An Ubiquitous 2.6 GHz Radio Propagation Model for Wireless Networks Using Self-Supervised Learning From Satellite Images

Abstract

The performance of any Mobile Wireless Network (MWN) is dependent on the appropriate level of radio coverage, with Path Loss (PL) models being a valuable resource for its evaluation. Recently, advancements in Machine Learning (ML) and Deep Neural Networks (DNNs) have been applied to radio propagation to produce new data-driven PL models. Notoriously, these advancements have also allowed the inclusion of non-classical inputs, such as satellite images. However, data-driven PL models are often developed under the assumption that training and test data distributions are similar, which is a weak assumption in real-world scenarios. Thus, generalization (i.e., the model's ability to perform on different data distributions) is a crucial aspect of data-driven PL models in the context of Mobile Network Operators (MNOs). This paper proposes a new data-driven PL model, the Ubiquitous Satellite Aided Radio Propagation (USARP) model, developed to enhance the geographical generalization capabilities of empirical PL models, by using satellite images. The USARP model considers self-supervised learning to extract general data representations of the radio environment from satellite images, improving the PL prediction Root Mean Square Error (RMSE) of the 3 rd Generation Partnership Project (3GPP) PL model in the order of 9 dB, and for a data distribution distinct from the training data. Moreover, it was demonstrated the potential of the USARP model in terms of geographical and radio environment generalization. Although the generalization capabilities of ML regression algorithms are limited, the chosen USARP architecture and the use of regularization techniques had a positive impact on its geographical generalization performance.

On the Use of Spatial Graphs for Performance Degradation Root-Cause Analysis Toward Self-Healing Mobile Networks

“…Secondly, regarding the SGCN architecture, it is envisaged to test a single graph implementation to higherorder analysis in the node convolutions. One possibility to explore involves using satellite imagery for terrain features, similar to the approach in [11], where the graphs can be embedded within image pixels. This approach facilitates the simulation of intricate scenarios, with individual pixels representing elements such as BSs, users, sensors, or other pertinent network variables.…”

“…The recent success of AI, namely Machine Learning (ML) and DL, has fostered an increasing adoption of learning-based solutions to plan, manage, and optimize wireless networks using path-loss predictions, such as in [11][12][13]. AI can contribute to solving cornerstone problems in self-healing operations, such as performance management, predictive fault detection, automatic root-cause analysis and self-healing regenerative mechanisms.…”

“…• elevation is the total antenna installation height, including the terrain elevation plus the height of the installation infrastructure; • azimuth is the angle between the axis of the BS's antenna and the geographical north. The initial two features capture the Euclidean distance between the nodes, a critical variable in any wireless network modelling problem because of its direct influence on the Received Signal Strength Indicator (RSSI), as shown in [11]. The third component allows us to calculate the difference in antenna height between the source and target BSs by combining the antenna installation height with the terrain elevation.…”