mlsanityData to 14 Oct 2026

Safety of Embodied Navigation: A Survey

arXiv 2508.05855 ↗First posted 7 Aug 2025cs.AI cs.ROcited by 10

Not on the map. Why?

Abstract (arXiv)

As large language models (LLMs) continue to advance and gain influence, the development of embodied AI has accelerated, drawing significant attention, particularly in navigation scenarios. Embodied navigation requires an agent to perceive, interact with, and adapt to its environment while moving toward a specified target in unfamiliar settings. However, the integration of embodied navigation into critical applications raises substantial safety concerns. Given their deployment in dynamic, real-world environments, ensuring the safety of such systems is critical. This survey provides a comprehensive analysis of safety in embodied navigation from multiple perspectives, encompassing attack strategies, defense mechanisms, and evaluation methodologies. Beyond conducting a comprehensive examination of existing safety challenges, mitigation technologies, and various datasets and metrics that assess effectiveness and robustness, we explore unresolved issues and future research directions in embodied navigation safety. These include potential attack methods, mitigation strategies, more reliable evaluation techniques, and the implementation of verification frameworks. By addressing these critical gaps, this survey aims to provide valuable insights that can guide future research toward the development of safer and more reliable embodied navigation systems. Furthermore, the findings of this study have broader implications for enhancing societal safety and increasing industrial efficiency.

Key facts. Cited by 10 papers in the map (data through 14 Oct 2026). Builds on 40 of its 50 references that are in the map. First posted on arXiv 7 Aug 2025.

Builds on

40 of 50 references.

References · 40

  1. 2016Adversarial examples in the physical worldcited by 1,769
  2. 2023Jailbroken: How Does LLM Safety Training Fail?cited by 1,651
  3. 2022Inner Monologue: Embodied Reasoning through Planning with Language Modelscited by 1,081
  4. 2017Adversarial Patchcited by 517
  5. 2020Privacy and Robustness in Federated Learning: Attacks and Defensescited by 166
  6. 2018ShapeShifter: Robust Physical Adversarial Attack on Faster R-CNN Object Detectorcited by 159
  7. 2024BEHAVIOR-1K: A Human-Centered, Embodied AI Benchmark with 1,000 Everyday Activities and Realistic Simulationcited by 148
  8. 2024SafeAgentBench: A Benchmark for Safe Task Planning of Embodied LLM Agentscited by 116
  9. 2024GOAT-Bench: A Benchmark for Multi-Modal Lifelong Navigationcited by 94
  10. 2023A Survey of Safety and Trustworthiness of Large Language Models through the Lens of Verification and Validationcited by 93
  11. 2024Vision-and-Language Navigation Today and Tomorrow: A Survey in the Era of Foundation Modelscited by 79
  12. 2024Safeguarding Large Language Models: A Surveycited by 78
28 more
  1. 2021FCA: Learning a 3D Full-coverage Vehicle Camouflage for Multi-view Physical Adversarial Attackcited by 74
  2. 2017Adversarial Attacks Beyond the Image Spacecited by 67
  3. 2021Memory-Augmented Reinforcement Learning for Image-Goal Navigationcited by 66
  4. 2023LLM-Assist: Enhancing Closed-Loop Planning with Language-Based Reasoningcited by 59
  5. 2025EmbodiedEval: Evaluate Multimodal LLMs as Embodied Agentscited by 57
  6. 2024LoTa-Bench: Benchmarking Language-oriented Task Planners for Embodied Agentscited by 51
  7. 2023Physically Realizable Natural-Looking Clothing Textures Evade Person Detectors via 3D Modelingcited by 41
  8. 2024Exploring the Robustness of Decision-Level Through Adversarial Attacks on LLM-Based Embodied Modelscited by 33
  9. 2019Physical Adversarial Textures that Fool Visual Object Trackingcited by 31
  10. 2023ACTIVE: Towards Highly Transferable 3D Physical Camouflage for Universal and Robust Vehicle Evasioncited by 31
  11. 2022PatchZero: Defending against Adversarial Patch Attacks by Detecting and Zeroing the Patchcited by 30
  12. 2024EARBench: Towards Evaluating Physical Risk Awareness for Task Planning of Foundation Model-based Embodied AI Agentscited by 27
  13. 2023Towards Transferable Targeted 3D Adversarial Attack in the Physical Worldcited by 25
  14. 2021Generative Dynamic Patch Attackcited by 19
  15. 2023Towards Benchmarking and Assessing Visual Naturalness of Physical World Adversarial Attackscited by 18
  16. 2020SPAA: Stealthy Projector-based Adversarial Attacks on Deep Image Classifierscited by 17
  17. 2024ReALFRED: An Embodied Instruction Following Benchmark in Photo-Realistic Environmentscited by 16
  18. 2024Robot Collapse: Supply Chain Backdoor Attacks Against VLM-based Robotic Manipulationcited by 15
  19. 2022Adversarial Laser Spot: Robust and Covert Physical-World Attack to DNNscited by 11
  20. 2025From Screens to Scenes: A Survey of Embodied AI in Healthcarecited by 10
  21. 2023Reward Certification for Policy Smoothed Reinforcement Learningcited by 7
  22. 2022Consistent Attack: Universal Adversarial Perturbation on Embodied Vision Navigationcited by 6
  23. 2024Towards Physically Realizable Adversarial Attacks in Embodied Vision Navigationcited by 6
  24. 2024Understanding Impacts of Electromagnetic Signal Injection Attacks on Object Detectioncited by 5
  25. 2024Hijacking Vision-and-Language Navigation Agents with Adversarial Environmental Attackscited by 5
  26. 2023Embodied Laser Attack:Leveraging Scene Priors to Achieve Agent-based Robust Non-contact Attackscited by 4
  27. 2024Embodied Active Defense: Leveraging Recurrent Feedback to Counter Adversarial Patchescited by 4
  28. 2023Distributional Modeling for Location-Aware Adversarial Patchescited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 2; 2026: 8.

10 papers.

  1. 2026Safety in Embodied AI: A Survey of Risks, Attacks, and Defensescited by 7
  2. 2025CluCERT: Certifying LLM Robustness via Clustering-Guided Denoising Smoothingcited by 6
  3. 2026ForesightSafety Bench: A Frontier Risk Evaluation and Governance Framework towards Safe AIcited by 3
  4. 2026LABSHIELD: A Multimodal Benchmark for Safety-Critical Reasoning and Planning in Scientific Laboratoriescited by 3
  5. 2026ForesightSafety-VLA: A Unified Diagnostic Safety Benchmark for Vision-Language-Action Modelscited by 3
  6. 2025Measuring Physical-World Privacy Awareness of Large Language Models: An Evaluation Benchmarkcited by 2
  7. 2026EmbodiedGovBench: A Benchmark for Governance, Recovery, and Upgrade Safety in Embodied Agent Systemscited by 2
  8. 2026Security of World-Model-Based Embodied AI: A Lifecycle of Threats, Defenses, and Evaluationcited by 2
  9. 2026A Comprehensive Survey and Systematic Real-World Evaluation of Embodied Vision-and-Language Navigationcited by 1
  10. 2026AdvNav: Behavior-Guided Black-Box Adversarial Attacks on Vision-Language Navigationcited by 0