Improved Indoor Location System Design And Implementation Based On Wi-Fi

Location Based Service(LBS)plays a very important part in modern production and life.With the rapid development of real-time positioning with GPS and the wide spread of tele-communication,the current outdoor positioning technology has been greatly developed.Whether navigation for car or mobile terminal,in the outdoor environment can get positioning accuracy very high.In contrast,the development of indoor positioning technology is still in the stage of continuous exploration.The main reason is because of the completely indoor environment,GPS signal is weaken in the building,and the various types of signal transmission source used in indoor positioning are affected by multipath propagation,cannot clear a widely applicable signal transmission model.In the commercial,medical,education and other industries a strong demand,indoor positioning technology has become a researching hot point.With the wide use of Wi-Fi technology,almost all of the indoor scene has Wi-Fi signal coverage.Its low cost,low power consumption,extensive coverage of the characteristics,making it a popular indoor positioning researching point.However,due to the interference of the changing of environment in indoor communication process,and the use of signal strength as the positioning calculation,the data itself has a certain accuracy error,so how to overcome this kind of interference on the Wi-Fi signal positioning technology,became a key point of researching.Based on the comparative study of the existing indoor positioning technology,this paper uses Bayesian estimation algorithm to fuse Wi-Fi fingerprint positioning dates and mobile sensor positioning dates.Within fusion,we get a higher positioning accuracy.Under a large number of experiments,the algorithm was well tested.Firstly,this paper studied the performance of Wi-Fi based indoor positioning technology,especially for indoor location with Wi-Fi fingerprint recognition,the characteristics of the principle and technology and the influencing factors in positioning process are more important.K-nearest neighborhood(KNN)method,K-weighted nearest neighborhood(WKNN)method and Bayesian algorithm were used to compute the common algorithms,including the nearest neighborhood(NN)method,the precision of MATLAB simulation,the learning and understanding of the Algorithm differentiation and positioning performance accuracy.For comparison WKNN algorithm is selected for fingerprint location algorithm.Secondly,based on the sensor data localization algorithm,including the data collection,analysis,processing,combined with the sensor data correction Wi-Fi fingerprint positioning was used in location system design.Then the basic structure of the positioning system was designed,and the functional modules as well as the whole flow of the positioning system were designed.Finally,under an actual laboratory environment the paper built to build the experiment including the database creation,online positioning experiments,and positioning results recording,by analyzing and comparing the algorithm performance,results showed that the accuracy of the Wi-Fi fingerprint positioning system combined with the sensor data mentioned in this paper is 36.4% higher than the traditional Wi-Fi fingerprint positioning system,which basically achieves the expected target.