aiqbench analysis
This is an embodied-AI robotics paper. It pairs a learned visual navigation policy (X-NavDP) with a topological memory and human-robot memory updating, so the learned component is central, though the contribution is mostly system design. The evidence is a single robot (Unitree Go2), one route, and 5 runs per condition, so the headline "superiority" claim is preliminary. Readers should look for the ablations and the full comparison details before drawing broader conclusions.
AVALON: Zero-Shot Indoor Navigation with Human-Seeded Topological Memory
arXiv:2610.21586v1 Announce Type: cross Abstract: Indoor service robots often enter unmapped buildings whose layouts and task constraints are already familiar to users. We address this human-familiar, robot-unseen setting with AVALON, a topological navigation framework organized around human-seeded layered spatial memory, node-verified topological execution, and task-triggered, evidence-gated human--robot memory co-updating. The robot has no prior on-site traversal or site-specific training. Deployment requires only a floor plan, place clues, and an initial topology-odometry alignment. Visual depth point-goal navigation leverages verified node arrivals to perform translational re-anchoring, without requiring a robot-constructed global SLAM map. Human--robot memory co-updating fuses complementary sources of information. Human operators specify task constraints that are unobservable via vision and report offline environmental modifications, whereas the robot provides physical-state evidence through active in-situ validation. Task relevance governs whether human-reported environmental changes undergo robot validation during the active mission. Field-collected perceptual evidence gates memory updates, maintaining an explicit separation between human-specified permissions and robot-observed traversability. Our experiments on a Unitree Go2 show that AVALON with X-NavDP completes a 12.81 m Intersection-to-Dining mission in 46.4 s (5/5) with no mapping, 40.9% less robot time than DimOS-Go2 (78.5 s, 4/5), while policy-only X-NavDP times out in all of the 5 runs. These demonstrate the superiority of the proposed topological navigation framework.
Abstract:Indoor service robots often enter unmapped buildings whose layouts and task constraints are already familiar to users. We address this human-familiar, robot-unseen setting with AVALON, a topological navigation framework organized around human-seeded layered spatial memory, node-verified topological execution, and task-triggered, evidence-gated human--robot memory co-updating. The robot has no prior on-site traversal or site-specific training. Deployment requires only a floor plan, place clues, and an initial topology-odometry alignment. Visual depth point-goal navigation leverages verified node arrivals to perform translational re-anchoring, without requiring a robot-constructed global SLAM map. Human--robot memory co-updating fuses complementary sources of information. Human operators specify task constraints that are unobservable via vision and report offline environmental modifications, whereas the robot provides physical-state evidence through active in-situ validation. Task relevance governs whether human-reported environmental changes undergo robot validation during the active mission. Field-collected perceptual evidence gates memory updates, maintaining an explicit separation between human-specified permissions and robot-observed traversability. Our experiments on a Unitree Go2 show that AVALON with X-NavDP completes a 12.81 m Intersection-to-Dining mission in 46.4 s (5/5) with no mapping, 40.9% less robot time than DimOS-Go2 (78.5 s, 4/5), while policy-only X-NavDP times out in all of the 5 runs. These demonstrate the superiority of the proposed topological navigation framework.
Robotics (cs.RO); Artificial Intelligence (cs.AI)
arXiv:2610.21586 [cs.RO]
(or arXiv:2610.21586v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2610.21586 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 299 words
Related coverage
HM3D-M3: A Benchmark for Multi-Agent, Multi-Modal, Multi-Object Goal Navigation
arXiv · Oct 15, 2026
Runway’s WorldPrompt and the Engineering of Real-Time Worlds
Latent Space · Sep 25, 2026
World Models and Yann LeCun’s Bet Against Large Language Models
Quantum Zeitgeist · Jul 17, 2026
DPed-VLN: A Benchmark for Socially Compliant Vision-and-Language Navigation in Dynamic Pedestrian Environments
arXiv · Sep 21, 2026