Improving a user experience is an important part of many operations, especially when it’s students trying to use a lab for their university work. Instead of giving students endless surveys, English phonetics lecturer Emmanuel Ferragne decided to use the Raspberry Pi to track how the labs were being used, in an effort to make them better.
What are you using the Raspberry Pi for?
I call it the Lab Monitor for want of a better label. Basically I had all those expensive, state-of-the-art instruments from a number of research projects, but they were stored in a cupboard because no room was available. Things have changed very recently and I now have a small lab where my students can use the hardware for their own research. Rather than imposing strict rules, I thought I would collect data to understand how students actually use the lab. So the Lab Monitor is really there to learn from the students’ habits and ensure that the lab meets their needs, and not the other way around!
I must insist that this is not about surveillance at all. I have no means of knowing who exactly is using the lab at any given time – and I don’t want to know! If the Lab Monitor tells me there’s activity going on in the lab, then I’m more than happy.
What kind of research are you doing?
I’m involved in a number of projects related to the acquisition of English by French learners. So I use techniques like acoustic speech processing, ultrasound tongue imaging, electroencephalography (EEG), etc., to better understand how learners produce and perceive the sounds of a foreign language. More generally, I’m interested in all aspects of the human voice: what it says about you, how it changes depending on who you are talking to, to what extent somebody’s voice can be said to be unique, etc.
Why the Raspberry Pi?
I’m always on the lookout for new hardware, and over the years I’ve developed a keen interest in tweaking the instruments I use for my research. One problem with the professional-grade instruments I work with is that some of them come with limited hackability, which I find ever so frustrating! But I knew the Raspberry Pi would let me be creative.
How does the system work?
In the current version of the Lab Monitor, on the hardware side, there’s a Raspberry Pi 3, a Sense HAT, an infrared motion detector, and a Mini Black HAT Hack3r. The Sense HAT gathers temperature and humidity data, and students can move the joystick if they don’t want to be disturbed when they run experiments. The Sense HAT is connected to the Pi via a Black HAT Hack3r so as to obtain more accurate temperature readings.
On the software side, the Nmap program scans the IP addresses of the computers in the lab at regular intervals to check their status (on/off); this is achieved with a shell script. All remaining operations are managed by a Simulink (a visual programming language by MathWorks) model that I deployed to the Pi. It collects the output from the IP scan, the temperature, humidity, joystick, and motion sensors, and sends the data to ThingSpeak, which is the MathWorks IoT platform. The IoT channel is public, so all users can check if, for example, the computer that is connected to the special instrument they are planning to use is available.
I have developed a Matlab GUI that lets me import the data feed and analyse it offline. For instance, I can check if some computers are over-/under-used and modify the lab setup accordingly. Or I can identify empty time slots to plan new lab meetings. Temperatures in the building can go pretty high in the summertime so I’ll keep an eye on them too.
Do you have any future plans for the research setup?
Well, strictly speaking, this project is more about workplace ergonomics than research. But if it makes life easier for my students, then it will have a positive impact on their research. And yes, I have plans to extend the system to other rooms here at the university, provided I get money to buy the hardware. But before that, two aspects should be improved. First, some downsizing is in order: for example, the Sense HAT here is overkill, and also, I might consider using a Pi Zero instead. And second, I have yet to find or build a suitable enclosure.
Do you have any future plans for using the Raspberry Pi in other research?
Yes, among other things, I sometimes work with a wonderful, tiny piece of hardware for biosignal acquisition called BITalino. In our current setup, the BITalino gathers data and sends it to a PC over Bluetooth. If we replace the PC with a Pi, the system becomes totally wearable and we can run our experiments anywhere. And the Pi could process the data and trigger actions based on certain features of the electrophysiological signal. In an experiment we are carrying out right now, we’re recording people’s heart rate and electrodermal activity in response to emotionally loaded words. It is easy to imagine how we could program the Pi to wait for these two signals to return to some ‘resting state’ before playing back the next audio stimulus. The Pi could also sound a buzzer if a participant’s heart rate goes above or below certain thresholds.