Our design team landed on an interactive floor plan concept, where visitors would be given a lanyard with an IoT enabled device that would track their position in real time. As visitors walked through the space, the lanyards automatically detected areas of special interest nearby and fired off proximity events, which were presented in a timeline feed.

All of this activity was presented in a 3D digital floor plan that could be manipulated and used to search for people or objects using a typical search function or through an Amazon Alexa skill we built to compliment the AWS cloud services we used to engineer some of the IoT solution. Users could also view the historical paths of objects — a feature that could potentially help optimise interior layouts.

We built a fast, cost-effective and scalable backend solution using AWS serverless technology and AWS IoT services. The position tracking data was calculated using Lambda functions and pushed through using MQTT. Position history data was saved periodically to Elastic search, with proximity event states saved to DynamoDB, which Alexa could query simultaneously.

Our frontend team developed the virtual floor plan using three.js with 3D models and animations created in-house. This allowed us to easily deploy to different devices, and quickly push out updates directly over our server without dealing with mobile app update processes.

We chose Decawave sensors as the backbones of our position tracking, but the base Decaware units only operated via Bluetooth, which isn’t the ideal choice for real-time accurate tracking. So we added a feather board with our custom firmware to the tag unit so we could connect and send position data via WiFi. We also created a custom-designed circuit board and 3D printed case to house these components.