Few things are worse than going to exercise, coming back home, and then realizing that you have been nose blind the entire time to your own odor. In order to detect the potential stench before anyone else does, Luke Berndt and his daughter, Elena, teamed up to create the Smelling Fresh, Feeling Fresh! project.
Their idea was to take a Nicla Sense ME board along with one of K-Way’s jackets as part of our recent collaboration and use it to recognize when the outerwear developed a foul smell. Data was gathered using already stinky clothes from dirty laundry bins and trash, with the BME688 four-in-one gas sensor picking up the slight differences in CO2, humidity, and volatile organic compounds (VOCs) between clean and smelly samples. All of the data was then uploaded to the Edge Impulse Studio and used to train a model, and after a few more rounds of gathering more data, it was finally accurate enough to deploy.
The original plan involved sending an alert over Bluetooth® Low Energy to an accompanying phone app and displaying the message to the user, but this proved too difficult because of low-memory issues. So instead, the duo simply made the code illuminate the RGB either red, yellow, or green to indicate the current air cleanliness.
For more details, you can check out their proof of concept on the Arduino Project Hub.
Read more about this on: Arduino Blog