An efficient indoor location system in WLAN based on Database Partition and Euclidean Distance-Weighted Pearson Correlation Coefficient

Abstract

This paper proposes an indoor location system in WLAN based on fingerprint Database Partition and Euclidean Distance-Weighted Pearson Correlation Coefficient which use a new method of partitioning the fingerprint database is PWNN(Nearest Neighbors based on Pearson correlation coefficient and Distance-weighted). This system includes three stages: offline data collection and pretreatment; Online positioning; Fixed positioning. The first stage partitions the fingerprint database in accordance with the maximum signal strength AP (Access Point) to improve the speed of matching. The second stage uses Pearson correlation coefficient to match the signal fingerprint and select the probability of the collection points, then applies NN algorithm and weighted Euclidean distance to estimate the position. The actual system test proves that the fusion algorithm can effectively improve positioning accuracy and greatly shorten positioning time. Thus, it is an effective and valid indoor positioning method.