Adaptive Neural-Fuzzy Systems for Adaptive Signal Coverage Power Estimation in Cellular Broadband Networks

Abstract

 Accurate estimation of signal coverage Power is paramount for efficient network planning, optimization, and resource management in cellular broadband networks. Traditional empirical and deterministic propagation models often lack the adaptability to dynamic wireless environments, resulting in suboptimal coverage and capacity. This paper proposes a novel Neural-Fuzzy (NF) method for adaptive signal coverage power estimation, combining the powerful learning capabilities of Artificial Neural Networks (ANNs) with the robust reasoning under uncertainty provided by Fuzzy Logic Systems (FLS). The proposed hybrid method is termed the Adaptive Neuro-Fuzzy Inference Systems (ANFIS) and it leverages ANNs to learn complex, non-linear relationships between environmental parameters and signal power attenuation, while the FLS refines these estimations by incorporating real-time feedback and heuristic rules to account for dynamic, unmodeled factors such as atmospheric conditions. The methodology details the architecture, training, and adaptive mechanisms of the NF system. Performance metrics, attained results, and comparative analysis against conventional models demonstrate the superior accuracy, adaptability, and robustness of the proposed ANFIS approach, making it an effective tool for next-generation cellular network management.