We envisioned SensorX to be a standalone system that is capable of achieving its intended purpose without the need for regular human intervention.
SensorX is a system which incorporates sensors, a processor and an alarm. The built-in accelerometer will be used to detect seismic activity which accompanies a volcano eruption. These sensors will relay information to the Raspberry Pi, which processes the signal. We will use machine learning to train an algorithm to detect if the signal represents an imminent eruption. It will also be connected to a ball bearing hung on a metal string with a metal cup below, which is intended to detect tilt. Tilt detection will be handled by an Arduino.
If the algorithm detects vibrations similar to the desired levels, it will send a positive signal to the Arduino. The Arduino then effectively checks if the degree of tilt is consistent with that of the ground in an imminent eruption. If so, the ball bearing will touch the metal cup and close the circuit. An LED will light up and the bluetooth module attached to the Arduino will send a signal to nearby smartphones.
Final Design
With the need for simplification of the design in order to increase reliability, reduce size and reduce power consumption, it was decided that the Team consolidated all mechanisms within the Raspberry Pi with SenseHAT module. As such, the tilt sensing mechanism is now replaced by the detection of fluctuant behavior observed when SensorX experiences tilt.
The machine learning algorithm has also been simplified by reducing the number of features processed and optimising the types of features processed, allowing for higher accuracy in determining if the vibrations detected is the result of an imminent volcanic eruption.
To learn more about the capabilities of the devices used, visit Technical Specifications.