mlsanityData to 7 Oct 2026

The Surprising Effectiveness of Visual Odometry Techniques for Embodied PointGoal Navigation

arXiv 2108.11550 ↗First posted 26 Aug 2021cs.CV cs.AI cs.LGcited by 27

Not on the map. Why?

Abstract (arXiv)

It is fundamental for personal robots to reliably navigate to a specified goal. To study this task, PointGoal navigation has been introduced in simulated Embodied AI environments. Recent advances solve this PointGoal navigation task with near-perfect accuracy (99.6% success) in photo-realistically simulated environments, assuming noiseless egocentric vision, noiseless actuation, and most importantly, perfect localization. However, under realistic noise models for visual sensors and actuation, and without access to a "GPS and Compass sensor," the 99.6%-success agents for PointGoal navigation only succeed with 0.3%. In this work, we demonstrate the surprising effectiveness of visual odometry for the task of PointGoal navigation in this realistic setting, i.e., with realistic noise models for perception and actuation and without access to GPS and Compass sensors. We show that integrating visual odometry techniques into navigation policies improves the state-of-the-art on the popular Habitat PointNav benchmark by a large margin, improving success from 64.5% to 71.7% while executing 6.4 times faster.

Key facts. Cited by 27 papers in the map (data through 7 Oct 2026). Builds on 45 of its 58 references that are in the map. First posted on arXiv 26 Aug 2021.

Builds on

45 of 58 references.

References · 45

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  3. 2017Proximal Policy Optimization Algorithmscited by 14,441
  4. 2015Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shiftcited by 6,979
  5. 2015High-Dimensional Continuous Control Using Generalized Advantage Estimationcited by 1,916
  6. 2018Group Normalizationcited by 1,514
  7. 2019SuperGlue: Learning Feature Matching with Graph Neural Networkscited by 1,437
  8. 2017Matterport3D: Learning from RGB-D Data in Indoor Environmentscited by 1,354
  9. 2017SuperPoint: Self-Supervised Interest Point Detection and Descriptioncited by 1,313
  10. 2019Habitat: A Platform for Embodied AI Researchcited by 1,163
  11. 2017Vision-and-Language Navigation: Interpreting visually-grounded navigation instructions in real environmentscited by 1,013
  12. 2019The Replica Dataset: A Digital Replica of Indoor Spacescited by 885
33 more
  1. 2017AI2-THOR: An Interactive 3D Environment for Visual AIcited by 850
  2. 2019ALFRED: A Benchmark for Interpreting Grounded Instructions for Everyday Taskscited by 773
  3. 2018On Evaluation of Embodied Navigation Agentscited by 526
  4. 2018Gibson Env: Real-World Perception for Embodied Agentscited by 477
  5. 2020Beyond the Nav-Graph: Vision-and-Language Navigation in Continuous Environmentscited by 429
  6. 2019DD-PPO: Learning Near-Perfect PointGoal Navigators from 2.5 Billion Framescited by 352
  7. 2017Embodied Question Answeringcited by 335
  8. 2020Learning to Explore using Active Neural SLAMcited by 308
  9. 2019Vision-and-Dialog Navigationcited by 246
  10. 2020ObjectNav Revisited: On Evaluation of Embodied Agents Navigating to Objectscited by 240
  11. 2017Cognitive Mapping and Planning for Visual Navigationcited by 209
  12. 2017DeepVO: Towards End-to-End Visual Odometry with Deep Recurrent Convolutional Neural Networkscited by 173
  13. 2015Modelling Uncertainty in Deep Learning for Camera Relocalizationcited by 133
  14. 2019DeepPruner: Learning Efficient Stereo Matching via Differentiable PatchMatchcited by 114
  15. 2018Building Generalizable Agents with a Realistic and Rich 3D Environmentcited by 109
  16. 2020Occupancy Anticipation for Efficient Exploration and Navigationcited by 104
  17. 2019Embodied Question Answering in Photorealistic Environments with Point Cloud Perceptioncited by 103
  18. 2020A Survey on Deep Learning Techniques for Stereo-based Depth Estimationcited by 90
  19. 2019Interactive Gibson Benchmark (iGibson 0.5): A Benchmark for Interactive Navigation in Cluttered Environmentscited by 89
  20. 2017Camera Relocalization by Computing Pairwise Relative Poses Using Convolutional Neural Networkcited by 79
  21. 2017MINOS: Multimodal Indoor Simulator for Navigation in Complex Environmentscited by 66
  22. 2021Cooperative Exploration for Multi-Agent Deep Reinforcement Learningcited by 57
  23. 2020A Cordial Sync: Going Beyond Marginal Policies for Multi-Agent Embodied Taskscited by 52
  24. 2019PIC: Permutation Invariant Critic for Multi-Agent Deep Reinforcement Learningcited by 46
  25. 2019Two Body Problem: Collaborative Visual Task Completioncited by 45
  26. 2019PyRobot: An Open-source Robotics Framework for Research and Benchmarkingcited by 39
  27. 2020A Survey on Deep Learning for Localization and Mapping: Towards the Age of Spatial Machine Intelligencecited by 35
  28. 2020Seeing the Un-Scene: Learning Amodal Semantic Maps for Room Navigationcited by 34
  29. 2020Integrating Egocentric Localization for More Realistic Point-Goal Navigation Agentscited by 27
  30. 2021Semantic Tracklets: An Object-Centric Representation for Visual Multi-Agent Reinforcement Learningcited by 18
  31. 2017Generic 3D Representation via Pose Estimation and Matchingcited by 17
  32. 2020High-Throughput Synchronous Deep RLcited by 9
  33. 2019Learning Your Way Without Map or Compass: Panoramic Target Driven Visual Navigationcited by 7

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 1; 2022: 10; 2023: 4; 2024: 6; 2025: 2; 2026: 4.

27 papers.

Sort:

By area:

  1. 2023ETPNav: Evolving Topological Planning for Vision-Language Navigation in Continuous Environmentscited by 165
  2. 2024GOAT-Bench: A Benchmark for Multi-Modal Lifelong Navigationcited by 93
  3. 2022Cross-modal Map Learning for Vision and Language Navigationcited by 91
  4. 2022Housekeep: Tidying Virtual Households using Commonsense Reasoningcited by 57
  5. 2021Coupling Vision and Proprioception for Navigation of Legged Robotscited by 46
  6. 20223D-Aware Object Goal Navigation via Simultaneous Exploration and Identificationcited by 46
  7. 2025OctoNav: Towards Generalist Embodied Navigationcited by 33
  8. 2022Episodic Memory Question Answeringcited by 32
  9. 2022Is Mapping Necessary for Realistic PointGoal Navigation?cited by 32
  10. 2024DynaMem: Online Dynamic Spatio-Semantic Memory for Open World Mobile Manipulationcited by 32
  11. 2022Uncertainty-driven Planner for Exploration and Navigationcited by 31
  12. 2022PEANUT: Predicting and Navigating to Unseen Targetscited by 22
15 more
  1. 2023Resilient Legged Local Navigation: Learning to Traverse with Compromised Perception End-to-Endcited by 16
  2. 2024Memory-based Adapters for Online 3D Scene Perceptioncited by 13
  3. 2024FloNa: Floor Plan Guided Embodied Visual Navigationcited by 10
  4. 2025LookOut: Real-World Humanoid Egocentric Navigationcited by 9
  5. 2022Break and Make: Interactive Structural Understanding Using LEGO Brickscited by 8
  6. 2022Object Manipulation via Visual Target Localizationcited by 6
  7. 2023Modality-invariant Visual Odometry for Embodied Visioncited by 6
  8. 2023Robust Navigation with Cross-Modal Fusion and Knowledge Transfercited by 2
  9. 2026LangMap: A Human-Verified Benchmark for Hierarchical Open-Vocabulary Goal Navigationcited by 2
  10. 2026POINav: Benchmarking and Enhancing Final-Meters Arrival in Real-World Vision-Language Navigationcited by 2
  11. 2022Unsupervised Visual Odometry and Action Integration for PointGoal Navigation in Indoor Environmentcited by 1
  12. 2024The Empirical Impact of Forgetting and Transfer in Continual Visual Odometrycited by 1
  13. 2024MPVO: Motion-Prior based Visual Odometry for PointGoal Navigationcited by 0
  14. 2026Insect-inspired Visual Point-goal Navigationcited by 0
  15. 2026Macro-Action Topological Navigation under Noisy Localization using Reinforcement Learningcited by 0