The New RFID

After my first day of research, I realized that my RFID system would have to change, again. While RFID can have a range of up to three meters with passive tags, the only readers and tags I found that would work with Arduino had a range of about two inches. This shatters my plan of placing the container anywhere in the fridge without any consideration for the reader. Plus, the tag with the best range and best price is the blank ID card.

Time for some brainstorming. Rather than having the tags permanently attached to the Tupperware, I could have ID card slots on the containers so the tags can be moved around. While the tags would be movable, they would still correspond to only one food item and would be labeled as such. To organize the cards that are not in use, my sister and I came up with two methods. One would be to have a card box that could either attach to the fridge with a magnet or sit on a table next to the fridge. The other would be to have an organizer on the front of the fridge that would layout the cards so you could tell with a glance what you’re out of and where a tag is. If magnets didn’t affect the RFID, I’d have each tag have a magnet so it could stick to the fridge on its own. Ah well.

In terms of use, the user would have to wave the card/container in front of the sensor when that item is being checked in or out. The reader I’m purchasing beeps once it reads the tag, though, so the user can tell when the tag has been read. However, a card would only need to be scanned when the item is being checked out or checked in. It wouldn’t have to be scanned when it is just coming out for normal use. The reader would need to be carefully placed so that it’s easy to scan in/out a new item, while also allowing someone to take an item out without it being scanned. But that’s simply trial and error.

The First Major Revision

While the shelf scale seemed awesome, it needed some revisions. It was pointed out to me that most people don’t need or even desire to know how much is left of every item in their fridge. Most things are easy to remember when they are running low such as mustard. It takes a while to get through the last 10% of a bottle so you only really need to know if you have it at all. The staple items such as milk and eggs are the ones that run out quickly and without much warning. It would be really nice to know if someone else drank half that gallon of milk you bought three days ago. This is where the next revision comes into play.

The shelf scale has disappeared and been replaced by a standalone milk scale with another system to count how many eggs remain.

Screen shot from Tide Pods Commercial

I also decided to revise the RFID system. My original idea had the user scanning each item and attaching a throw away or possibly reusable RFID sticker to the item. This would reference a database to figure out what the item is and assign the unique RFID code to it. However, it was brought to my attention that this may not be the most user friendly. So another idea was brought to the table. Rather than attaching an RFID tag, the user would have relabeled and RFID tagged Tupperware containers for each food item they want tracked. For example, there’d be a “lunch meat” or “cheese” box and when the user purchased it, she would simply put the new lunch meat into the Tupperware and place it in the fridge. There’d be no scanning necessary and no stickers to deal with falling off or loosing.

While the RFID revision is not my personal favorite, I understand that my love of scanning things is not shared by the general population and that this system is much simpler and easier to understand. Plus, it’ll be easier to code. 😛 🙂

The First Idea

Before I wrote my proposal, I had several ideas floating around my head about how to implement a smart fridge. I knew from the beginning that I wanted it to require minimal user input. Ideally, you wouldn’t have to change the way you use a fridge at all. However, the only way to truly to that would be with cameras and image recognition which is extremely difficult, slow, and unreliable as of right now. I knew I would have to use some sort of sensor other than cameras.

I wanted some way to track what was in the fridge. The answer seemed obvious to me. RFID! I had seen a demo of RFID in high school at the Motorola Innovation Studio and it was epic. They were able to read all the tags on a set of clothing on a rack in about one second. First problem, solved.

But I wanted to know more than just if the item was in the fridge. I wanted to know how much of the item I had. So what’s the easiest way to measure how much of something is there? By weight. My initial idea was to have the entire shelf be a weight sensor which would measure the weight before an item was removed and compare it to the weight once the item was returned. The removal would be measured by the RFID part of the system. So basically the fridge would know exactly what was inside and how much there was of that item.

This sounds awesome to me. However, there are a number of issues with this system. First, the scale needs to be sensitive enough to measure the weight difference from one serving of ketchup while measuring an entire shelf full of food. The we
ight of each container, the density of each food item, and amount of food in a container all need to be known. Some of this information already exist in various databases but not all of it. Also, each food item needs to be removed and added one at a time so the system can get the new weight and figure out which item is gone. This system also requires the user to scan each item they buy and put a RFID sticker on it. While these problems don’t make this system impossible, there is a better way to get similar results.