Team Braintrust will develop a Wi-Fi-based device positioning system, augmenting an existing system capable of determining a device's location to the highest accuracy possible. The system will receive sets of a client's visible Wi-Fi devices and their relative signal strengths. With this, it determines the device’s location by comparing the received sets to a series of historically visible Wi-Fi devices and their relative signal strengths at known locations. The location returned will be latitude, longitude and altitude and/or a meaningful label about the location. There will be a label-only location, such as a room number, when GPS coordinates for the historical readings are unavailable, a frequent occurrence inside buildings.


The system will be capable of detecting data sets that differ significantly from other data sets from the same client at the same location, preventing the system from using outlier data when determining location. It will also "learn" and "self-heal" by adding new Wi-Fi devices to the database after they have been consistently detected by clients from a known location and by ignoring in calculations Wi-Fi devices that are consistently no longer detected by clients at a known location.

The Project