We wanted to build a platform that utilized IoT and traditional forms of communication (web apps and phones). We looked to a couple projects and finally landed on one that we noticed one day in a local supermarket. We noticed that the form of communications or organizing employees and setting up new checkout lines relied on roving managers and walkie-talkies. We wanted to build an automated system that cut the excess work put upon managers and outdated technology of non-cellular radio communications.

What it does

RowController is a "devkit" for store owners and retailers to build an semi- to wholly automated storefront with tools that monitor and manage customers and employees. RowController at this moment features three distinct products:

(1) a firmware for cameras that tracks customer movement patterns and reports information to (2) a webapp that allows managers to track customer and employee movements in their store and create custom recipes with the camera to automate notifying employees through a (3) mobile phone app that employees use to receive notifications on where to be present at

How I built it

We built three separate programs that inter-mesh with each other using a realtime database. Our camera firmware was built using C++ and uses cameras that track depth (such as the kinect) to track the amount of people within a particular region of a store.

Our webapp was built using a node server and html5/es6/css3. Our node server is also our logic control center, that figures out what employees can be automatically assigned projects based on customer density or line length.

Our iOS app was built with swift.

Challenges I ran into

We ran into a huge challenge in the first day with our camera firmware, originally we were going to use facial tracking, but the issue arises-ed when someone turned their face around. So we decided to build a depth based tracking system that filtered out static objects.

Accomplishments that I'm proud of

We are proud of how much stuff we were able to implement in such a short time.

What I learned

We learned a lot about computer vision when developing the firmware for the camera. We also learned a bit about iOS design, which neither of us really ever studied before.

What's next for Rowcontroller

We want to wholly expand and flesh out the product into a true devkit for retailers, so they could integrate various firmware into things like security cameras and lights as well as expand into better monitoring employees including time worked and other cost saving measures.

Share this project: