arxivcs.RO2026-07-20
VLN-AVP: Zero-Shot Vision-Language Navigation with Hybrid Long-Short-Term Memory for Autonomous Valet Parking
Yijian Li, Xiangru Mu, Changze Li, Hantian Shi, Jiyuan Cai, Jia Cai, et al.
Existing methods in Autonomous Valet Parking (AVP) typically rely on pre-built maps, which severely restricts their scalability to unseen environments and open-vocabulary targets. Inspired by the application of Vision-Language Models (VLMs) in Vision-Language Navigation (VLN) tas…