Earthquake Detection Using Accelerometer Moving Average Filtering

The patent applies preliminary action by performing moving average filtering on acceleration data before earthquake detection. The control module filters raw acceleration data from the three-axis accelerometer to generate filtered acceleration data, which removes environmental noise and improves reliability. This preprocessing step is executed continuously in the background, so when an earthquake occurs, the filtered data is already prepared and ready for immediate comparison and detection, thus maintaining rapid response while improving reliability.

The patent applies dynamics by using moving average filtering instead of static average filtering. The control module continuously updates the filtered acceleration data by calculating the moving average of recent acceleration values, allowing the filter to adapt to changing vibration patterns in real-time. This dynamic filtering approach maintains reliability by smoothing noise while preserving the ability to detect continuous ground vibrations, as the filter responds dynamically to new acceleration inputs rather than using a fixed averaged value.

The patent applies feedback by implementing a control module that continuously monitors filtered acceleration data and compares it with predetermined thresholds to determine earthquake occurrence. The system provides feedback by continuously updating the filtered acceleration data based on new acceleration inputs and adjusting earthquake detection decisions based on the filtered data trends. This feedback mechanism ensures that the apparatus operates reliably by making detection decisions based on processed, reliable data rather than noisy raw data, reducing false failures while maintaining operational stability.