Radio Environment Map Construction by Residual Kriging Based on Bayesian Hierarchical Model.

This article focuses on the development and validation of the BHM-RK algorithm, a novel method for constructing radio environment maps (REMs) in multi-transmitter scenarios. The algorithm integrates spatial clustering for transmitter partitioning, Bayesian hierarchical model (BHM) posterior inference using Gibbs sampling for pathloss estimation, and residual Kriging interpolation for shadow fading estimation. Experimental results from both synthetic and real-world data demonstrate that BHM-RK outperforms existing methods—including BHM without shadow fading, ordinary Kriging, generalized regression neural networks, natural neighbor interpolation, and inverse distance weighting—in terms of accuracy measured by root mean squared error (RMSE) and mean absolute error (MAE). This research contributes to enhanced precision in REM construction, which is critical for spectrum sensing, control, and sharing in applications such as unmanned systems, the Internet of Things, and 5G communications.