DevBuilt

Detecting Occupancy Inside Building

Monitoring building occupancy is essential for the future, whether it be for improving energy efficiency, helping businesses grow, or in case of emergencies. Unfortunately, GPS is not very usable indoors and many other methods require extensive hardware. This led us to develop a minimal friction, WiFi based system to map people in the building, with Wireless Access Points (WAPs) in a network serving as our beacons. We also worked on including a computer vision aspect for large scale settings where CCTV is already included.

How did we build this?

Frontend: Used React and the Mapbox API to create the main dashboard. Backend: Used ARP and a GUI Wifi Scanner to get signals/addresses. Did processing in Python and created an API for the data.

Challenges we ran into

A big challenge was actually detecting all the other MAC/IP addresses on the same network as well as their signal strengths. We tried several libraries and methods but none seemed to work as we hoped. In general, we learned more about how the device addresses, access points, and geolocation works.

Accomplishments that we're proud of & What we learned

We are proud that we have done so much in a limited amount of time and learned something with a different technology background. We learned about networking and also played around with 3D map visualizations + CV.

What's next for DevBuilt

Due to time limitations, we weren't able to make the whole pipeline an automatic service and also had to reduce our scope in terms of number of devices to survey. We plan on clarifying these goals further and completing the prototype to product and market it as a service or API.

Applications

1. This can be used for crowd tracking and analysis for a smooth transition of the workforce in office spaces after pandemic struck world 2. This can be used with online restaurant booking apps as an additional feature to monitor and output crowd densities in restaurants 3. This can be used as analytics for business busy hour tracking and help monetize the business better 4. Search and rescue operations in case of disasters would be better with help of this app 5. This could be used for checking up abnormal patterns in cases of future protests