Learning Indoor Localization using Radio Received Signal Strength

With this research we will investigate a novel machine learning approach to the prediction of location from received signal strength indicators (RSSI) values obtained from these transmitting access points. Indoor localization has been a long-standing problem in recent times and gaining popularity among researchers. In this research we aim to solve this problem in an indoor environment like office buildings using radio received signals strengths. The most popular approach for positioning has been GPS (Global Positioning System). But we all know that it is inadequate when we consider indoor environments. Hence to solve this issue; we make use of the radio received signal strengths. The most common technology used for indoor positioning is Wi-Fi, which uses radio signals as its signal propagation medium. In this research we are proposing to create an indoor localization system radio signal strengths from as low- energy BLE technology from Bluetooth as access points that were easily available, where the locations of these access points will be unknown. The RSSI obtained from these beacons will be used to predict the locations using machine-learning algorithms. For evaluating our theory we are using the classic fingerprinting method as our baseline for the evaluations. To evaluate this we considered the classic algorithm of nearest neighbor, which is used as a classic method for implementing fingerprinting.