Caribe Wave Sensor pusher

Retrieves data from different sensors on a Raspberry Pi, and output them to stdout for consumption for the next stage (pheromon, see this repo).

Requirements

The wiringPi library must be installed before hand and the spi_bcm2835 module must be enabled (and not blacklisted).

To compile the code, we use g++ (Raspbian 4.9.2-10) 4.9.2. A Makefile is provided.

Hardware

We use prototyped boards with MPC3002 + analog accelerometers, or SPI enabled accelerometers directly, connected to the SPI pins of a Raspberry Pi B+

Supported accelerometers

Nearly all analog accelerometers should be supported out of the box since they are connected through a MCP3002 10bit ADC.

The code has been tested and optimized for the following analog sensors :

• ADXL33 (See https://www.sparkfun.com/products/12803)
• MMA7260Q (See https://www.sparkfun.com/products/retired/308)

16bit SPI accelerometers should be supported too, if they comply to the LIS331 set of instructions (which is pretty obscure if you don't read the datasheet carefully).

The code has been tested and optimized for the following SPI sensors :

• LIS331 (See https://www.sparkfun.com/products/10345)

Operation

./main [accelerometer_type] [sampling_delay_in_ms] [trigger] [time span] [debug flag]

Arguments

All arguments are mandatory.

• Accelerometer type : Either "LIS", "MMA" or "ADXL". No default. Uppercase only.
• Sampling rate : in milliseconds, between 0 and 1000.
• Trigger: minimum trigger value in milli g.
• Trigger time span: the time span during which we continue to output data after a trigger, in millis. Must be a multiple of the sampling rate.
• Debug flag : Either 1 or 0.

Example :

# Starts the main process in debug mode, with a sampling rate of 1000 milliseconds, for a LIS331 type accelerometer
./main LIS 1000 800 1

The sampling rate is ajusted to the time taken to do the actual sampling via a call to clock(). The precision is sub-ms.

Output format

We output delta values to take into account the fact that the physical device is not a real seismograph (not bolted to the ground, always in the same direction).

We calculate an l²-norm and compare it against the squared trigger value in g² to decide if we output the values or not (so as not to clutter the output with irrelevant data).

Licence

MIT.