Impact of Least Absolute Deviation-Based Calibration on Cell Radius and Coverage Estimation in Cellular Network Planning

Abstract

The expansion of 5G and future 6G networks necessitates dense deployments where coverage estimation is critical to minimize inter-cell interference and energy consumption. Path loss modeling provides the mathematical foundation for these estimations. While models like Hata or COST-231 offer a baseline, they require site-specific calibration to maintain accuracy in nonhomogeneous urban or suburban topographies. Measurement-based signal strength calibration is particullarlly very important for adjusting empirical path loss models to suit a particular radio environment as well as providing solutions for accurate estimation of cell coverage area and radius as well. In this work, the statistical robustness of the path loss tuning methods, namely the Absolute least deviation (LAD) and Least square regression (LS) have been employed focusing on their capability for accurate cell radius and coverage area determination with efficient geometric haxogonal celular grid layouts. It has been shown through compuational MATLAB software design and interface implementation that LAD offers much more robust estimation of cell radius planning compared to LS. This research shows that, although LS is effective for propagation modeling in perfect conditions, LAD is a far better tool for estimating values in real cellular deployments. Using LAD, engineers can get a more accurate value for the effective cell radius, thereby avoiding problems like "coverage holes" or over-deployment of base stations.