analiza aiqbench
RoboAtlas combines frontier-based exploration with vision-language model reasoning, using a bandit to decide when to switch from exploring to semantically guided navigation. The reported gains are on GOAT-Bench Val Unseen and a Unitree Go2 deployment, and the strongest claim is that a smaller open model (Qwen2.5-VL-7B) with the semantic map outperforms GPT-4o baselines without it. Readers should check how much of that gain comes from the map versus the evaluation setup, and whether the 100% real-world success rate generalizes beyond the tested environments.
RoboAtlas: Contextual Active SLAM
arXiv:2606.26046v3 Announce Type: replace Abstract: We present RoboAtlas, a contextual Active SLAM framework that adaptively balances geometric exploration and semantic reasoning using a scalable 3D semantic mapping system, OpenRoboVox. RoboAtlas integrates frontier exploration, global semantic-map reasoning, and egocentric VLM-based reasoning through a contextual multi-armed bandit that transitions from exploration to semantically guided navigation as scene understanding improves. We evaluate the system in simulation and on a Unitree Go2 robot in large-scale real-world environments exceeding 1800 m2 with approx. 30k mapped semantic instances, achieving a 100% task success rate. On the GOAT-Bench "Val Unseen" benchmark, RoboAtlas achieves state-of-the-art performance with highest reported success rate (SR) of 90.6%, using GPT-4o, improving over the strongest prior baseline by 17.8 percentage points in SR. Using the much smaller Qwen2.5-VL-7B model, it still achieves 88.8% SR, outperforming all baselines using GPT-4o in SR, and revealing the importance of the information gained by our semantic mapping framework over simply replacing the underlying foundation model. The results demonstrate that grounding foundation models with large-scale 3D semantic maps enables robust and efficient contextual Active SLAM.
Abstract:We present RoboAtlas, a contextual Active SLAM framework that adaptively balances geometric exploration and semantic reasoning using a scalable 3D semantic mapping system, OpenRoboVox. RoboAtlas integrates frontier exploration, global semantic-map reasoning, and egocentric VLM-based reasoning through a contextual multi-armed bandit that transitions from exploration to semantically guided navigation as scene understanding improves. We evaluate the system in simulation and on a Unitree Go2 robot in large-scale real-world environments exceeding 1800 m2 with approx. 30k mapped semantic instances, achieving a 100% task success rate. On the GOAT-Bench "Val Unseen" benchmark, RoboAtlas achieves state-of-the-art performance with highest reported success rate (SR) of 90.6%, using GPT-4o, improving over the strongest prior baseline by 17.8 percentage points in SR. Using the much smaller Qwen2.5-VL-7B model, it still achieves 88.8% SR, outperforming all baselines using GPT-4o in SR, and revealing the importance of the information gained by our semantic mapping framework over simply replacing the underlying foundation model. The results demonstrate that grounding foundation models with large-scale 3D semantic maps enables robust and efficient contextual Active SLAM.
Alexander Schperberg and Shivam K. Panda made equal contribution
Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
arXiv:2606.26046 [cs.RO]
(or arXiv:2606.26046v3 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2606.26046 Focus to learn more arXiv-issued DOI via DataCite
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.
Text integral sub licența deschisă a publicației · 278 cuvinte
Subiecte conexe
Game Design Creativity evals
LessWrong · 12 oct. 2026
[AINews] not much happened today
Latent Space · 10 sept. 2026
A Survey on Reinforcement Learning Applications in SLAM
arXiv · 30 sept. 2026
ECHO: Embodied Camera Observations of Human Object Carrying
arXiv · 8 oct. 2026