Development of an Automated Warehouse System Using Autonomous Mobile Robots
Mia, Md Rakib (2025)
2025
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2025120332046
https://urn.fi/URN:NBN:fi:amk-2025120332046
Tiivistelmä
In today’s dynamic and rapid supply chain, the need for flexible, efficient and affordable warehouse solutions has never been more needed particularly for the SME market which don’t often get exposed to high-end automation. Although automation models have been increasingly developed, a warehouse automation model that is affordable, practicable, and scalable is still largely lacking. This thesis fills this gap by the proposition of a planning framework for AMR systems with an integration of Autonomous Mobile Robots and the development of a dynamic simulation-based automated warehouse system solution for modern intralogistics.
The research was based on a Research-Based Development (RBD) approach. It started with a comprehensive literature review to understand present technologies, and to identify constraints of the conventional AGV systems, specifically their non-flexibility in our SME situations. Based on these results, the system has been implemented on Robot Operating System (ROS) and simulated in a warehouse in Webots, to model AMR movements, path planning (A*, DWA), SLAM-based localization and task coordination logic.
The architecture of the system was iteratively designed using an Agile process, and performance data from simulation testing enabled ongoing refinement of the system. Major functionality components including obstacle avoidance, multi-robot coordination, task scheduling and layout adaptability were thoroughly tested in tailor scenario settings, providing a close replication of real-world SME use cases and testbed constraints. Security, scalability and easy integration were given prime importance.
As a consequence, a working man-model was implemented, model and evaluated in simulation. The system shows high potential in decreasing the amount of manual labeling intervention, increasing the throughput and providing small and medium enterprises (SMEs) a more affordable way to intelligent warehouse automation. It closes not only some identified lacunae in existing literature, but also delivers an easily reproducible model for logistically operating real-world, scale-based, and simulation-validated deployment.
The research was based on a Research-Based Development (RBD) approach. It started with a comprehensive literature review to understand present technologies, and to identify constraints of the conventional AGV systems, specifically their non-flexibility in our SME situations. Based on these results, the system has been implemented on Robot Operating System (ROS) and simulated in a warehouse in Webots, to model AMR movements, path planning (A*, DWA), SLAM-based localization and task coordination logic.
The architecture of the system was iteratively designed using an Agile process, and performance data from simulation testing enabled ongoing refinement of the system. Major functionality components including obstacle avoidance, multi-robot coordination, task scheduling and layout adaptability were thoroughly tested in tailor scenario settings, providing a close replication of real-world SME use cases and testbed constraints. Security, scalability and easy integration were given prime importance.
As a consequence, a working man-model was implemented, model and evaluated in simulation. The system shows high potential in decreasing the amount of manual labeling intervention, increasing the throughput and providing small and medium enterprises (SMEs) a more affordable way to intelligent warehouse automation. It closes not only some identified lacunae in existing literature, but also delivers an easily reproducible model for logistically operating real-world, scale-based, and simulation-validated deployment.