Our approach to build a Seismometer on a smartphone presents different challenges
in using the capabilities of accelerometer sensors in smartphones. The primary
challenge is to deal with the volume of the raw data that is sensed by the accelerometer
over a continuous period, a smart phone accelerometer sensor reading the acceleration
values at the fastest sampling rate which is 200 samples per second could maintain a
history of its observations by logging them which produces 25 Gigabytes of raw data
per month. This is a large volume of data to store and analyze, so with much less
storage the detection of events is made possible by implementing triggering algorithms
like the seismometers do [9]. However, the challenge is the computation of algorithm
in real-time while the smartphone is continuously reading the accelerometer values
at a high sampling rate (mentioned earlier) for fine grain detection of weak motion
events whose probability of detection is very low by the sensing devices. Most of the
Seismometers use simple average or threshold based triggering approach for detection,
some of the other approaches use adaptive technique, neural network methods
or sophisticated ones based on pattern recognition