Hydrogeological Risk Management in Smart Cities: A New Approach to Rainfall Classification Based on LTE Cell Selection Parameters

The sudden climate change, that has taken place in recent years, has generated calamitous phenomena linked to hydrogeological instability in many areas of the world. An accurate estimate of rainfall levels is fundamental in smart city application scenarios: it becomes essential to be able to warn of the imminent occurrence of a calamitous event and reduce the risk to human beings. Unfortunately, to date, traditional techniques for rainfall level estimation present numerous critical issues. This paper proposes a new approach to rainfall classification based on the LTE radio channel parameters adopted for the cell selection mechanism. In particular, this study highlights the correlation between the set of radio channel quality monitoring parameters and the relative rainfall intensity levels. Through a pattern recognition approach based on neural networks with Multi-Layer Perceptron (MLP), the proposed algorithm identifies five classes of rainfall levels with an average accuracy of 96 % and a F1 score of 93.6 %.