AvocadoToast pulls accelerometer data off the LIS3DH MEMS module and creates a web interface for the data. It is capable of acting as a sleep monitor, and is trivially extensible to add alternate functionality or sensing ability.
The client/ directory contains the microPython
firmware that executes on the board itself,
which read from the sensor and publish events
on the EEERover MQTT server.
The web/ directory contains the backend
and frontend of our marketing website
and API. The service can be initialized
by executing the included start_server.sh
script. An MQTT listener daemon is
created, and a Flask application is started
up. The Flask application serves our marketing
website (TCP port 5000, on the root directory)
and a number of readings the sensor is capable
of taking (exported on /sensor/<cmd>, where
<cmd> is x, y, z, etc. This allows
server extensions to be implemented in Python
and custom clients to be developed with an
easy-to-use API. In fact, our index page
uses the API itself to access the JSON
database containing the raw sensor data.
A trivial extension has already been implemented
for an appliance (in this case a washing machine)
as a proof of concept.
The data/ directory contains data of a user's
sleep patterns collected using the device.
The plot_data python files were used to graph
the sensor readings for testing purposes.
Plotting the data off-board allows for understanding
of what operations should be performed on-board
to get an appropriate metric for sleep activity