A New Approach in Wireless Sensor Network Fingerprinting Localization by Affinity Propagation and K-Nearest Neighbor Method

Due to the increasing development of wireless sensor network, localization using fingerprint has particular importance, Which includes database called Receive Signal Strength (RSS) vectors as the basis of indoor positioning, In this paper a clustering method called affinity propagation used for clustering datapoints, And then in the online phase, coarse localization and fine localization is utilized, in coarse localization an adaptive method is used with exemplars, eventually in fine localization the similarity of nearest neighbor and k-nearest neighbors is applied. Finally, the simulation results show the improvement of our theory compared with the methods of without clustering.