Mid Sweden University

Earthquakes are natural phenomena whose consequences can be catastrophic. The introduction of new technologies such as intelligent transducers, which, being equipped with a communication interface, allow the exchange of information with other devices connected to the network, prompted researchers to study systems able to exploit data related to the first moments of the seismic event in an attempt to predict its characteristics. This idea is the base of Earthquake Early Warning (EEW) systems that aim to estimate the upcoming event's fundamental parameters to issue an alarm signal in case of danger, allowing the necessary precautions to be taken. There are many problems concerning parameters' estimation representing an indication of the phenomenon's danger in this context. In particular, the magnitude estimation and the compromise between the processing of reliable results and the time window's brevity in which the system algorithm must acquire and process the data to produce such results represent challenging aspects. This paper present new results of an approach, proposed by the authors that use a quotient of iterative filters. By choosing the right filter parameters, this quotient correlates with the earthquake magnitude by processing the incoming wave's first three seconds. © 2021 IEEE.