Inspiration

We wanted to create a Full Stack integrated IoT product that bettered the Consumer Shopping Experience and provides more value to both customer and retailer.

What it does

Hardware unit is attached to a shopping cart and has a touchscreen display and a barcode scanner. Customer scans items and finds out if they're on sale, related items they might want and gets suggested recipes using the product. The store can track popular items/sales/etc, and the server/DB uses machine learning to track what items are being bought with what to continually improve it's recommendations.

How I built it

We have a raspberry pi that simulates the hardware components by taking 12 integer arrays representing UPC barcode values (i.e. 028400091565 = Lay's Classic Potato Chips 1.5 oz bag) since we didn't have a barcode scanner to work with, and it sends the value to a MySQL DB on an AWS cloud server, which identifies the product and checks its relation to other products. The DB uses machine learning in the background to find correlations between product sales to determine what items to recommend. It then sends the product data and related product information back to the Smart Car client device via a web interface with the GUI interface.

Challenges I ran into

We used the Raspberry Pi 3 instead of the Dragonboard 410c because the Dragonboard we checked out was non-functional, and no other working ones were available.

Accomplishments that I'm proud of

Our Full Stack work! We successfully were able to take data in from the Pi, process it in our cloud infrastructure and get the item to show up on the GUI with recommendations.

What I learned

As a team, we all learned new skills for each of our parts of the project. I hadn't ever used Python before, so I had to learn it as I went in order to take input from the keypad connected to the GPIO pins of the Raspberry Pi and push it to the database. Our Database Engineer learned how to create and use an AWS EC2 instance that hosts our database. Our GUI designer learned to use the Flask web service API. Overall we learned a lot about IoT.

What's next for Smart Cart

Integration with customer loyalty programs (such a Kroger Card, for example) that allows the cart to display a custom shopping list tied to the customer, tailor recommended products to the individual customer based not only on overall trends but their personal purchase history, display coupons connected with the card, etc.

Potentially, Smart Cart 2.0 could use sensors to detect the products in the cart and could automatically charge a customers account when the leave the store without having to stop at a checkout line, similarly to Amazon Go.

Share this project:
×

Updates