Comparison of 2D & 3D LiDAR SLAM Algorithms Based on Performance Metrics

Abstract

Robotics has evolved significantly from its industrial origins to encompass advanced autonomous mobile systems capable of operating in complex and unstructured environments. A key catalyst for this rapid advancement in autonomous navigation has been the evolution of robotic localization and mapping capabilities, with SLAM algorithms at the forefront. This paper presents a comparative analysis of 2D and 3D LiDAR-based Simultaneous Localization and Mapping (SLAM) techniques, evaluating their performance against key metrics such as CPU load, algorithm speed, RAM usage, and map accuracy, leveraging mapping and computation profilers for detailed insights. The findings support enhancing the adaptability of SLAM applications, including high real-time processing and precise mapping for indoor robotics, as well as optimized speed, memory efficiency, and edge computing integration for mobile AGV applications. This analysis serves as a foundation for selecting the most suitable SLAM algorithm for specific use cases, guided by essential performance metrics for practical implementation.