Design And Implementation Of Fingerprint Location Method Based On 5G Communication System

With the development of society,the daily activities of human beings tend to places such as shopping malls and office buildings,and location-based indoor information services become more and more important.In the 5G fingerprint positioning technology,the gross error caused by the influence of environment,equipment and observation means,the construction of dense point fingerprint database will consume a lot of manpower and material resources,and the large estimated position offset in dynamic positioning and other issues affect the development of 5G indoor fingerprint positioning technology.Based on the 5G fingerprint positioning method,this paper focuses on the design of 5G positioning platform,the method of eliminating observation errors,the construction of a low-cost fingerprint database,and the reduction of dynamic positioning offset,which carries out the following research:(1)This paper analyzes the advantages and disadvantages of different 5G indoor location methods in the actual application scenarios.Combining the distribution characteristics of 5G base stations,a location fingerprint location scheme based on 5G signals is designed,and K-nearest neighbor(KNN)algorithm is used as the fingerprint matching algorithm.According to the characteristics of 5G indoor fingerprint positioning,this paper designs and builds a 5G positioning system platform.This paper analyzes the transmission characteristics of 5G signal,and demonstrates the feasibility of selecting synchronous signal as positioning signal.In this paper,two indoor static location scenes,including LOS scene A and NLOS scene B,verify the effectiveness of the related algorithms.(2)Aiming at the problem of gross error in the observed 5G signal,based on the analysis of the time distribution characteristics of 5G signal,this paper proposes a preprocessing method of 5G signal based on chi square increment robust Kalman filtering algorithm(CI-RKF),which can effectively correct the gross error of 5G signal and improve the stability and reliability of the observed data.The experiment shows that compared with the direct use of the original data,the average positioning accuracy of the two experimental scenarios can be improved by 27.2% and 29.6%respectively by using CI-RKF to preprocess the raw data..(3)In order to reduce the cost of establishing offline fingerprint database,this paper uses Universal Kriging Interpolation(UK)to reconstruct dense point fingerprint database and proposes a Universal Kriging Spatial Interpolation(PLM-UK)based on path loss model.PLM-UK modifies the estimated value of interpolation points through the path loss model.The experiment shows that in the two experimental scenarios,the average positioning accuracy of the dense point fingerprint database reconstructed by UK is increased by 21.1% and 23.4% respectively;Using the dense point fingerprint database reconstructed by PLM-UK,the average positioning accuracy has been improved by 30.5% and 28.2% respectively.(4)Aiming at the problem of large position offset and unstable positioning in dynamic positioning,this paper proposes a weighted KNN algorithm(WKNN-HLI)that integrates historical position information.By extracting the historical position information of mobile devices,the positioning accuracy can be effectively improved.At the same time,this paper designs indoor and outdoor dynamic positioning scenes for experimental verification.Compared with traditional KNN and WKNN algorithms,the average positioning accuracy in indoor environment is improved by 26.7% and16.2% respectively.In the outdoor environment,the average positioning accuracy has increased by 47.2% and 36.4% respectively.