WiFi indoor locating method based on domain clustering

Abstract

The invention discloses a WiFi indoor locating method based on domain clustering. The method comprises the following steps: firstly selecting a plurality of reference points in two different indoor environments to serve as calibration points, collecting the WiFi signal strength information at the calibration points, and associating the signal strength information with the position information of the calibration points to form position fingerprint so as to obtain a position fingerprint library; then collecting the WiFi signal strength information of a test point, performing pre-matching on the WiFi signal strength information of the test point with the position fingerprint library, and finding K adjacent calibration points closest to the test point by using Euclidean distances of the calibration points and the test point; counting the boundary values of the K adjacent calibration points, and figuring out a central point; then locating the test point by using a WKNN or Naive Bayesian classifier; and m AP are observed by the calibration points and the test point in total, obtaining possible positions of the m test points; and finally determining a final position of the test point. Locating results tested in different indoor environments prove that the WiFi indoor locating method provided by the invention has better performance.

Description

technical field [0001] The invention belongs to the technical field of indoor positioning, and in particular relates to a new WIFi fingerprint-based method for indoor position positioning through area clusters. Background technique [0002] In the last two decades, location-based services for more and more personalized indoor environments have stimulated the development of various indoor positioning systems (IPS). A range of different positioning technologies are deployed for indoor position estimation, such as GNSS, ultrasound, ultra-wideband (UWB) and WiFi. There are two main challenges in WiFi indoor positioning technology: multipath effect and non-line-of-sight (NLOS) effect. Therefore, most WiFi indoor positioning systems use fingerprint technology for position estimation, which consists of two stages: a calibration stage (offline stage) and a positioning stage (online stage). The calibration phase builds a fingerprint of the location area by correlating the WiFi sign...

AI Technical Summary

Problems solved by technology

[0004] In general, the performance of the AP selection algorithm is affected by different environments, and the definition of the number of optimal AP subsets is often vague and has a certain impact on the accuracy of position estimation.

Images