What it does

How I built it

Our project was built around the arduino and a network of ultrasonic sensors. An individual ultrasonic sensor was used for each row of cans and keeps track of the distance from the back of the cooler to the closest can. The distance is used to calculate the number of cans remaining in each row and allows us to get predictive drink shortage data. We used presentation foam to mount the sensors and microprocessor and tape to anchor the mount to the back of a cooler tray.

Challenges I ran into

We initially had issues getting consistent readings. We thought the cans were perhaps causing interference, or the arduino was not delivering sufficient power to the sensors, that the sensor lines were experiencing crosstalk, that the different cans had varying reflective properties... Eventually, we determined that getting accurate readings was mostly a matter of orienting the sensors at a certain tilt and doing some calibration.

Accomplishments that I'm proud of

We are most proud of the moment when our project was giving us back what we were working for all along. There were many moments of frustration, so it was nice to finally get some validation.

What I learned

Some of the members of this team learned from their teammates to become more patient in the effort of troubleshooting. Sometimes things get frustrating and we just want to rip all of our work apart and start from scratch. But that isn't always the prudent approach.

What's next for UltrasonicCounter

The UltrasonicCounter has great potential for Coke and could provide them with a good approach to remedying their current system of outage notification and introduce a new method for shortage prediction.

Built With

Share this project:


posted an update

Main selling point still remains, being an ultra-sonic sensor, it overcomes the errors in reading from outside light and sound and that it could send out real time information to the cloud using any available infrastructure. It has the capability to store the data and send it across in batches (SD card on board, if need be).

As our team tested that, it can that it can be retrofitted in most coolers without much effort and cost, the manpower cost per cooler is almost negligible.

The model we displayed to the judges was working 100% and ready to go live on the final day. We took care to have no moving parts in our project, thus leaving no room for damaging it while restocking.

The costing of the product goes down to $5 (wholesale) per cooler, including man-power cost to get it in place, which is another great positive.

Have a quick look at the live video of how it works below:

Being a team that didn't know each other before the project, and being from completely varied background, it was a great achievement to accomplish this in couple of days (especially when you start from scratch)

Log in or sign up for Devpost to join the conversation.