Processing Technology of Multi-sensor Position Fingerprint Information

Aiming at the poor stability of the fingerprint library constructed using single sensor location fingerprint information and low positioning accuracy due to insufficient depth of fingerprint feature mining in traditional location fingerprint information processing technology, a new multi-sensor location fingerprint processing technology is proposed. First, the CSI phase and amplitude of the WiFi signal, the RSS of the 4G cellular network and the geomagnetic information are used as the joint location fingerprint information. Then, the convolutional neural network is used to further process the location fingerprint. Experiments show that the proposed method is superior to the commonly used neural network methods in terms of positioning accuracy and stability.

I. Introduction

Location fingerprint technology [1] is very popular in the field of indoor positioning because of its low-cost and high-precision characteristics. This method uses a set of location fingerprint data collected in different environments to characterize the signal characteristics of the current location, and maps it to the actual environment, and then constructs a location fingerprint library corresponding to the actual environment. Finally, the current location is obtained by matching the collected location information to the fingerprint database. Therefore, how to deal with location fingerprint information is a problem that needs to be solved urgently.