A Calibrated Permittivity Modeling Approach for Cross-Area Path Loss Prediction

• Published in: IEEE wireless communications letters 2023-08, Vol.12 (8), p.1299-1303
• Main Authors: Jiang, Yuanhao, Zhou, Shidong, Zhong, Xiaofeng
• Format: Article
• Language: English
• Subjects: Algorithms; Buildings; calibrated permittivity modeling; Calibration; Data models; Loss measurement; Message passing; Modelling; Optimization; Path loss prediction; Permittivity; Permittivity measurement; Predictive models; radio propagation; Ray tracing; Reference signals; Wavelength measurement; Wireless communications
• ISSN: 2162-2337; 2162-2345
• DOI: 10.1109/LWC.2022.3232380

Accurate path loss prediction (PLP) in target areas (TAs) is required in many wireless communication applications. However, in real-world deployment, there only may be available path loss data outside TAs. In this letter, a framework based on calibrating the modeling of buildings' permittivity for cross-area PLP is proposed, with the advantage of not requiring data from TAs. The proposed framework requires only the Reference Signal Received Power (RSRP) data from other measured areas (MAs, no intersection with TAs), to learn more precise permittivity modeling and indirectly enhance the prediction in TAs. Furthermore, we exploit a message-passing based algorithm and its low-complexity implementation in the optimization of calibrating permittivity modeling, to improve the prediction accuracy and reduce the iterations required. Real-world measurement is performed and the results have revealed a better performance improvement of the proposed method, compared with Random Forest (RF) and ray-tracing (RT).