When Engineering Outruns Intelligence: Rethinking Instruction-Guided Navigation
On the map: LLM reasoning → Spatial and embodied reasoning
Abstract (arXiv)
Recent ObjectNav systems credit large language models (LLMs) for sizable zero-shot gains, yet it remains unclear how much comes from language versus geometry. We revisit this question by re-evaluating an instruction-guided pipeline, InstructNav, under a detector-controlled setting and introducing two training-free variants that only alter the action value map: a geometry-only Frontier Proximity Explorer (FPE) and a lightweight Semantic-Heuristic Frontier (SHF) that polls the LLM with simple frontier votes. Across HM3D and MP3D, FPE matches or exceeds the detector-controlled instruction follower while using no API calls and running faster; SHF attains comparable accuracy with a smaller, localized language prior. These results suggest that carefully engineered frontier geometry accounts for much of the reported progress, and that language is most reliable as a light heuristic rather than an end-to-end planner. Code available at: https://github.com/matinaghaei/instructnav-scrutinized
Key facts. No citing papers in the map yet (data through 13 Oct 2026). Placed in Spatial and embodied reasoning, one of 17 directions in LLM reasoning. Builds on 19 of its 32 references that are in the map. First posted on arXiv 26 Jul 2025.
Builds on
19 of 32 references.
Closest ancestors · 1
Same direction, newest first
Other references · 18
- 2022Do As I Can, Not As I Say: Grounding Language in Robotic Affordancescited by 2,321
- 2023Voyager: An Open-Ended Embodied Agent with Large Language Modelscited by 2,135
- 2017Matterport3D: Learning from RGB-D Data in Indoor Environmentscited by 1,360
- 2021Habitat-Matterport 3D Dataset (HM3D): 1000 Large-scale 3D Environments for Embodied AIcited by 583
- 2018On Evaluation of Embodied Navigation Agentscited by 531
- 2020Object Goal Navigation using Goal-Oriented Semantic Explorationcited by 421
- 2023On the Planning Abilities of Large Language Models : A Critical Investigationcited by 310
- 2023Habitat 3.0: A Co-Habitat for Humans, Avatars and Robotscited by 221
- 2023VLFM: Vision-Language Frontier Maps for Zero-Shot Semantic Navigationcited by 201
- 2022CoWs on Pasture: Baselines and Benchmarks for Language-Driven Zero-Shot Object Navigationcited by 194
- 2022ZSON: Zero-Shot Object-Goal Navigation using Multimodal Goal Embeddingscited by 177
- 2024InstructNav: Zero-shot System for Generic Instruction Navigation in Unexplored Environmentcited by 152
6 more
- 2023ESC: Exploration with Soft Commonsense Constraints for Zero-shot Object Navigationcited by 126
- 2023L3MVN: Leveraging Large Language Models for Visual Target Navigationcited by 118
- 2023Navigation with Large Language Models: Semantic Guesswork as a Heuristic for Planningcited by 88
- 2023General Object Foundation Model for Images and Videos at Scalecited by 71
- 2024VoroNav: Voronoi-based Zero-shot Object Navigation with Large Language Modelcited by 55
- 2025FrontierNet: Learning Visual Cues to Explorecited by 11
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026R2F: Repurposing Ray Frontiers for LLM-free Object Navigation
- 2026LLMs for Text-Based Exploration and Navigation Under Partial Observability
- 2026Exploring Bottlenecks in VLM-LLM Navigation: How 3D Scene Understanding Capability Impacts Zero-Shot VLN
- 2024Open-Nav: Exploring Zero-Shot Vision-and-Language Navigation in Continuous Environment with Open-Source LLMs
- 2024Balancing Performance and Efficiency in Zero-shot Robotic Navigation
- 2025Run, Ruminate, and Regulate: A Dual-process Thinking System for Vision-and-Language Navigation
- 2026AdaGeoVLN: Selective Geometry Across Representation Depth and Navigation Time for Vision-Language Navigation ↗
- 2026LocalNav: Distilling Frontier VLMs and Embodied RL for On-Device Object Goal Navigation