Cross-Embodiment Transfer in Vision-Based Robot Navigation: Challenges, Methods, and Open Problems

Zhang, Yiliang

Cross-Embodiment Transfer in Vision-Based Robot Navigation: Challenges, Methods, and Open Problems

Yiliang Zhang

Eastern Institute of Technology, Ningbo, Zhejiang, China

Identifier: KAGG:2607.00001

Resolver: https://academy.kaggleyes.top/id/KAGG:2607.00001

DOI: None

Abstract

The ability to deploy a trained navigation policy across robots with substantially different physical embodiments remains one of the central challenges in mobile robotics. This survey reviews configuration-driven cross-embodiment navigation, including height-conditioned perception, dimension-aware local planning, and safety assurance across heterogeneous platforms.

References

T. Zang, S. Cheng, H. Huang, S. Wang, and W. Zhang. AgniNav: Configuration-Driven Cross-Embodiment Local Planning for Robot Navigation. arXiv:2606.10903, June 2026. https://arxiv.org/abs/2606.10903

KAGG identifiers are local persistent identifiers issued by Kaggleyes Academy. They are not DOI Foundation DOI names.