Abstract
An earthquake early warning system has been developed to obtain real-time profile of impending ground motion in order to minimize earthquake hazard The system is designed to analyze the P-wave portion of earthquake signal and instantaneously evaluate earthquake parameters such as magnitude, epicentral distance and focal depth. Neural networks are employed to add self-learning functions to the existing system and to reduce the error of estimation for site-specific applications. The neural network has been successfully trained with recorded data to produce accurate and informative earthquake parameters.
