Simulator for autonomous robot navigation

• Main Author: Ng, Zheng Jie
• Other Authors: Lam Siew Kei
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Computer and Information Science; Engineering; Simulator
• Online Access: https://hdl.handle.net/10356/175137
• Institution: Nanyang Technological University

Simultaneous Localization and Mapping (SLAM), a fundamental aspect of robotics and autonomous navigation systems, is comprised of two essential components localization and mapping. Localization involves the ability to navigate and determine the position of a robot or device within an unfamiliar environment, while mapping pertains to the creation and maintenance of a representation of the environment. Currently, the prevailing method for localization heavily relies on Global Positioning System (GPS) sensors. However, the effectiveness of GPS is often constrained to scenarios with a clear view of the sky, and it introduces significant errors when used for indoor navigation, underground exploration, or in densely built urban areas with tall buildings [1]. This limitation has spurred the exploration of alternative solutions such as Visual-SLAM. Visual-SLAM presents a promising alternative by harnessing visual information captured through cameras for localization and mapping purposes. Unlike GPS, visual-based approaches are not reliant on external signals. They can thus operate effectively in GPS-denied environments, making them particularly suited for indoor navigation, underground exploration, and autonomous vehicles navigating urban canyons [1]. The versatility of Visual-SLAM extends beyond robotics; it finds applications in augmented reality, virtual reality, and indoor positioning systems. The proposed project aims to develop a modular Graphical User Interface (GUI) tailored specifically for Visual-SLAM applications. This GUI will facilitate the visualization and analysis of various real-time Visual-SLAM algorithms, providing users with insights into their performance under different conditions. The GUI's modularity will enable easy integration with different Visual-SLAM algorithms and frameworks, fostering collaboration and innovation. Leveraging the capabilities of Gazebo GUI and the Robot Operating System (ROS), the project aims to study a user-friendly interface that simplifies the deployment and evaluation of Visual-SLAM solutions across diverse robotic platforms and simulation environments. Through this initiative, the project aims to accelerate research and development in Visual-SLAM, paving the way for enhanced navigation capabilities in robotics, augmented reality applications, and beyond