Dormmate

Our door unlocking mechanism attached to the door.
Our app used to approve or reject requests to enter the dorm room.
Screenshot of real-time facial recognition system using Viam

Inspiration

It can be frustrating when you're locked out of your dorm room but your roommate is in class, or when you've left something in a close friend's room but can't get it back until they're home. It can also be useful to know when a friend's stopped by so you can catch up with them later when you have time. We were therefore inspired to build a new rendition of the traditional doorbell, adapted for dorm rooms.

What it does

Traditional doorbells tell you someone is at the door, but you can't see who it is or let them in remotely. If you're not home, your guest has to wait outside. That’s why we built Dormmate: a doorbell with facial recognition that sends push notifications to your phone with the visitor's name, and lets you unlock the door remotely, even when you're not home.

How we built it

Inspired by the Ring doorbell, Dormmate has a camera that detects the person at your door. Once the visitor presses the doorbell, which is run by a Raspberry Pi, the host will receive a notification on their phone. The Viam algorithm will identify who is there and prompt you on the Dormmate app with the choice to approve or deny the visitor’s entry.

Inside of the door frame, there is a tray preventing the door from completely locking, which makes the lock much more accessible for a rotating arm attached to the motor. Once the app sends a signal to the motor indicating that the door should open, the rotating arm will “scoop” the door lock and push it out of the hole in the door frame. Thus, the door opens without the need of the owner’s presence or illegal keycard copy.

Challenges we ran into

The Raspberry Pi in particular brought us a lot of challenges. Because of faulty hardware, lack of storage, and other issues, we were unable to use the originally planned Logitech camera and had to use a laptop instead. We also had some issues with the dimensions of the CAD models regarding the door unlocking mechanism, but we eventually created a prototype that worked.

Accomplishments that we're proud of

We successfully managed to use facial recognition, build an app with a sleek UI, and connect both. We also managed to create a prototype of a door unlocking mechanism and link the Raspberry Pi so that it can send notifications to a phone.

What we learned

We learned a lot about how to read through documentation and learn how to use new tools quickly. We also learned how to plan out our work so that we didn't spend too much time focusing on one problem while neglecting others. We especially focused on trying to create as deliverable of a product as possible, even if it didn't totally fulfill every single one of our initial goals, which proved to be helpful.

What's next for Dormmate

In the future, we plan on improving the precision and quality of the camera with a Raspberry Pi with more storage. We also hope to make a stronger arm with more torque to push down on even the most stubborn locks.