Research On Ship Personnel Location Algorithm Based On WIFI Fingerprint

Currently,countries around the world are vigorously promoting the development of ship intelligence,and ship personnel positioning is an important foundation for intelligent ship personnel management.Only by accurately knowing the location of personnel on board can intelligent services such as personnel control,personnel scheduling,and providing navigation for personnel be realized.Compared with traditional ships,most intelligent ships have a perfect wireless network environment,which provides conditions for using WIFI fingerprint positioning methods to locate ship personnel.This paper focuses on the research of WIFI fingerprint positioning algorithms,combines the characteristics of ship personnel positioning,and optimizes the three stages of WIFI fingerprint positioning: fingerprint collection,fingerprint database construction,and fingerprint matching one by one.A hierarchical WIFI fingerprint positioning algorithm suitable for ship personnel positioning is proposed,providing a new idea for intelligent ship personnel positioning.The main work accomplished in this article is as follows:(1)The technical means and positioning methods used for indoor personnel positioning are summarized,and the advantages and disadvantages of each technical method are analyzed based on the characteristics of ship personnel positioning.(2)Based on the experimental site,a WIFI fingerprint data collection method for setting reference points in different regions is proposed in view of the fact that reference points cannot be set using longitude and latitude coordinates in ship space.Divide the experimental area according to the spatial structure,establish a separate coordinate system in each area,and set the reference point spacing based on the situation of obstacles in each area.(3)For the construction of reference point WIFI fingerprint label,the normal distribution test is used to determine the characteristic AP that conforms to the RSSI value normal distribution,the 3-test principle is used to eliminate residual data,and the weighted average method is used for the residual data to construct reference point fingerprint label,and improve the accuracy of fingerprint label.(4)For the construction of fingerprint database,a method of partitioning is proposed from two levels: spatial structure and data characteristics.According to the spatial structure characteristics of the experimental area,it is divided into three sub regions.For data feature partitioning,a K-means clustering algorithm using Spearman correlation coefficient as a parameter optimization criterion is proposed to divide each sub region into smaller sub fingerprint databases.(5)For fingerprint matching,a WIFI fingerprint hierarchical positioning algorithm based on XGBoost model and KNN algorithm is proposed.The XGBoost model is used to achieve target sub fingerprint library label prediction,and then an improved KNN algorithm is used to achieve accurate positioning in the sub fingerprint library.Using measured data,experiments were conducted on the proposed WIFI fingerprint classification positioning algorithm and traditional WIFI fingerprint positioning algorithm.The experimental results showed that the proposed positioning method can effectively achieve personnel positioning in experimental areas similar to ship structures,and has improved positioning accuracy compared to traditional algorithms.At the same time,it has significantly shortened positioning time and stronger real-time performance.After personnel interference testing,it has been verified that the hierarchical positioning algorithm proposed in this article can still ensure performance even with a small amount of personnel interference.Overall,the hierarchical positioning algorithm proposed in this article is more suitable for ship personnel positioning scenarios.