Jul 2026· International Conference on Edge Computing [Services Society]· pp. 171-181· 0 citations· 42 references
TL;DR
Lang2Graph is presented, an experimental framework for indoor topological graph inference from natural-language navigational instructions that isolates four governing factors: instruction structure, metadata clarity, prompting strategy, and model size and reasoning capability, and Reasoning-aligned open-source models of moderate scale (14B parameters) outperform larger proprietary models on the most challenging instruction categories.
Abstract
Edge-deployed systems such as autonomous robots, AR/XR devices, and emergency-response handhelds require accurate indoor topological representations, yet existing sensor-based and expert-curated mapping methods are impractical for crowd-sourced, resource-constrained deployment. Additionally, current large language model (LLM) approaches to indoor topology inference lack a systematic framework for evaluating the factors that govern performance. We present Lang2Graph, an experimental framework for indoor topological graph inference from natural-language navigational instructions that isolates four governing factors: instruction structure, metadata clarity, prompting strategy, and model size and reasoning capability. We propose the Independent Prompt Executor (IPE), a prompting strategy that decomposes graph construction into independent per-instruction reasoning steps, preventing error propagation. To support factor-level evaluation, we develop a fully synthetic dataset and an augmented benchmark (R2R-AUG) covering idealized and real-world conditions. Our evaluation across multiple model families show that structured instructions, clear metadata, and IPE improve precision, recall, and F1 by 44%, 45%, and 47%, respectively. Reasoning-aligned open-source models of moderate scale (14B parameters) outperform larger proprietary models on the most challenging instruction categories, indicating that reliable indoor topology inference is achievable without cloud-scale dependencies and establishes a viable path toward on-device edge deployment.
Foundation-model-based vision-language navigation (VLN) has advanced autonomous robot navigation by enabling robots to interpret natural-language instructions, identify semantic goals, and follow user-specified behavioral rules. However, existing VLN systems rely heavily on cloud-hosted foundation models for language understanding and semantic grounding, limiting their applicability where network connectivity is unavailable and reliable metric goal localization is required. Although recent small language models (SLMs) enable fully onboard inference, their suitability for navigation instruction decomposition has not been systematically evaluated. This paper makes three contributions toward fully onboard VLN for outdoor environments. First, we present the first systematic benchmark of 17 edge-deployable SLMs against 4 online APIs for robotic navigation instruction decomposition, evaluating accuracy and latency on human-annotated instructions across three computing platforms and providing practical guidance for selecting onboard language models. Second, we propose a lightweight hybrid semantic-geometric goal localization framework that combines open-vocabulary object detection, prompted segmentation, and LiDAR geometry to estimate metric goals, while maintaining visual bearing guidance when reliable geometric observations are unavailable. Third, we integrate these advances into Edge-BehAV, a fully onboard extension of the BehAV architecture that enables cloud-independent behavior-guided navigation. Experimental results show that the best offline SLM matches the instruction decomposition performance of the strongest cloud API while running approximately 9x faster and without network connectivity. The proposed goal localization framework reduces mean goal-distance error from 2.05 m to 0.20 m at lower computational cost, and the complete system succeeds in 31 of 32 closed-loop outdoor trials.
Ali Salmasi, Xianjia Yu, Tomi Westerlund· 0 citations
This framework combines simultaneous localization and mapping (SLAM), visual-language feature extraction, incremental semantic and instance label fusion, and spatial graph construction to enable a construction robot navigation framework that supports open-vocabulary language queries.
Charles M. Raines, I. Fernandez, Mandy Sun et al.· Journal of computing in civi...· 0 citations
Instance-Enriched Semantic Maps is proposed, a unified framework with three key contributions: instance-level two-and-a-half-dimensional rich information mapping, storage-efficient semantic representation that achieves approximately 96% reduction compared to three-dimensional scene-graph approaches while preserving sufficient spatial information for navigation, and robust query processing via LLM-based target selection.
J. Hong, Eunae Kang, Sanghyun Kim et al.· Engineering applications of...· 0 citations
As Multimodal Large Language Models (MLLMs) are increasingly deployed in decision-critical pipelines such as robotics, embodied AI, and safety monitoring, the opacity of their spatial judgments limits operator trust and auditability. MLLMs demonstrate strong reasoning but often struggle with fine-grained spatial understanding and object hallucination. Prior work, ByDeWay, introduced Layered-Depth-Based Prompting (LDP), a training-free framework that mitigates hallucinations by structuring prompts using monocular depth estimation. However, coarse depth layering falls short in resolving object-to-object spatial relationships within the same geometric plane, such as projective ("left of","above") and topological ("inside","touching") relations. We propose ByDeWay-V2, which integrates explicit spatial relational context alongside depth cues, expressed as human-readable predicates that serve as auditable evidence for downstream decision support. Using an open-vocabulary object detector (YOLO-World-L), our framework computes pairwise geometric relations between detected objects and injects them as structured spatial predicates into the MLLM prompt, bridging 3D scene depth and 2D spatial semantics without any training. We evaluate ByDeWay-V2 on the Visual Spatial Reasoning (VSR) and BLINK benchmarks across multiple MLLMs, with hallucination grounding assessed via POPE. On the BLINK spatial subset, ByDeWay-V2 achieves a 46 percent relative F1 improvement over LDP for Qwen2.5-VL, and recovers BLIP-Base's spatial reasoning on VSR from near-random performance to a competitive F1 of 0.53. Our lightest configuration operates under a strict 40-token context budget on CPU, showing the framework's suitability for resource-constrained, real-time decision-support settings.
Piyush Jain, Kousik Dasgupta, Rajarshi Roy et al.· 0 citations
Vision-and-Language Navigation (VLN) has progressively expanded from indoor to outdoor environments. However, existing outdoor VLN datasets still rely on fixed discrete topological graphs for construction. It fails to align with the rapidly changing real-world outdoor environments and impedes the sim-to-real transfer of VLN agents. To address this limitation, we propose DaViNCi (\textbf{D}yn\textbf{a}mic \textbf{Vi}sion-and-Language \textbf{N}avigation in \textbf{C}ont\textbf{i}nuous Environment), the first outdoor VLN dataset that simultaneously introduces both continuous and dynamic factors. The agent not only moves in the outdoor environment using continuous actions but is also required to handle unpredictable dynamic elements. The dataset encompasses six distinct maps with a total of 6,933 trajectories. Through comprehensive comparative experiments, we find that the success rate on DaViNCi decreased by more than 10\% in discrete environments compared to previous datasets. And there is an even greater decline in continuous settings, demonstrating the challenge of DaViNCi. Furthermore, we clarify the impact of action granularity and dynamic elements. These results demonstrate the practical value of DaViNCi in advancing outdoor VLN toward more realistic environments. The website is https://xzh0312.github.io/DaViNCi/.
Zihao Xie, Pingrui Lai, Yitong Wu et al.· 0 citations
Recent Multimodal Large Language Models (MLLMs) struggle to bridge the representational gap between 2D semantic understanding and 3D spatial geometry. Existing 3D-aware models either rely on costly 3D-specific data or utilize RGB-only inputs with heuristic sampling and monolithic, shallow fusion, which respectively disrupt essential spatiotemporal connectivity and induce modality contention across diverse spatial tasks. To overcome these bottlenecks, we introduce SpaR3D-MoE, an end-to-end framework that enables adaptive spatial reasoning by equipping MLLMs with geometry-aware capabilities from only sparse RGB inputs. First, we propose an adaptive spatiotemporal manifold sampling mechanism that constructs a geometry-aware spatiotemporal graph to extract informative keyframes, effectively mitigating sequence redundancy while preserving the scene's topological connectivity. Second, we introduce the heterogeneous geometry-inductive Mixture-of-Experts driven by an instruction-pose aware router, which adaptively routes multimodal tokens to specialized experts, resolving the cross-modal contention inherent in monolithic fusion. Extensive experiments on VSI-Bench, ScanQA, and SQA3D demonstrate that our method achieves state-of-the-art performance. Notably, SpaR3D-MoE achieves the highest average score of 63.5 on VSI-Bench, outperforming the strongest baseline by 7.8 absolute points, alongside relative improvements of 35.4% and 51.4% in Route Plan and Relative Direction tasks, respectively.
Haida Feng, Hao Wei, Haolin Wang et al.· 0 citations