WiFi fingerprint indoor positioning method

Abstract

The invention discloses a WiFi fingerprint indoor positioning method including the following steps of offline processing including (11) defining a positioning area, establishing a location fingerprint database LFDB of L sampling points, (12) performing k-mean clustering on the location fingerprint database LFDB and by taking the European distance as the evaluation criterion of similarity degree, dividing the fingerprint database into fingerprint database with k subclasses KF; and online positioning including (21) calculating the distance from the actual measurement fingerprint lf=(rssi1, rssi2, ...rssin) to each class center, and writing as DIS=[d1, d2,...dk]; (22) finding the class corresponding to the smallest value in DIS, writing as G<SPECIAL>; (23) calculating the distance between the actual measurement fingerprint lf and each fingerprint in the G<SPECIAL>; (24) selecting the reference fingerprints; (25) calculating the weight coefficient of each reference fingerprint; and (26) calculating the position coordinate of the actual measurement fingerprint. The method clusters the RSSI value acquired in the data offline sampling phase by means of the k-mean algorithm and reduces the computational complexity of the fingerprint matching process; and in the online positioning phase, a dividing method is adopted to reduce the positioning error.

Description

technical field [0001] The invention relates to a positioning method, in particular to a WiFi fingerprint indoor positioning method. Background technique [0002] For the research of indoor positioning system, there are already typical indoor positioning systems such as RADAT system, HORUS system, LANDMARC system and so on. The RADAR indoor positioning system is based on the RSSI (Received Signal Strength Indication) indoor propagation model method. By analyzing the correlation between indoor sampling points and RSSI, the attenuation caused by the wall is introduced, and the attenuation factor is estimated by linear regression for compensation. Indoor propagation loss model. Through this compensation model, calculate the angle between the RSSI receiver and the transmitter, and combine the triangulation positioning algorithm to locate the position of the receiver. [0003] The HORUS wireless positioning system includes a clustering module, a discrete space estimation module...

AI Technical Summary

Problems solved by technology

[0006] In order to solve the nearest neighbor algorithm used in the existing WiFi fingerprint indoor positioning, the present invention uses the fingerprint with the largest similarity between the measured fingerprint and the fingerprint database as the positioning result. The algorithm is relatively simple, but because the reference position is single, the interference of irrelevant items is not eliminated. There is a lot of noise, which will cause a great error in the matching of location fingerprints. A WiFi fingerprint indoor positioning method is proposed to solve the above problems.

Images