Indoor positioning method based on SVM and K mean value clustering algorithm

Abstract

The invention adopts a method based on a SVM (Support Vector Machine) and a K mean value clustering algorithm for realizing indoor positioning and aims to improve positioning precision and stability of indoor positioning and to reduce complexity of a positioning algorithm. The main positioning process includes a first step of selecting a plurality of reference points uniformly in a positioning area and collecting RSSI samples on the points, and forming fingerprints based on the RSSI samples and the position coordinates of the reference points, wherein the fingerprints form a fingerprint database; a second step of adopting a K mean value clustering algorithm for the fingerprint database and dividing the positioning area into a plurality of sub areas with smaller ranges automatically; a third step of adopting an SVM algorithm for the positioning sub areas and constructing positioning sub models for each sub area; a fourth step of calculating the mean value of RSSI in each sub area and determining the mean value as the clustering center of said sub area; a fifth step of calculating Eucidean distance between a real time RSSI signal and all the clustering centers of the sub areas, determining the nearest sub area and realizing precise positioning by utilizing the positioning model constructed in the third step.

Description

technical field [0001] The invention relates to an indoor positioning method based on a support vector machine (Support Vector Machine, SVM) and K-means clustering, and belongs to the application field of the Internet of Things. Background technique [0002] With the rise of the Internet of Things technology, location-based services have attracted more and more attention. For outdoor positioning, there are already very mature positioning technologies, and the applications are also very successful, such as the Global Positioning System and Beidou Navigation System. Compared with outdoor positioning, the indoor environment is more complicated, especially when indoor wireless signal propagation, such as multipath effect, shadow effect and other particularities, make mature outdoor positioning technology cannot be directly applied to indoor positioning. For this reason, scholars have studied many positioning methods to achieve indoor positioning. [0003] The support vector mac...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide an indoor positioning method based on SVM and K-means clustering, which is mainly used to solve the problem of high complexity of positioning learning algorithms in large-scale positioning areas, large variation range of RSSI signals, complex mapping between RSSI signals and physical positions, etc. problems, thereby improving the accuracy of indoor positioning and the stability of the results, and at the same time reducing the complexity of the positioning algorithm, speeding up the positioning time, and more suitable for actual scenarios

Images