We wanted to create a WiFi based indoor tracking system that would change the way buildings are managed.
What it does
omnifi keeps track of everyone walking around your building and provides realtime data about how many people are in each room and how they move about.
It works by installing omnifitrakrs around the building, the installation simply requires a power outlet. omnitrakrs are fun little gadgets that sit around and collect data about movement inside your building.
The use cases are endless. This can be used anywhere from museums to offices or even department stores.
It can be used by department stores to keep an accurate record of traffic throughout the day or week, this means managing staff more efficiently and perhaps having sales only on specific times of the day or moving products that aren't selling to more high traffic locations in their store.
Museums can use this to help with maintenance of exhibits by knowing when to expect low traffic to certain areas in the museum.
omnifi is completely passive meaning that any device that has WiFi enabled gets tracked without any packets sent from our smart omnitrakrs.
How we built it
We utilised Raspberry Pis as our omnitrakrs. We added WiFi dongles to them and set them to monitor mode. We then sniff packets and filter for any changes in the MAC address bank.
Any change is sent to the omnifier, that is, the brain of the operation. The omnifier gets the aggregated data from all omnitrakrs and creates a precise image of movement in the building.
It is also possible to set up a map and see the data in real time spread across the map.
Challenges we ran into
We had a bunch of problems analyzing the network traffic in a way that would give accurate enough results. This was also more difficult since we insisted on not having the client connect to our WiFi AP or even sending any packets to the client. We knew it had to be completely passive for it to have any appeal.
Accomplishments that we're proud of
We're super proud of having overcome the challenges of the tracking accuracy and reliability. Making a tracking device that is completely passive in any environment is always a fun feat to achieve.
What we learned
We learned a lot about the 802.11 protocol and it's intricacies.
What's next for omnifi
Pinpoint accuracy. We already have a game plan (and some code) as to how to achieve this. The only reason it won't be shown within the confines of the hackathon is because we need a bit more time to get a demo of that up and running.
Essentially we want to track people much more accurately than just the room or small area they are in.
Another step forward is by connecting the MAC addresses to actual people. This can be done easily in any store as we can track the MAC address to the checkout and then tie their credit card name to that address. Also, if someone logs in to the store's website from one of the MAC addresses in the database we can make the connection there too. Once we have a name tied to the MAC address we can scrape their social network accounts and see what we want to sell them.