Frigid: Finalist at HackPrinceton

Objective

We wanted to tackle food waste and create a better experience with the Frigid. Analyzes the image of the food placed in its basket using Microsoft's Computer Vision API, Raspberry Pi, and Arduino. 

It places food in specific compartments of the fridge using a laser-cut acrylic railing frame and basket, which is actuated by motors. The plan for a full product is to maintain a specific temperature for each compartment (our prototype uses LEDs as an example of different temperatures) that allows the food to last as long as possible, maintain flavor and nutrition, and either freeze or defrost foods such as meats from the tap of a finger. This also makes it much easier to arrange groceries after buying. 

Using the identification from the vision system, Frigid can also warn you when you are running low on supplies or if the food has gone bad. Notifications over the internet allow you to easily order more food or freeze food that is getting close the expiration date. 

How it was built

As a hardware project, we spent most of our time trying to build our product from scratch and maintain structural rigidity. Material selection is limited as laser cut acrylic is all we had to build the fridge frame, basket holding the food, and compartments for food to be held in. Plastic was stacked and impromptu buttresses were used to increase structure strength due to material limitation.

The trap door is actuated a pulley system and required a torque of 10 lb-ft. Our original servo wasn’t strong enough so we had to replace it.

The Arduino was used to control the cart and door, while the Raspberry Pi with Pi-cam was used for identification of the food.