Finally the HR sensors from PulseSensor (which now graduated from KickStarter and has now its own website) arrived: the package is really nice, and contains both the ear-clip and the velcro to attach in inside a glove or any other band.
It also comes with software to test it’s working and track the heart beat with a Processing sketch.
We already assembled everything on the small protoboard, created the finger-caps to hold the sensors, and added the SD to log all the values tracked.
At this point we are now encountering a few problems:
- since the HR sensor depends on the light reflected by blood capillaries, it doesn’t work very well with people who don’t have lots of them (the ones that usually get cold fingers). Also it depends a lot of the pressure put on the fingertip: not enough and the sensor will move and give unstable reading, too much and the fingertip will be “compressed”, thus impeding free flow of blood, and giving wrong readings
- the GSR is as well not very accurate, we cannot get it to really change the readings by inducing change in the emotional state (pinching, laughing, asking personal and lie-prone questions): probably it’s because we are testing in a very dry and cold room, so sweating is less likely to happen
- finally, the two sensors don’t work very well together: individually they work, but when together it seems like the readings become too much correlated not to be caused by some electrical interference.
For the moment we think GSR is a bit too much unreliable, so we are probably just going to use the HR beat to test our hypotesis.
We also recorded a short video of us testing the heart beat detector:
In the meantime, together with Geoffroy at the Académie, we are refining the core ideas behind the project, and better understanding how we could propose the installation to the public.