kinematic constraints. The acquired results show that the idealized solutions significantly improve the robot’s ability to navigate complex environments, working as a promising initial approach to the problem, to be further explored in the future. With that said, this work is playing a part as the first step towards advancing industrial robotics technology, by offering alternative path planning solutions which make navigation both more fluid and autonomous.
The growing search for automation and robotics solutions applicable to the industry, especially in the logistics and warehouse sector, has been pushing for the development of techniques that optimize the operation and efficiency of transportation processes and the handling of goods. In this context, the Autonomous Mobile Robot (AMR) stands out as it’s capable of independent movement, without the need to adapt to its current environment, and because it offers innovative solutions for repetitive high risk tasks. This project addresses the development of a solution capable of substantially improving the navigation algorithms used for AMR, more specifically applied to an autonomous stacker. The need to develop this project is bound by the vehicle’s limitations in reacting to certain unexpected obstacles and small environments. The proposed solution aims to improve the AMR’s adaptability in warehouse environments, where it is vital to ensure manouverability in tight spaces and the replanning of alternate routes through that same environment. To achieve these goals, new obstacle avoidance algorithms were developed, capable of planning and adjusting trajectories autonomously, incorporating features such as reversing, to deal with mobility restrictions. This algorithm was implemented in C++ and integrated in the company’s ROS architecture, utilizing 3D Gazebo as a simulator for preliminary testing. The development strategy included an initial phase for the analysis of trajectory planning algorithms and a later one, to implement compatible techniques with 2D mapping and with the stacker’s kinematic constraints. The acquired results show that the idealized solutions significantly improve the robot’s ability to navigate complex environments, working as a promising initial approach to the problem, to be further explored in the future. With that said, this work is playing a part as the first step towards advancing industrial robotics technology, by offering alternative path planning solutions which make navigation both more fluid and autonomous.
Description
Keywords
Planeamento de trajetória ROS AMR Empilhadora Autónomo