Review of Autonomous SLAM Mapping and Navigation for Unknown Environment Exploration

Abstract

Simultaneous Localization and Mapping (SLAM) allows autonomous robots to navigate and explore unknown environments by constructing a map while positioning their location. Mapping is essential for real-time decision-making, obstacle avoidance, and flawless movement of an autonomous robot. There are two primary techniques for performing SLAM: Visual SLAM, which uses a camera to detect surrounding features, and LiDAR SLAM, which relies on a laser sensor for distance measurement. Visual SLAM is typically used in well-lit environments with rich surface textures and struggles in low-lit or featureless areas. LiDAR SLAM, on the other hand, produces higher accuracy and reliability in mapping with or without the presence of light but requires more computational power. This paper comprehensively reviews Visual SLAM and LiDAR SLAM techniques, analyzing their advantages, limitations, and applications. The study highlights key SLAM implementations across different robotic platforms, evaluating their localization accuracy, computational efficiency, and adaptability to dynamic environments. Furthermore, the integration of SLAM with the Robot Operating System (ROS) is explored, emphasizing its role in optimizing mapping, sensor fusion, and autonomous navigation. The findings provide insights into the most effective SLAM strategies for various application domains, paving the way for future advancements in autonomous robotic systems.