NavHarness improves lifelong navigation by remembering past searches
NavHarness: Towards Lifelong Embodied Navigation
Robotics
Summary
Robots that navigate spaces often forget what they learned in earlier trips, making repeated tasks harder. The authors created NavHarness, a system that helps these robots keep track of maps and previous attempts during navigation, updating their knowledge as they go. This memory is saved and used to improve future navigation, helping the robot learn from experience across different tasks and locations. Their method shows better success rates than treating each task independently, demonstrating improved long-term navigation.
What this means in practice
- •For robotics engineers: Implement navigation systems that improve over time by incorporating memory of past tasks and environment changes.
- •For home automation developers: Develop devices that navigate homes more reliably by remembering prior searches and adapting to new tasks.$Commercial implications: Enables commercial robotic assistants to perform repeated household tasks with improved accuracy using lifelong navigation memory.
Authors
Xunyi Zhao, Jian Zhou, Sihao Lin, Gengze Zhou, Zerui Li, Xinyu Yan, Jiajun Liu, Anton van den Hengel, Qi Wu
Abstract
Frontier models can now perform well on individual embodied navigation tasks through multi-round multimodal reasoning with simple tools. Across successive tasks, however, an agent must also rely on an evolving map and earlier search records, both of which may be incomplete or conflict with new observations. We present NavHarness, a training-free embodied harness towards lifelong navigation that makes memory processing part of the navigation loop. During navigation, its multi-round agentic session draws on maps, task records, and house knowledge, checking them against observations and recording corrections to guide its actions. NavHarness preserves this experience across fresh conversations for new tasks or recovery attempts, while outcome verification and run-end summaries support its later reuse. On GOAT-Bench, NavHarness improves s-SR over context-only independent sessions by 18.6 points with Astra and 22.6 with Opus 5. Using SLAM-estimated poses, NavHarness with GPT-6 Astra achieves state-of-the-art task success of 83.7 s-SR with 36.9 e-SR on GOAT-Bench and 85.9 s-SR on IR2R-CE. To understand these gains, we examine how experience is carried between sessions and find that structured recovery handovers outperform length-matched summaries. In extended deployments across houses, consolidation improves navigation beyond retaining maps and task records, with case studies showing how agents use earlier experience to interpret new goals, investigate unresolved questions, and resume failed searches. We suggest that progress towards lifelong navigation depends on how successive reasoning sessions build on prior experience, alongside improvements in single-task capability.