Research On Self-Calibration Fusion Position Algorithm Based On WLAN And PDR

Location technology is a widely applied high-tech technology.It has huge demand in transportation and commerce,and has unlimited development space in the future.With the development of mobile Internet,mobile device positioning technology has become the focus of people's attention.The outdoor positioning system,such as the Global Positioning System,cannot penetrate the building and is not suitable for indoor positioning.However,indoor location is necessary in some shopping malls,hospitals,warehouses,large parking lots and so on,so indoor location technology has become a research hotspot in recent years.In this paper,a fusion location algorithm based on WIFI/PDR is proposed,and a Wi-Fi fingerprint database updating algorithm is put forward.The effectiveness of the algorithm is verified by building the simulation of the actual experimental environment and software platform.This paper studies the WLAN location technology using fingerprint database,by setting up an experimental environment to established the fingerprint database at offline stage,focus on the positioning effect of different K values in KNN algorithm,realized the high precision positioning algorithm at online stage;researches the PDR algorithm and selects the appropriate step estimation model.Aiming at the problem that PDR system cannot provide users with initial location,this paper proposes a method to determine initial location,which can accurately estimate the initial location of users.In this paper,a positioning algorithm based on variable length sliding window is proposed.Experiments show that the algorithm can achieve smaller positioning error at inflection point,thereby reducing the cumulative error of PDR algorithm and improving positioning accuracy.At offline stage,updating the fingerprint database needs to consume huge manpower and time,so this paper proposes a fingerprint database updating algorithm,using the improved IDW method based on variogram to estimate the fingerprint information of other points in the fingerprint map,according to fingerprint information of the beacons provided by crowd sourcing data,so as to achieve the purpose of rapid updating of fingerprint database.Through the experimental environment,the proposed fusion location algorithm and the fingerprint database updating method are tested and verified by experiments.The simulation analysis shows that the positioning error of the sliding window method proposed in this paper is 1.28 m,which improves the positioning accuracy of about 40% than the PDR algorithm and can reduce the cumulative error made by PDR effectively and improve the accuracy of fingerprint database.In the range of 2m positioning error,the algorithm proposed in this paper increases the cumulative positioning error probability by about 15% than the old fingerprint database before updating,similar to the positioning performance of current measured fingerprint database.Through experimental verification,the fusion positioning algorithm proposed in this paper improves the positioning accuracy.The fingerprint database updating algorithm has the characteristics of small overhead and good positioning effect.