Skip to Main Content
Location estimation as one of the most popular research areas has been recently attended because of wide range of its applications. K Nearest Neighbor (KNN) is a basic deterministic algorithm for locating which is widely used in fingerprinting approach. The performance of the KNN can be improved extensively by employing appropriate selection algorithm. In this paper, a novel algorithm called Dynamic KNN (D-KNN) which uses Angle of Arrival (AOA) and KNN as a hybrid method is proposed. This method comparing with KNN algorithm with constant K, selects the best number of nearest neighbors dynamically. It utilizes the adaptive antenna system to determine the user locative area by intersection of several obtained AOA. The best K neighbors which are located in the determined area can be selected to employ in the KNN. Analysis and simulation results are reported the best overall performance of D-KNN in different conditions.