Seismo — Geophone system

Project technical lead, building and managing a team for the development of a geophone system, which, after research, was divided into 2 independent projects:

PoE array of geophones. The array of sensors analyzed the “sound” of the earth, sensors consisted of Rpi4, 24bit ADC, PoE, a geophone with a differential power amplifier. The sensors had a synchronization system and could determine in the monotonous noise of the earth, which is the same in all sensors, carry out signal correlation, and detect under the noise signal, thanks to correlation, we could detect the phase of the signal offset and determine the direction and type of noise source. To accurately determine the position of the source, it was necessary to know the exact position of the sensors. Since the speed of sound is uneven, and can change from the weather, or from the removal of the target, an algorithm was developed that built the geometry of the array itself. The system had a server where the source of noise can be seen on the map.

Since the system must have a high speed of operation, large amounts of data, it could not be autonomous, a line of battery-powered sensors was made.

The system was built on the basis of nrf52. They have excellent consumption characteristics. The sensor collected data from a precision 16-bit ADC, which has excellent dynamic range. The fast Fourier transform was obtained by the hardware module of the microcontroller, autocorrelation was performed, and the type of interference was determined.

The device was configured via BLE, from the phone. The sensor had a GPS receiver and a LORA module. The devices were networked, and at the end there was a server with a LoRa receiver and a telegram bot, where you could subscribe to certain devices, configure notifications, etc. An offline modem was made that could receive all the data, and BLE notifications were received by an employee who could be outside the cellular connection.

Synchronized network from geophones. Rpi4, STM32, nRF52, LoRa, LowPower, FFT Python.