The Smart Fridge

Arduino Prototyping

Summer of 2015 I received the Shanahan Summer Projects Grant from Harvey Mudd College to work on my idea for a Smart Fridge. I used Arduino, Twilio, Tembo, and various sensors to build a version 1.0 prototype.

The Challenge

A Smart Fridge that is more than just a tablet on a fridge door. I decided to build one that will track food contents and allow you to get updates when you need it the most, at the grocery store.

The Journey

Learning how to use the Arduino Yun, Twilio, Tembo, a letter scale, RFID tags, and light sensors, and figuring out where to find answers and how to use them.

The Outcome

A Smart Fridge prototype that tracks fridge contents, milk remaining, and a current egg count and will text you the current status. A confidence in Arduino and the knowledge that I can build whatever I put my mind to with enough time.

Texting the Fridge

The Smart Fridge came about from my frustration of always running out of ketchup especially just after I got back from the store. I wanted a way of checking what I had in the fridge while I was at the store. Most of the commercially available “smart fridges” at that point only had WiFi and Pandora. Personally, I consider something to be made “smart” when it integrates technology in order to better achieve its main function. The main function of a fridge is to keep your food fresher for longer. Therefore, any smart features should be focused around your food, not your entertainment.

The Smart Fridge’s main goal was to track the food it contains and report it to an app. My goal for the project was to allow you to text the fridge and have it text back it’s contents. Like so.

I was able to accomplish this through a combination of Twilio, Tembo, and the Arduino Yun. Twilio gave me the account to text from, Tembo gave me the code to get it working, and the Arduino Yun ran everything from its linux side.

The RFID Tracking

In order for the fridge to know what was inside, I used RFID tags which would be attached to each item. For this prototype, the user would have to tap the tag on the RFID Reader when it ran out in order to check out the item. Ideally, the tags wouldn’t need to be tapped but rather the reader would be able to sense all the contents at once. At the time, I wasn’t able to find a tag reader strong enough that could work with Arduino.

The Letter Scale Turned Milk Scale

One of my most frustrating shopping experiences is when I get back home from the grocery store and realize that I only have 1/4 cup of milk left. This is where my milk scale comes in. I built it by retrofitting a letter scale to connect to the Arduino and then using a standard 1 gallon milk crate to set 100%. From there it simply reads the weight of the milk and reports back the percentage remaining.

The Egg Sensor

Finally, I was working on an egg sensor which would count the number of eggs left. This ended up being a big challenge for the Arduino. I tried using 13 light sensors (one measures ambiant light) with a multiplexer to get all the inputs to the Arduino, but it ended up being a mess of wires. I was also using slightly different sized light sensors which made everything just that much more complex. I became especially disheartened when I found an off-the-shelf smart egg container. I bought one to try to retrofit and connect to the Arduino like I did with the letter scale, but it was much more sophisticated than the scale. I ended up running out of time to fix and finish the egg sensor.

Final Outcomes

Overall, I identified a problem, found a solution, and created a version 1.0 prototype. I’m proud of what I was able to accomplish. So much of this I had never attempted before and I had my doubts during the projects that I’d ever get any of it working, but I did. It's funny looking back on it now because I've learned so much more and know that I could do it so much faster and better, but this is how I got here.

For more info and a step by step of the work involved visit which was my blog for the project.