Extended Weighted ABG: A Robust Non-Linear ABG-Based Approach for Optimal Combination of ABG Path-Loss Propagation Models

This paper proposes a robust non-linear generalized path-loss propagation model, the Extended Weighted ABG (EWABG), which efficiently allows generating model by combining several available datasets (from measurements campaigns) and other previously proposed state-of-the-art 5G models. The EWABG works integrating individual models into one single in least-squares sense, allowing to extend knowledge from frequencies distances covered or is first extension of common ABG-based approach, surpasses non-uniformity problem between low high (as most campaigns have taken place frequencies). also addresses removing outlier measurements, step not included previous In this case, we compared recent techniques for avoiding outliers, adopted Theil-Sen method, due its strong robustness demonstrated experiments carried out. addition, specifically considers attenuation atmospheric gases, order improve estimations. good performance has been tested against including WABG exhaustive experimentation out includes non-line-of-sight environment different scenarios, UMiSC, UMiOS UMa. obtains best accuracy, specially noisy environments with reporting negligible increment error rates (with respect non-outliers situation), lower than 1%, WABG.