A comparative analysis of alpha-beta-gamma and close-in path loss models based on measured data for 5G mobile networks

Abstract

Mobile coverage is crucial for the fifth-generation (5G) network since it affects the network's accessibility and dependability in various locations. With a wider coverage, more people and devices will access the 5G network, allowing them to experience the benefits and capabilities of this important technology. However, the development and construction of the networks need a lot of time and effort to maximize network coverage and service delivery while utilizing the fewest possible infrastructure components. Path loss models are commonly employed to predict network coverage. Therefore, it is expedient to adopt a path loss model that is suitable for the specified geographical area. In this paper, we investigated and analysed two empirical models: the Close-in free space reference distance (CI) model and the alpha-beta-gamma (ABG) model at the low- and mid-spectrum bands. Secondly, the path loss parameters were derived to characterize the two empirical models based on data from the measurement campaign. Finally, Root Mean Square Error (RMSE) was employed for the comparison of the models. It was found that at frequencies 0.9 and 1.8 GHz, the CI model's least RMSE values were 3.6 and 4.1 at location 1 and 4.7 and 4.2 dB at location 2, respectively, outperforming the ABG. However, at 3.4 GHz, the ABG with 3.25 and 1.75 dB outperformed the CI at both locations. For future work, the integration of machine learning techniques to dynamically predict and adapt path loss models in real time based on continuous data collection will be appropriate.