Abstract. Aiming at the problem of low precision in the positioning stage of fingerprint location positioning technology based on WiFi, through the comparative study of the nearest neighbor classification algorithm (NN),K nearest neighbor classification algorithm (KNN) and Bias algorithm, and then we present a kind of weighted KNN algorithm of based on triangle correction. On the one hand, the algorithm uses the difference between the AP signal intensity value of the node and the fingerprint node to be taken as the weighting factor, and the positioning accuracy of KNN is improved by the contribution ratio of different fingerprint nodes; On the other hand, we further improve the positioning accuracy by selecting three nearest neighbor points which the unknown node must be in the triangle composed of these three points. Finally, the simulation results showed the effectiveness of the algorithm.

Abstract. Aiming at the construction stage of off-line fingerprint database of fingerprint positioning technology based on WiFi, for the construction of traditional fingerprint database needs field acquisition of fingerprint signal widely, it is difficult to achieve in a large area, a method is proposed to construct the fingerprint database based on optimal attenuation factor model. Firstly, the traditional attenuation factor model was optimized, the free space model and the attenuation factor model were combined, and through the signal acquisition of several nodes, it realized the self-adaption of the path loss index in attenuation factor model. Secondly, the method of constructing the fingerprint database by using the optimized attenuation factor model was given. Finally, the optimization is confirmed by simulating with a variety of classical propagation model.