Uncovering Age Progression in Wireless Signal Propagation Modeling Using Decisions of Machine Learning Classifiers : (Poster)

While the path-loss models of propagating wireless signals may provide different results if they are employed in an environment that differs from the original settings that they were structured for, their formulation may also become irrelevant as a result of topographic and rapid civil growth of urbanization. In this paper, we analyze the aging effect in path-loss models of wireless signal propagation in a deployed 3G network. The study utilizes measurement data form a first-phase roll-out in 2010 in Amman, Jordan to conduct comparative analysis in conjunction with the data of a similar measurement campaign in 2018 in the same area of interest. Machine Learning (ML) techniques, including Artificial Neural Networks and Decision Trees classifiers, were utilized to uncover and verify the age progression of the environment on path-loss modeling. The collected measurements are then compared to the Log-distance propagation model. ML classification algorithms proved to be a powerful tool for analysis through numerical results in verifying these progression trends.