Using machine learning and old smartphones to provide intelligent smart home services without the requirement of installing a lot of sensors.
What it does
Users can find misplaced items by speaking to a smart assistant
How we built it
- As a client we use the Microsoft Arduino board as a smart speech assistant.
- Android application to send images and other sensor values of the smartphone to the cloud
- For classification we trained machine learning models on Azure for image classification to locate items
- For building dynamic data pipelines we use the open source tool StreamPipes.
Challenges we ran into
Working with lots of different technologies and bringing them together
Accomplishments that we are proud of
Bringing it all together to a running prototype.
What I learned
How cognitive APIs work and creating a smart assistent
What's next for Fog-Yeah
Sleep, then conquer the world