usage.txt

Note: I only know how to do this in Linux. YMMV.

There's lot to be polished here, but this should work for now.

First connect to your Beddit - it has a RFCOMM (Bluetooth serial port)
service. Either use some GUI tool or do something like this:

------------------------------------------------------------
# hcitool scan
Scanning ...
        xx:xx:xx:xx:xx:xx       Beddit NNNN
------------------------------------------------------------

Note the Bluetooth MAC. You need it below.

------------------------------------------------------------
# /etc/init.d/bluetooth restart
------------------------------------------------------------

Now you should be able to connect to your Beddit:

------------------------------------------------------------
# rfcomm connect xx:xx:xx:xx:xx:xx
------------------------------------------------------------

If this command succeeds, you will have a /dev/rfcomm0 (or other
number) character device you can use to communicate with Beddit. Now
you can start the measurement:

------------------------------------------------------------
$ ./measure.py /dev/rfcomm0
------------------------------------------------------------

The command will produce a timestamped file named 'data.out' (it will
refuse to run if one exists). That file will show the data sent by
Beddit in hexadecimal format. Terminate the program with ctrl-c when
you wish to stop the measurement. It will produce about 100 megabytes
of data in one night.

The decoding utility will run on a binary file without timestamps.
to_binary.py converts data.out to data.bin, which only contains the
raw data your Beddit sent:

------------------------------------------------------------
$ ./to_binary.py
$ ./decode.py
[you may see some warnings about CRC errors and inconsistent packet
numbers if the connection had problems.]
------------------------------------------------------------

decode.py produces a file named data.txt which has the raw sample
data in numeric format. The first column is the "normal" channel,
while the second column is the "gained" channel. If you had any missed
packets, the data from those packets is currently simply dropped (so
the data in data.txt will skip forward without any indication that
anything is wrong).

140 samples (lines) correspond to one second of data. If you plot a
minute (8400 samples) of the data at a point when you are in deep
sleep, you should be able to see your heartbeats and your breathing
there.