Recognition of Electric Earthquake Precursors using Neuro-Fuzzy Methods: Methodology and Simulation Results

A. Konstantaras and M.R. Varley (UK), F. Vallianatos (Greece), and G. Collins and P. Holifield (UK)

Keywords

Non-Linear DSP, Neuro-Fuzzy Systems, Seismic Signal Processing, Adaptive Filtering

Abstract

Electric Earthquake Precursors (EEPs) are transient electric potential anomalies appearing on the Earth’s electric field prior to an earthquake. EEP recognition is a key problem for the seismological community as EEP signals are usually very weak and are often unobservable within the electric background. Furthermore, noise of magnetotelluric origin, rain and anthropogenic noise, are only some of the types of disturbances that make reliable recognition of EEPs even more difficult. This paper describes a hybrid neuro-fuzzy system for recognition of EEPs within noisy electric field signals. The model effectively acts as a non-linear inverse adaptive filter, and has been trained to recognise EEPs from the electric background, using simulated signals in order to be able to confidently assess system performance. Preliminary results are presented, demonstrating the effective performance of the system.

Important Links:



Go Back