OptiSight: Bridging Semantic Reasoning and Geometric Control for Embodied Navigation
On the map: LLM reasoning → Spatial and embodied reasoning
Abstract (arXiv)
Autonomous indoor navigation requires both semantic understanding and precise geometric control. We propose OptiSight, a hybrid framework that combines Vision-Language Model reasoning with deterministic visual servoing through a finite-state Chain-of-Thought architecture. Grounded-SAM localizes open-vocabulary targets, while camera projection geometry converts visual observations into navigation commands without requiring dense mapping. The VLM is queried only at key decision points, reducing computational overhead while geometric control handles continuous navigation. Experiments in AI Habitat demonstrate reliable zero-shot navigation across diverse indoor scenarios, including obstacle avoidance and semantic ambiguity, while operating within an 8~GB VRAM budget. The source code is available at https://github.com/avanalperen/OptiSight-Python-Multimodal-CoT-for-Visual-Reasoning.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Spatial and embodied reasoning, one of 17 directions in LLM reasoning. Builds on 14 of its 29 references that are in the map. First posted on arXiv 24 Aug 2026.
Builds on
14 of 29 references.
Closest ancestors · 5
Same direction, newest first
- 2026FantasyVLN: Unified Multimodal Chain-of-Thought Reasoning for Vision-Language Navigation20 Jan
- 2025Logic-Guided Socially-aware Robot Navigation World Model27 Oct
- 2025CoT-VLA: Visual Chain-of-Thought Reasoning for Vision-Language-Action Models27 Mar
- 2024Robotic Control via Embodied Chain-of-Thought Reasoning11 Jul
- 2024NavCoT: Boosting LLM-Based Vision-and-Language Navigation via Learning Disentangled Reasoning12 Mar
Other references · 9
- 2024Grounded SAM: Assembling Open-World Models for Diverse Visual Taskscited by 929
- 2026Qwen3.5-Omni Technical Reportcited by 174
- 2024Uni-NaVid: A Video-based Vision-Language-Action Model for Unifying Embodied Navigation Taskscited by 156
- 2025Nav-R1: Reasoning and Navigation in Embodied Scenescited by 38
- 2025MSGNav: Unleashing the Power of Multi-modal 3D Scene Graph for Zero-Shot Embodied Navigationcited by 13
- 2025Imaginative World Modeling with Scene Graphs for Embodied Agent Navigationcited by 6
- 2026ALRM: Agentic LLM for Robotic Manipulationcited by 6
- 2026Habitat-GS: A High-Fidelity Navigation Simulator with Dynamic Gaussian Splattingcited by 5
- 2026SceneGraphVLM: Dynamic Scene Graph Generation from Video with Vision-Language Modelscited by 1
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026AnchorVLN: Geometry-Anchored Vision-Language Grounding Reasoning for Open-Vocabulary Navigation
- 2026IROS: A Dual-Process Architecture for Real-Time VLM-Based Indoor Navigation
- 2026OmniVLN: Omnidirectional 3D Perception and Token-Efficient LLM Reasoning for Visual-Language Navigation across Air and Ground Platforms
- 2026CapNav: Benchmarking Vision Language Models on Capability-conditioned Indoor Navigation
- 2026A Semantic Autonomy Framework for VLM-Integrated Indoor Mobile Robots: Hybrid Deterministic Reasoning and Cross-Robot Adaptive Memory
- 2026A Deployable Embodied Vision-Language Navigation System with Hierarchical Cognition and Context-Aware Exploration ↗
- 2026Bridging the 2D-3D Gap: A Hierarchical Semantic-Geometric Map for Vision Language Navigation
- 2026SysNav: Multi-Level Systematic Cooperation Enables Real-World, Cross-Embodiment Object Navigation