Passive Fingerprint Positioning Based on Multi-Dimensional Signal Feature Fusion

To address the limitations of active positioning methods requiring active terminal participation, this study proposes a passive positioning scheme based on base station Measurement Report (MR). By exploiting multi-dimensional features in terminal MR data, including reference signal received power (RSRP) and timing advance (TA), we construct a lightweight fingerprint database using sparse GPS sampling points from historical MR data, and design a weighted K-nearest neighbor (WKNN) positioning model based on signal spatial similarity. Experimental results demonstrate that the proposed method achieves favorable positioning accuracy and stability in typical urban scenarios without requiring real-time positioning assistance from terminals.