Abstract:Accurate path loss prediction is crucial for wireless network planning and optimization in suburban environments with complex terrain variation and diverse land cover. This paper proposes a model assisted hybrid path loss prediction method that introduces an environment adaptive compensation on top of the classic close-in free-space reference distance (CI) path loss model. By jointly predicting the path loss exponent and a compensation term, the proposed approach dynamically adjusts the empirical trend. To improve the effectiveness of environmental representation, three environmental image organization schemes are constructed and evaluated. Experiments on measurement data collected in Pingtan Island show that the proposed method outperforms the CI model and a conventional model assisted baseline, achieving a test root mean square error of 4.04 dB.
From: Chenlong Wang [view email]
[v1]
Tue, 10 Mar 2026 15:37:20 UTC (12,750 KB)
[v2]
Mon, 7 Sep 2026 02:38:57 UTC (15,291 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。