Research And Implementation Of Indoor Positioning System Based On Extended Kalman Filter

With the rapid popularization of wireless LANs and smart mobile devices, location-aware services are changing the usage patterns of mobile devices, and are playing an increasingly important role in people’s social life. Compared to other indoor location algorithms, Wi-Fi indoor positioning technology based on fingerprint with high positioning accuracy, wide coverage, reliability and low cost, meets the requirements of most indoor positioning and it is widely used in the engineering practice. Moreover, the integration of micro-inertial devices(such as acceleration sensors, gyroscopes, etc.) on the mobile devices makes it possible for indoor positioning using inertial technology. Unlike fingerprint positioning, inertial positioning is hard to be interrupted by external environmental factors.In this paper, according to the characteristics of Wi-Fi fingerprint positioning and dead reckoning positioning, Wi-Fi/DR fusion positioning algorithm has been proposed using extended kalman filter algorithm. Several points are included in this paper as follows:1. Analying characteristics of received signal strength and improving the offline database building algorithm based on Gaussian filtering algorithm for higher credibility of fingerprint. Improving the proportional relation between Naive Bayes and reference node spacing and proposing the probability weighted Bayes algorithm to improve the accuracy of indoor positioning.2. Based on extended kalman filter, proposing Wi-Fi/DR fusion positioning algorithm and establishing the state equation and observation equation.3. Based on Netty framework and HTTP persistent connection, building the high concurrency positioning server which supports heartbeat and SSL encrypted transmission on Tencent cloud. The Android client based on html5 is used to test the actual indoor positioning environment.Positioning performance experimental results show that the performance of the proposed WiFi/DR fusion indoor positioning system is better than the Wi-Fi or DR subsystem and the indoor positioning accuracy is improved. System performance experimental results show that the indoor positioning system provides with high concurrency and the maximum number of concurrent access reaches twenty thousand in the stand-alone environment(CPU: core i3 330 M, RAM: 4GB).