Is Semantic SLAM Ready for Embedded Systems ? A Comparative Survey

2025-05-18

AI Summary

Semantic visual SLAM deployment on embedded robotic platforms faces tight constraints in accuracy, efficiency, and power consumption. This work presents the first systematic, multi-metric evaluation of geometric semantic SLAM, NeRF-based, and 3D Gaussian Splatting architectures on edge hardware—specifically NVIDIA Jetson AGX Orin—assessing mapping accuracy, semantic segmentation quality, memory footprint (≤2.1 GB), power draw (≤25 W), and end-to-end latency. Results demonstrate that geometric semantic SLAM achieves the optimal trade-off between localization accuracy and real-time performance, satisfying edge deployment requirements; in contrast, NeRF and Gaussian Splatting deliver finer semantic representation but exceed resource budgets by significant margins. Building on these findings, we propose a novel algorithm–hardware co-design paradigm, providing empirical validation and concrete design guidelines for deploying lightweight semantic SLAM in resource-constrained robotics applications.

Technology Category

• Machine Learning: Hardware-aware ML
• Intelligent Robots: Localization, Mapping, and Navigation
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics

Abstract

In embedded systems, robots must perceive and interpret their environment efficiently to operate reliably in real-world conditions. Visual Semantic SLAM (Simultaneous Localization and Mapping) enhances standard SLAM by incorporating semantic information into the map, enabling more informed decision-making. However, implementing such systems on resource-limited hardware involves trade-offs between accuracy, computing efficiency, and power usage. This paper provides a comparative review of recent Semantic Visual SLAM methods with a focus on their applicability to embedded platforms. We analyze three main types of architectures - Geometric SLAM, Neural Radiance Fields (NeRF), and 3D Gaussian Splatting - and evaluate their performance on constrained hardware, specifically the NVIDIA Jetson AGX Orin. We compare their accuracy, segmentation quality, memory usage, and energy consumption. Our results show that methods based on NeRF and Gaussian Splatting achieve high semantic detail but demand substantial computing resources, limiting their use on embedded devices. In contrast, Semantic Geometric SLAM offers a more practical balance between computational cost and accuracy. The review highlights a need for SLAM algorithms that are better adapted to embedded environments, and it discusses key directions for improving their efficiency through algorithm-hardware co-design.

Problem

• Evaluating Semantic SLAM feasibility for embedded systems
• Comparing accuracy and efficiency of SLAM architectures
• Identifying resource-efficient solutions for constrained hardware

Innovation

• Comparative review of Semantic Visual SLAM methods
• Analyzes Geometric SLAM, NeRF, and Gaussian Splatting
• Highlights need for embedded-optimized SLAM algorithms