WiFi indoor localization based on K-means

Abstract

A large number of studies show that in complex indoor propagation environment, parameters of indoor positioning method for typical applications, such as localization performance of TOA, TDOA, AOA, RSSI method is often less than ideal. In order to reduce the influence of indoor environmental factors on the indoor wireless positioning, improve the positioning accuracy and expand the location area, the indoor wireless positioning method based on WiFi K-means is proposed. The improved distance formula is used to take into account the effect of attribute values, and the difference between different objects can be calculated more accurately. The AP in the position of each room is established by testing the signal strength of different signals. The experimental results show that the precision in location probability of 3 meters is more than 80%, which relative than hard clustering algorithm, positioning accuracy is improved.