Abstract
In modern wireless communication systems, radio propagation modeling to estimate pathloss has always been a fundamental task in system design and optimization. The state-of-the-art empirical propagation models are based on measurements in specific environments and limited in their ability to capture idiosyncrasies of various propagation environments. To cope with this problem, ray-tracing based solutions are used in commercial planning tools, but they tend to be extremely time-consuming and expensive. We propose a Machine Learning (ML)-based model that leverages novel key predictors for estimating pathloss. By quantitatively evaluating the ability of various ML algorithms in terms of predictive, generalization and computational performance, our results show that Light Gradient Boosting Machine (LightGBM) algorithm overall outperforms others, even with sparse training data, by providing a 65% increase in prediction accuracy as compared to empirical models and 13x decrease in prediction time as compared to ray-tracing. To address the interpretability challenge that thwarts the adoption of most ML-based models, we perform extensive secondary analysis using SHapley Additive exPlanations (SHAP) method, yielding many practically useful insights that can be leveraged for intelligently tuning the network configuration, selective enrichment of training data in real networks and for building lighter ML-based propagation model to enable low-latency use-cases.
Highlights
E MERGING cellular networks are anticipated to witness a dramatic growth in connected devices and exciting new vertical services
While the proposed model yields better accuracy than empirical models, our analysis shows that its computational complexity and implementation cost is much lower than the highly sophisticated ray-tracing based tools that are being widely used in commercial cell planning tools, because it only uses the key features as input to the trained Machine Learning (ML)-based model to predict the Received Signal Strength (RSS), as compared to ray tracing, which approximates the interactions of all rays with the neighboring environment to estimate the pathloss, computationally inefficient
The results show that LightGBM outperforms other ML tools, including Deep Neural network (DNN), in terms of computational complexity and robustness to extremely sparse training data, as often is the case in real networks
Summary
E MERGING cellular networks are anticipated to witness a dramatic growth in connected devices and exciting new vertical services. In practice approximate methods such as ray tracing are used to model signal propagation, by taking into account the interactions of rays with the environment and using the dominant ray path to calculate the pathloss. These models can be very accurate depending on the resolution of available topographical database, but are computationally inefficient. Empirical and semi-empirical models such as COST-Hata [5], Stanford University Interim (SUI) [6], Standard Propagation Model (SPM) [7] and ITU-R P.452-15 [8] can be efficiently computed. AI can replace classical mathematical models with a robust data-driven pathloss prediction model, that is more accurate than empirical propagation models and more computationally efficient than deterministic models, for system level intelligent network planning and post-deployment optimization and automation in cellular networks
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
