Based on the RSSI readings recorded, it is quite evident that there is a huge degree of variance in RSSI readings from the same iBeacon at the same recorded position.
So, an efficient straight forward formula was ruled out. After careful examination of the RSSI readings and its distribution, we felt that with the available variance in recording, only an approximation of the user’s current location can be made.
A machine learning approach was thought out to be a realistic approach to solve this problem. This method contains two parts