Evaluating SLAM based Scene Construction for Resource-Constrained Platforms

Simultaneous Localization and Mapping (SLAM) plays a crucial role in enabling autonomous navigation across diverse environments, including indoor, outdoor, and industrial settings. However, implementing SLAM on low-end systems presents significant challenges due to computational constraints, sensor limitations, and real-time processing requirements. This paper evaluates various SLAM methodologies with a focus on their feasibility for resource-constrained platforms. Specifically, we analyze the trade-offs between accuracy, computational efficiency, and hardware requirements, comparing geometric SLAM approaches with modern semantic-based and sensor fusion techniques. . Experimental results using benchmark datasets demonstrate the impact of algorithmic optimizations and sensor fusion strategies on SLAM performance in low-end systems. Our findings provide insights into selecting and optimizing SLAM algorithms for real-world applications where computational resources are limited.