We wanted to tackle the frustrations of inefficient grocery shopping as hungry university students with very little free time. We saw the opportunity to re-invent the everyday grocery cart into an advanced tool for efficient navigation through grocery stores. Due to personal experiences and frustrations having to search every aisle in a grocery store, we were inspired to create a better overall grocery shopping experience. Therefore, we created the InstaKart along with a mobile Android application to incorporate a seamless front-end UI design.
- Daniel Wang, CS at the University of Toronto
- Vanessa Seto, Systems Design Engineering at the University of Waterloo
- Bader Taher, Electrical Engineering at the Florida Institute of Technology
- Rahul Jain, CS at UC Berkeley
What it does
Our solution is a mobile grocery application where the user enters a list of food items to be purchased from the store. Based on the list, our application generates a unique QR code for our raspberry pi cam to take a picture and scan the QR code. Using this code, it would find the shortest path in the grocery store that would pick up all the food listed on the user's food list. This would provide for a better experience in which the user could shop for groceries in a much more efficient manner.
How we built it
We saw many ways to approach this optimization problem, such as Dijkstra's' algorithm for shortest path problem. Since this was handled in our backend in python, we decided to instead use a greedy approach as a heuristic to the Traveling Salesman problem. From the algorithm, we generated arrows in a map.jpg file and created a HTML server hosted on Google cloud platform that displayed the list of grocery items.
In addition, we created a mobile application in which generates a QR code and created a prototype of the grocery cart with the raspberry pi mounted on top of the handles. The materials we used for this part of the project were 80 20 extrusions to build the outer structure of the shopping cart.
Challenges we ran into
- Hosting the HTML server on Google cloud was quite difficult.
- Trying to get the raspberry pi to take a picture of the QR code and scan it correctly took a while because we had to be sure to be still and make sure the lighting was good.
- Generating the map of the supermarket was quite difficult because we had to see where our things were located as various supermarkets had various layouts.
Accomplishments that we're proud of
We are proud that we were able to create a QR code generation on the mobile phone application that we created. Furthermore, it was quite cool to see a prototype of our project and see how we could see it played out in real life. Also, we were proud with the fact that we designed a website in HTML from scratch by only using stack overflow and online resources.
Youtube link: https://youtu.be/LqXaWmNqgH8