Title:Multi-Sources Information Fusion Learning for Multi-Points NLOS Localization

Abstract:Accurate localization of mobile terminals is crucial for integrated sensing and communication systems. Existing fingerprint localization methods, which deduce coordinates from channel information in pre-defined rectangular areas, struggle with the heterogeneous fingerprint distribution inherent in non-line-of-sight (NLOS) scenarios. To address the problem, we introduce a novel multi-source information fusion learning framework referred to as the Autosync Multi-Domain NLOS Localization (AMDNLoc). Specifically, AMDNLoc employs a two-stage matched filter fused with a target tracking algorithm and iterative centroid-based clustering to automatically and irregularly segment NLOS regions, ensuring uniform fingerprint distribution within channel state information across frequency, power, and time-delay domains. Additionally, the framework utilizes a segment-specific linear classifier array, coupled with deep residual network-based feature extraction and fusion, to establish the correlation function between fingerprint features and coordinates within these regions. Simulation results demonstrate that AMDNLoc significantly enhances localization accuracy by over 40\% compared with traditional convolutional neural networks on the wireless artificial intelligence research dataset.