Aiming at the problem of low localization speed and big positioning fluctuations of indoor Wi Fi fingerprints positioning

use the method of Gauss fitting and averaging by multimetering and smoothing processing on received signal. Use distance as similarity measurement

set a threshold to classify fingerprint database. By improving K nearest neighbor algorithm and on the basis of classification

implement fast matching of K nearest neighbor. The experimental results show that the time efficiency of the classified localization system is greatly improved

with an average decreasing anplitude of 62. 8 %; average error of Wi Fi fingerprint positioning decrease from 4. 17 m to 2. 12 m.