Conventionally statistical path loss models are high-dimensional data-based without utilizing specific environment features. In this letter, a novel environment features-based model (EFBM) for path loss prediction is presented. We connect the propagation environment and channel by representing the environment with low-dimensional features: distance, deviation, volume, and blockage. The features are propagation-related, which can predict path loss directly by utilizing the Random Forest (RF) method. Compared with the data-based method, the proposed method can reduce the Root Mean Squared Error (RMSE) by 0.33 and 0.89 dB at 6 and 28 GHz and provide closer results to the Ray-Tracing (RT)-based ground-truth values.

中文翻译：

---

基于环境特征的路径损耗预测模型


传统的统计路径损耗模型是基于高维数据的，没有利用特定的环境特征。在这封信中，提出了一种用于路径损耗预测的新颖的基于环境特征的模型（EFBM）。我们通过用低维特征表示环境来连接传播环境和通道：距离、偏差、体积和阻塞。这些特征与传播相关，可以利用随机森林（RF）方法直接预测路径损耗。与基于数据的方法相比，所提出的方法可以在6和28 GHz处将均方根误差（RMSE）降低0.33和0.89 dB，并提供更接近基于光线追踪（RT）的地面实值的结果。