Research On Indoor Intelligent Positioning Technologies Based On 5G Fingerprint Signal

With the rapid development of the Fifth Generation Mobile Communication(5G)system,people have entered the era of the Internet of Things.Internet of Things(Io T)technologies are widely used in various scenarios.Positioning technology is an important research topic in Io T application research.Currently,global navigation satellite systems can solve most outdoor positioning problems,but achieving high-precision indoor positioning still remains challenging.The characteristics of 5G,such as high traffic density,ultra-high connection density and low delay,can provide better data support for indoor positioning technology,thereby improving the accuracy of positioning.To address the issue of poor validity of historical fingerprint database,high cost of updating fingerprint database and low positioning accuracy in multiple scenes,this thesis conducts research on indoor intelligent positioning algorithms based on 5G fingerprint signal.The relevant work is as follows:(1)The methods and technologies for indoor positioning are studied.Firstly,traditional indoor positioning methods are researched,including time-based positioning methods,angle-based positioning methods,and signal propagation model-based positioning methods.Finally,the fingerprint-based indoor positioning method is studied,introducing the processing method of the offline phase fingerprint database and the real-time positioning method of the online phase.(2)An indoor positioning algorithm based on fingerprint database migration and reconstruction is proposed to improve the effectiveness of historical fingerprint database and reduce fingerprint collection costs.Firstly,based on fusion indicators,multiple source domain fingerprint databases are obtained through reference point selection from different time periods of the historical fingerprint database.Then,based on data analysis,the common features of the source domain are transferred to the unique features of the target domain to reconstruct a new fingerprint database.Finally,the results of each new fingerprint database are combined to achieve high-precision and low-cost indoor positioning.The proposed algorithm is compared and tested using a dataset,and the results are analyzed to verify its effectiveness.(3)An indoor positioning algorithm based on region division and adaptive Weighted K-Nearest Neighbors(WKNN)is proposed to solve the problem of poor real-time positioning and low positioning accuracy in the online phase.Firstly,a region partitioning algorithm based on greedy clustering is proposed.Based on the position and feature distance between reference points,the number of subregions and the region center are determined.And then the sub-region division of the fingerprint database is completed.Then,an adaptive WKNN positioning algorithm based on statistical law correction is proposed.Based on the weighted feature distance and coordinate distance,the optimal K value and the set of neighboring reference points of real-time data are determined.And the statistical law between the neighboring reference points is used to correct the weight to achieve target positioning.Finally,the proposed algorithm is compared and tested in actual deployment scenarios and simulation scenarios to verify its advantages.