Research On Construction Method Of Indoor Radio Fingerprinting Map

With the advancement of information technology and the development of pervasive computing,there is a growing demand for location-based applications and services.In recent years,wireless sensor networks(WSN)have developed rapidly in indoor positioning technology,providing us with an indoor positioning method with low cost,high positioning accuracy and easy implementation.The indoor positioning method based on the radio frequency fingerprint is to match the radio frequency fingerprint collected during online positioning with the location fingerprint map established when offline,so as to realize the position estimation of the target to be located.In the positioning process,the design of the RF fingerprint in the offline training phase will have a direct impact on the positioning accuracy of the online positioning phase,which is the basis of the RF fingerprinting technology.This paper will study the design of RF fingerprint and the dynamic update of fingerprint map.The main work is as follows:(1)In view of the multipath effect in the process of wireless signal propagation,the coupling of RF fingerprint and environment is low,the stability is poor,and the resolution is low.In this paper,the propagation characteristics,communication characteristics and attenuation rules of RF signals are deeply analyzed and on-site.Based on the test data,the correlation between packet reception rate and RSSI and LQI is analyzed.The map construction scheme of RSSI-LQI hybrid fingerprint based on packet reception rate is designed.Comparing different types of position fingerprints,this paper analyzes two performance indicators,stability and resolution,and compares the maximum probability value fingerprint with the typical RSSI mean fingerprint.The experimental results show that in terms of stability and resolution,the RSSI-LQI hybrid fingerprint based on packet reception rate is better than the traditional location fingerprint,which improves the online positioning accuracy of the system.(2)For the problem that the RF signal strength will change with time andthe positioning performance is significantly reduced,this paper presents a method for dynamically updating the RF fingerprint map based on the correlation matrix of the anchor node.The method uses the relationship between fixed anchor nodes to establish a similarity matrix to dynamically update the map,which eliminates the influence of the dynamic time-varying characteristics of the RF signal on the positioning performance.The experimental results show that compared with the static RF fingerprint map for online positioning,the average positioning error of the algorithm is reduced from 1.8m to 1.3m in a complex indoor environment,which further improves the system positioning accuracy.