Partition fusion positioning method based on WiFi and geomagnetic fingerprints

Abstract

The invention discloses a partition fusion positioning method based on WiFi and geomagnetic fingerprints. The partition fusion positioning method comprises the following steps: S1, reconstructing offline RSS data in a whole area according to a mean square error minimum principle by using a stacked auto-encoder; S2, after the offline RSS data are reconstructed, disconnecting a decoding layer of the stacked auto-encoder, and storing a coding layer; finely tuning the whole network according to a BP algorithm, so the classification precision is optimized; S3, converting the geomagnetic signal from the carrier coordinate system into a navigation coordinate system; S4, testing the test points; according to the sparsity of the RSS, extracting the features of the RSS by adopting a stacked auto-encoder; then, dividing a scene into specific areas, namely adding label information to RSS samples, and extracting RSS features according to a stacked auto-encoder for classification; and finally, in each region, inputting RSS and geomagnetism as mixed fingerprints into the regression neural network for fitting. According to the algorithm, very good reconstruction, classification and fitting errors can be obtained under the indoor scene condition.

Description

technical field [0001] The present invention belongs to the technical field of wireless communication, machine learning and indoor fingerprint positioning, and relates to a method for fusion positioning of secondary partitions based on WiFi and geomagnetic fingerprints. Background technique [0002] The rapid increase in the number of mobile device users worldwide has spawned many new services and applications, among which location-based services LBSs (Location Based Services) have become ubiquitous, and their demand is growing exponentially, accurate location information in Played an important role in practical application. Compared with other technologies, the ubiquitous computing and location-aware capabilities of mobile units are the most important factors for the rapid emergence of LBSs. [0003] The Global Positioning System (GPS) can provide high-quality and stable outdoor positioning accuracy. However, due to the obstruction of the wall, the indoor GPS signal will b...

AI Technical Summary

Problems solved by technology

[0007] The present invention proposes a partition fusion positioning method based on WiFi and geomagnetic fingerprints. The purpose of the present invention is to solve the following three problems that cause large errors in positioning results under complex large-scale indoor scene conditions: ① In indoor scenes The dimensional space of the fingerprints collected in the network is relatively large, and it is prone to "dimension disaster" in local areas; ②WiFi signal propagation is easily affected by complex indoor environments, resulting in phenomena such as multipath effects, co-channel interference, and non-line-of-sight transmission , that is to say, the signal space and the physical space are not a simple linear mapping relationship; ③ Only relying on WiFi fingerprints cannot maximize the information acquisition function of the rich sensors of smartphones, resulting in low positioning accuracy

Images