Research And Implementation Of Indoor Positioning Technology Based On WiFi

With the rapid development of the location based service, indoor positioning is widely used among the consumers and enterprises, which include navigation and recommendation in the mall, car parking and finding, the health-care of the elderly,self-help navigation in the exhibition hall, intelligent manufacturing, and mine rescue and so on. Because of the complex and variable indoor environment, traditional GPS signals are easy to produce multiple effects such as reflection, refraction and scattering due to the shielding of the building. And this causes that the GPS positioning system can not be normally used indoors. Compared with the outdoor positioning, indoor positioning has higher request for the reliability and stability of the location information. Meanwhile, it is more difficult to be done than outdoor positioning. WiFi technology has developed rapidly in recent years and it has been widely deployed in various public places. WiFi can help overcome the difficulties which GPS positioning can't complete in indoor positioning.Based on the above discussions, in this paper, we will study and analyze the indoor positioning technology, and summarize the advantages and disadvantages that exist in the existing positioning technology. Furthermore, we will advance a kind of indoor positioning design method which is based on WiFi fingerprint. All of the research work in this paper can be divided into five aspects. Firstly, in the off-line sampling phase of the passive positioning, the author puts forward and realizes multipoint sampling. Multipoint sampling can overcome some shortcomings of single point sampling, such as huge workload, difficulties in updating of fingerprint database and complexity of operation and maintenance. Secondly, in the stage of online positioning, in order to reduce the amount of calculation, clustering analysis is usually needed for fingerprint database before the positioning calculation. A number of grid points which have the largest correlation with the positioned user need to be selected and regarded as fingerprint to do matching calculation. According to the actual needs, the author designs a kind of online classified algorithm which is based on the real-time data of AP scanning on each reference point. Thirdly, in the process of positioning matching calculation, localization algorithm is used more traditionally, in which KNN, the probability method based on Gaussian model, the method based on nuclear, and the method based on neural network are included. However, from the perspective of real-time data and positioning, not all the above algorithms conform to the requirement. The author mainly uses the KNN weighted algorithm which is based on the difference value of the RSS. Fourthly, in order to solve the problem that the position coordinates are not continuous, the author designs and realizes the optimization algorithm of Kalman filter. In the end of the paper, the author puts the system into the software implementation. Finaly, the author introduces thesoftware architecture of the system as well as the way of realizing each subsystem.Numerous field experiments have assured the stability and reliability of the system used for indoor positioning. In the static case, the positioning can reach a precision of 3m- 5 m and control the time delay around 3s. This fully confirms the effectiveness of the algorithm design and the software implementation in this paper.