Map-Assisted LOS/NLOS Identification for 5G MIMO Systems

In Integrated Sensing and Communication (ISAC) systems, the identification of Line-of-Sight (LOS) and Non-Line-of-Sight (NLOS) signal propagation states is of great significance for achieving key technologies such as high-precision positioning, beamforming, and environmental sensing. Although existing methods based on channel data have achieved high accuracy in specific scenarios, they generally suffer from insufficient generalization capabilities across different scenarios. To address this issue, this paper proposes a LOS/NLOS identification method that integrates map information with channel data. Our method can leverage map information to introduce stable geometric priors, guiding a Convolutional Neural Network (CNN) based model to perceive the physical relationship between building obstructions and LOS/NLOS states, thereby enhancing cross-scenario identification capabilities. We validated the proposed scheme in multiple 5G Urban Microcellular (UMi) vehicular communication scenarios. Experimental results demonstrate that the proposed method achieves higher identification accuracy and stronger generalization in UMi scenarios with varying building densities and distributions.