Semantic Mapping in Indoor Embodied AI -- A Survey on Advances, Challenges, and Future Directions
Reviewed by Pith T0 review T1 audit T2 compute T3 formal T4 kernel pith:5VN2OLFErecord.jsonopen to challenge →
read the original abstract
Intelligent embodied agents (e.g. robots) need to perform complex semantic tasks in unfamiliar environments. Among many skills that the agents need to possess, building and maintaining a semantic map of the environment is most crucial in long-horizon tasks. A semantic map captures information about the environment in a structured way, allowing the agent to reference it for advanced reasoning throughout the task. While existing surveys in embodied AI focus on general advancements or specific tasks like navigation and manipulation, this paper provides a comprehensive review of semantic map-building approaches in embodied AI, specifically for indoor navigation. We categorize these approaches based on their structural representation (spatial grids, topological graphs, dense point-clouds or hybrid maps) and the type of information they encode (implicit features or explicit environmental data). We also explore the strengths and limitations of the map building techniques, highlight current challenges, and propose future research directions. We identify that the field is moving towards developing open-vocabulary, queryable, task-agnostic map representations, while high memory demands and computational inefficiency still remaining to be open challenges. This survey aims to guide current and future researchers in advancing semantic mapping techniques for embodied AI systems.
This paper has not been read by Pith yet.
Forward citations
Cited by 7 Pith papers
-
FrameVGGT: Coherence-Preserving Memory for Bounded Streaming Geometry
FrameVGGT replaces token-level KV retention with frame-level segments and prototypes to bound memory while preserving geometric coherence in streaming VGGT.
-
Vision-Language Model Reasoning for Contextual Semantic Mapping in Intralogistics
A zero-shot pipeline using SLAM, SAM segmentation, clustering and VLM multi-view reasoning produces semantic maps with object class and movability labels, reporting 98.93% mIoU and 89.17% mAcc on intralogistics data.
-
Make Your VLA More Robust Without More Data By Interleaving Motion Planning
MPVI interleaves model-based motion planning with VLAs via VLM completion checking to achieve 113% higher task progress on BEHAVIOR-1K without extra data.
-
Semantic-Aware Guided Drone Exploration for Language-Conditioned 3D Indoor Mapping
SAGE integrates CLIP embeddings into FALCON via object storage, temporal projection cache, object frontiers, and a bounded semantic-geometric cost, yielding 9-26x faster object discovery than FTU in simulation and hig...
-
Vision-Language Model Reasoning for Contextual Semantic Mapping in Intralogistics
A pipeline combining SLAM, SAM, instance clustering, and VLM multi-view reasoning produces contextual semantic maps with 98.93% mIoU semantic accuracy and 89.17% mAcc movability estimation in intralogistics.
-
Vision-and-Language Navigation for UAVs: Progress, Challenges, and a Research Roadmap
A survey of UAV vision-and-language navigation that establishes a methodological taxonomy, reviews resources and challenges, and proposes a forward-looking research roadmap.
-
When Multi-Robot Systems Meet Agentic AI:Towards Embodied Collective Intelligence
Introduces Embodied Collective Intelligence (ECI) as a conceptual framework for multi-robot systems with Co-Perception, Co-Action, and Co-Evolution, illustrated by a limited navigation study on shared memory inheritance.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.