A Novel Propagation Pathloss Model Calibration Tool

Abstract

This paper develops and comprehensively evaluates a very simple and remarkably efficient method, here referred to as the Quasi-Moment-Method (QMM), as a tool for the calibration of basic (classical) pathloss models, in various radiowave propagation scenarios. After a succinct description of the characterizing features of the method, the paper presents computational results involving comparisons with published data concerning Minimum Mean Square Error (MMSE) optimization, Adaptive Neuro-Fuzzy Inference System (ANFIS) pathloss modeling, and (for an indoor case), a dual-slope reference model. The results reveal that the QMM remarkably outperforms the MMSE-optimized and modified single-slope, close-in reference models, when evaluated in terms of the statistical measures of Root Mean Square Error (RMSE), Mean Prediction Error (MPE), and Standard Deviation Error (SDE). Computational results due to the QMM also compare favorably (and better, with some performance metrics) with published corresponding results available from the literature, in which heuristic (ANFIS and Artificial Neural Network (ANN)) as well as geospatial (Kriging) models were utilized. A number of inherent properties of the real, symmetric ‘model calibration matrices’ associated with QMM process are identified in the paper, as offering interesting possibilities for further investigations involving eigenvalues and corresponding eigenvectors.