Map as a Prompt: Learning Multi-Modal Spatial-Signal Foundation Models for Cross-scenario Wireless Localization
Artificial IntelligenceMachine Learning
Summary
The authors address the challenge of accurately locating wireless signals, which is important for technologies like self-driving cars and smart factories. They introduce SigMap, a new model that uses two main ideas: a way to hide parts of the signal based on its natural patterns to better learn from it, and a method to include 3D map information to help the model work well in different places. Their experiments show that SigMap performs better than previous methods, even when tested in totally new environments without extra training. This suggests their approach is both accurate and adaptable for wireless localization tasks.
Authors
Yong Chu, Xun Zhou, Zenglin Xu, Hui Wang, Yue Yu
Abstract
Accurate and robust wireless localization is a critical enabler for emerging 5G/6G applications, including autonomous driving, extended reality, and smart manufacturing. Despite its importance, achieving precise localization across diverse environments remains challenging due to the complex nature of wireless signals and their sensitivity to environmental changes. Existing data-driven approaches often suffer from limited generalization capability, requiring extensive labeled data and struggling to adapt to new scenarios. To address these limitations, we propose SigMap, a multimodal foundation model that introduces two key innovations: (1) A cycle-adaptive masking strategy that dynamically adjusts masking patterns based on channel periodicity characteristics to learn robust wireless representations; (2) A novel "map-as-prompt" framework that integrates 3D geographic information through lightweight soft prompts for effective cross-scenario adaptation. Extensive experiments demonstrate that our model achieves state-of-the-art performance across multiple localization tasks while exhibiting strong zero-shot generalization in unseen environments, significantly outperforming both supervised and self-supervised baselines by considerable margins.