An Iterative Weighted KNN (IW-KNN) Based Indoor Localization Method in Bluetooth Low Energy (BLE) Environment

With the dense deployment of low power node such as Internet of Things (IoT), indoor localization is growing rapidly, has gained a lot of interests in commercial applications. This paper proposes an Iterative Weighted KNN (IW-KNN) indoor localization method based on RSSI (Receive Signal Strength indicator) of the BLE (Bluetooth Low Energy), which has a low power consumption, a long life expectancy. An enormous amount of RSSI are collected to build a fingerprints database. IW-KNN has three principal improvement approaches. Firstly, Euclidean distance, Cosine similarity are combined to measure the similarity of two RSSI vectors, which can take both length, direction of vector into consideration. Secondly, unlike traditional KNN (K Nearest Neighbors) which estimates a position by a majority vote of its neighbors, a weighting factor is applied to neighbors for localization. Thirdly, IW-KNN selects different iBeacons to obtain RSSI at each iteration, calculates the mean position as the final result after several iterations. Compared with several indoor localization methods, namely, traditional KNN, Similarity improved KNN, Weighted KNN, IW-KNN can estimate positions effectively, decrease the mean error by 1.5 to 2.7 meters in our experimental environment.